HUAWEI COMMUNICATEof GPON and 10G GPON over the current OLT platform, passing all of the required tests for successful co-existence of GPON and 10G GPON in an established ODN. Yi Xiang,

HU

AW

EI CO

MM

UN

ICA

TED

EC 2010 ISSU

E 58

DEC 2010 ISSUE 58

New services call for new financial networks

Taming the tidal wave of HSI traffic

MSC Pool yields attractive ROI

Enterprise Ethernet gets a sharper edge

iVideoIPTV hit primetime

helps

Sponsor: Huawei COMMUNICATE Editorial Board,Huawei Technologies Co., Ltd.

Consultants: Hu Houkun, Xu Zhijun, Xu WenweiYu Chengdong, Ding Yun, Shen JingyangZha Jun, Zhang Hongxi, Zhu Yonggang

Editor-in-Chief: Gao Xianrui ([email protected])

Editors: Pan Tao, Li Xuefeng, Xu Peng, Xue Hua, Xu PingChen Yuhong, Huang Zhuojian, Yao Haifei, Long Ji Zhu Wenli, Fan Ruijuan, Ranajit Sankar DamMike Bossick, Gary Maidment, Zhou Shumin

Contributors: Tang Xinbing, Li Weishi, Gao Ji, Chen JianyunYu Tao, Zhai Junhui, Xue Yongbo, Xie Juan, Li BingLu Shuang, Xu Jianfeng, Zhang Xuehui, Cai TaoWang Kai, Wu Zhaojun, Li Yao, Wu YanningYang Feng, Peng Wei, Han Bin, Zhang Kejing

E-mail: [email protected]

Tel: +86 755 28789348, 28789343

Fax: +86 755 28787923

Address: B1, Huawei Industrial Base, Bantian, Longgang, Shenzhen 518129, China

Publication registration No.: Yue B No.10148

Copyright © Huawei Technologies Co., Ltd. 2010. All rights reserved.No part of this document may be reproduced or transmitted in any form or by any means without prior written consent of Huawei Technologies Co., Ltd.

NO WARRANTYThe contents of this document are for information purpose only, and provided “as is”. Except as required by applicable laws, no warranties of any kind, either express or implied, including but not limited to, the implied warranties of merchantability and fitness for a particular purpose, are made in relation to contents of this document. To the maximum extent permitted by applicable law, in no case shall Huawei Technologies Co., Ltd be liable for any special, incidental, indirect, or consequential damages, or lost profits, business, revenue, data, goodwill or anticipated savings arising out of or in connection with any use of this document.

A fundamental change has arisen in the world economy from the dramatic development of the Internet and mobile communications and IT-CT convergence. To a rapidly globalizing market, enterprises the world over have responded by implementing a global operational strategy – local resources are globally configured to ensure satisfactory service for customers.

Operating in wider geographic areas as a result of business globalization, multinational enterprises require easier communications, which necessitate building global ICT infrastructure to integrate their global operations and share information in-house.

Cloud computing and virtualization have come at the right time to enable data consolidation and facilitate global ICT infrastructure for these enterprises. Serving as a supply center for their ICT applications, a cloud is much more powerful and faster than any standalone computer, over which employees can access ICT resources anywhere at any time and enjoy more secure and smarter services.

ICT is playing an increasingly important role in business operations. A ubiquitous, secure and efficient enterprise network comprises multiple systems, such as CRM, order management, UC and telepresence. These are coherently integrated to raise capability across the board, improving efficiency and customer satisfaction and solidifying the path of sustainable development. For such sectors as government, finance, education, and energy in particular, ICT is definitely infrastructure for day-to-day operations.

ICT-based enterprise networks have gone IP and are becoming ubiquitous, as best embodied in the catchphrase “Simpler IT and easier communication”. In building their own ubiquitous networks, enterprises are often faced with such challenges as network and service expandability and OAM capability. To help them tackle these challenges, Huawei has developed the “less for more” design concept showing its in-depth understanding of enterprise network architecture from perspectives like network width, service depth and OAM.

Huawei’s enterprise network solution has been deployed worldwide, encompassing government, finance, education, and energy. To support its global operations, Huawei has built its own enterprise network serving over 90,000 employees and 200,000 terminals at more than 100 branches, 22 sub-regional offices, 17 R&D centers and 36 training facilities that are distributed in more than 100 countries and regions. The network has helped Huawei improve its operational efficiency and slash its operational costs.

With a wealth of experience in enterprise network research and construction on a global scale, Huawei anticipates strong partnerships with businesses with globalization aims, helping them streamline their IT applications and build state-of-the-art, ubiquitous IP networks.

Network ubiquity for global success

Shen Jingyang

President of Huawei Enterprise Network Market Dept.

02 Huawei showcases its major achievements in IP standardization

News

01 Huawei and Paris Chamber of Commerce and Industry form partnership

08 How to tackle backbone challenges in the post-Moore era

By Cai Tao & Wang Kai

11 Preparing network platformsfor new data centers

By Li Bing

18 Enterprise Ethernet gets a sharper edge

By Pan Haotao & Zhang Fei

21 New services call for new financial networks

By Wu Fei

23 Extranet takes China’se-government nationwide

By Bi Jianzhong & Fang Su

Industry Focus

15 Smart communications, smarter grid

By Bi Jianzhong & Zhang Qiong

13 Global business, global networkA case study of Huawei’s enterprise network

By Wu Fei

03 iVideo helps IPTV hit primetimeBy Chen Haibin

Main Topic

What’s inside:

P.03 P.43

Let’s COMMUNICATE beyond technology and share understandings of the latest industry trends,

successful operational cases, leading technologies and more. Based on in-depth analysis of the

matters that lie close to your heart, we will help you stay on top in the competitive telecom industry.

43 U2000: A fresh IP O&M experienceBy Shen Hong & Suo Weiwei

35 100G stands tall with a strong backbone

By Shen Anle

29 IP and OTN synergy creates enhanced backbone networks

By Liu Hongli

26 Synergizing IP and OTN transportnetworksIncreasing network convergence has sped up network architecture evolution, giving rise to Channelized OTN and Multi-channel Load Balance technologies. Both can be used to creatively synergize IP and optical transport networks while raising network performance.

By Yan Qinghua & Bai Lu

Bearer Network

45 SmartCare: Your intelligent O&M partner

By Yu Bo

41 Taming the tidal wave of HSI traffic

By Wu Gang

How to Operate

39 MSC Pool yields attractive ROIMSC Pool is increasingly popular with operators due to its numerous advantages, such as load balancing and enhanced network reliability. A closer look reveals it possesses a range of benefits that can contribute to an attractive ROI rate.

By Deng Weifeng & Wu Zhaojun

31 Synergetic protection for IP and optical bearer networks

By Yan Qinghua

P15 P.18

DEC 2010 . ISSUE 58

News

Huawei awarded the world’s first commercial frame contract covering LTE TDD technology in Poland

Warsaw, Poland, 18 November

2010, Huawei announced that

it had been awarded a frame

contract to deploy the world's

first commercial LTE TDD network

for Aero2, Poland's leading

mobile broadband operator. With

Huawei's end-to-end LTE TDD/

EPC solution, the network will

allow ultra-speed data rates and

deliver a rich experience and high-

quality mobile broadband services

to Aero2's subscribers. The

network will become operational

in early 2011.

"Aero2 i s commit ted to

providing high-quality mobile

broadband serv ices for our

customers and introducing

cutting-edge telecom technologies

in Poland," said Adam Kuria-

ski, president of Aero2. "We are

confident that with Huawei's

advantages in LTE technology, we

will be able to offer users a rich

communications experience with

the deployment of the LTE TDD

network."

"This milestone demonstrates

that LTE TDD technology is

already mature, stable and reliable

for large-scale deployment," said

Ying Weimin, president of LTE

network, Huawei. "Based on

Huawei's LTE unified platform

supporting both LTE FDD and

LTE TDD, we are confident that

the network will contribute to

Areo2's success in the mobile

broadband era."

Etisalat and Huawei complete 10G GPON full-service test

Abu Dhab i , Un i ted Arab

Emirates, 1 November, 2010,

Huawei announced that it had

successfully completed its 10G

GPON ful l -service test with

Etisalat on the existing network.

As a fully-compatible service,

the 10G GPON technology can

utilize the existing infrastructure

without affecting the ODN.

Esameel Al Hammadi, vice

president/ASR (Access, Satellite,

Radio) of Et isa lat sa id "We

are glad to explore the future

potential of the innovative 10G

GPON technology. Etisalat is

now ready to provide the best

services to the UAE community,

thereby supporting the vision of

contributing, to an e-City and

e-Society during the Knowledge

Age."

During this test, Etisalat and

Huawei verified the coexistence

of GPON and 10G GPON over

the current OLT platform, passing

all of the required tests for

successful co-existence of GPON

and 10G GPON in an established

ODN.

Yi Xiang, president of the

Middle East, Huawei, said, "It

is Huawei's honor to have a

long term strategic partnership

with Etisalat. In the last few

years, we have had successful

partnerships on GPON FTTx

projects. Today we are proud to

have collaborated with Etisalat in

completing the first 10G GPON

full-service test in the Middle

East market. "

Huawei showcases cutting-edge LTE TDD technology at the 16th Asian Games

Guangzhou, China, 24 November,


i t had exclus ively deployed

an LTE TDD trial network for

China Mobile at the 16th Asian

Games in Guangzhou. Visitors

to the Asian Games are able

to enjoy a range of cutting-

edge services, such as mobile

HD v ideoconferenc ing and

surveillance and HD video on

demand. Using a portable digital

video camera with the LTE TDD

module embedded, members

of the media can send real-

time high-resolution photos and

live broadcast video across the

world.

1

Huawei and Paris Chamber of Commerce and Industry form partnership

Paris, France, 5 November 2010,

Huawei signed a partnership

ag reemen t w i t h t he Pa r i s

C h a m b e r o f C o m m e r c e

and Industry (CCIP) that will

contribute to the efforts by

France and China to create more

business opportunities for both

countries.

Apart from helping companies

in Paris, Hauts-de-Seine, Seine-

Saint-Denis and Val-de-Marne

to expand their business in

China, this partnership also has a

special significance for Huawei.

It will help promote even closer

cooperation between Huawei

and the telecommunications

industry in France.

The agreement was signed

by Sun Yafang, chairperson

of the board of Huawei, and

Pierre Simon, president of Paris

Chamber of Commerce and

Industry, during the visit of

Chinese President Hu Jintao to

France.

"It is an honor for Huawei

to forge this very important

p a r t n e r s h i p w i t h s u c h a

prestigious institution in France,"

said Sun.

"The innovative nature of this

partnership should be stressed.

It is a very strong signal for

businesses in our two countries.

It will enhance our cooperation

with one of the world's largest

telecommunications markets, as

Huawei really wants to help French

SMEs to export to the booming

Chinese market," said Simon.

DEC 2010 . ISSUE 58

Huawei wins “Best Contribution to LTE R&D” award in North America

Dallas, Texas, 12 November,

2010, Huawei announced that it

has won the "Best Contribution

to R&D for LTE in North America"

award at LTE North America

2010.

T h i s a w a r d r e c o g n i z e s

Huawei's contributions to LTE

R&D, as wel l as to the LTE

ecosystem through patents,

customer innovation centers,

and labs worldwide.

Ying Weimin, president of

LTE network, Huawei, said,

"Huawei has supported LTE/LTE-

advanced through a growing

R&D undertaking. The award is

the best evidence of Huawei's

strong capability and leading

position in LTE R&D and our

continuous efforts to accelerate

LTE commercialization."

Huawei has invested heavily

in the development of the LTE

technology since 2004, including

a focused commi tment to

North American development

by opening R&D centers in the

United States and Canada.

A s o f O c t o b e r , 2 0 1 0 ,

Huawei was ranked No.1 in the

world with 18 commercial LTE

contracts and more than 70

LTE trials. To date, Huawei has

submitted over 5,800 LTE/EPC

standard proposals to 3GPP and

holds 270 essential LTE patents

in European Telecommunications

Standards Institute (ETSI), ranking

No.1 among all vendors.

Huawei now serves more than one billion GSM users worldwide

Shanghai, China, 1 November,


more than one billion people

– approximately one quarter

of wor ldwide GSM users –

now use Huawei's GSM radio

access networks for their mobile

communication needs. This key

milestone was reached through

the GSM network that Huawei

successfully constructed for MTN

Nigeria.

" I 'm exc i t ed to be one

of the one b i l l ion users of

Huawei's GSM technology,"

said Navi Naidoo, CTO of MTN

Group. "Huawei's partnership

has been integral to MTN's

success over the past few years.

Due in large part to Huawei's

GSM infrastructure, we are

able to provide high-quality

GSM services to 120 mill ion

subscribers across the entire

African continent and several

countries in the Middle East."

"This milestone was achieved

due to the close and mutually

benef ic ia l re lat ionships we

have with our customers," said

Wan Biao, president of wireless

networks, Huawei . "MTN's

subscriber base, among which

GSM users are predominant, has

grown annually by 25 percent,

demonstrating that GSM remains

the best choice to satisfy voice

service demand for mobile users.

We hope that our partnership

with forward-looking operators

like MTN will lead to another

billion users."

2

Huawei & Option sign acquisition agreement for M4S

B r u s s e l s , B e l g i u m , 1 0

November, 2010, Huawei, and

Option the wireless technology

company, officially signed the

agreement for the sale of M4S,

Option's wholly owned 4G RF

semiconductor company, to

Huawei.

Under this agreement, and

in line with the terms of the

announcement made on October

27th, 2010, Huawei will acquire

100% of M4S' outstanding

equ i ty in te res t s on a fu l l y

diluted basis. Jan Callewaert,

CEO of Option commented,

"The signing of this agreement

concludes the process of finding

an industrial partner for M4S

we initiated earlier this year.

M4S will benefit from Huawei's

vast R&D capabilities and be

in the optimal environment to

spearhead the industry with

the most innovative 4G chipset

solutions." Guo Ping, executive

vice president and chairman of

the board of Huawei Device,

said, "The acquisition of M4S

will enable Huawei to provide

even more innovative mobile

broadband devices, accelerate

the commercialization of 4G

services and support Europe's

Digital Agenda ambition. "

Huawei showcases its major achievements in IP standardization

Beijing, China, 8 November

2010, Huawei successfully hosted

the inaugural 2010 IP Technology

Gala in Beijing on Nov.6 and

Nov.7. The gala attracted over

320 participants from more than

50 operators worldwide, including

France Telecom, Comcast and

China Mobile.

Attendees discussed the

key trends in the IP industry,

including IP/MPLS architecture

a n d v i r t u a l i z a t i o n , I P v 6 ,

mobile broadband and video

applications. Huawei shared its

latest developments in these

areas and pointed out that

network flattening, virtualization,

IP intelligence to network edge,

and IPv4-to- IPv6 t rans i t ion

would be the future of IP/MPLS

architecture.

Main Topic

DEC 2010 . ISSUE 58

With the development of video services, users want to watch high quality video content through a terminal of their choice, and at a place and time of their convenience. With more broadband networks being built, more operators are providing IPTV services to meet the diversified video and media program requirements, as IPTV supports strong and real interactions between users and media providers. The Huawei video bearer network solution (iVideo) provides iVSE (including FCC and RET functions) and iRSM to fully guarantee smooth operation of IPTV services.

By Chen Haibin

3

iVideoIPTV hit primetime

helps

iVideo helps IPTV hit primetime

Huawei Communicate

DEC 2010 . ISSUE 58

IPTV trends

P television (IPTV) services have grown fast in recent years. According to a Frost & Sullivan report, there were about 40 million IPTV users at the end of 2009, and the

number will reach 140 million in 2012. At present, the number of IPTV users is growing most rapidly in the United States and Western Europe, while the number of IPTV users in the Asia-Pacific region is expected to grow rapidly in the next few years.

Operators provide IPTV services usually because of the following reasons: Competition from other operators; competitive pressures that mobile broadband services bring to fixed broadband services; providing video as a basic service, binding users with multiple services, and increasing user bandwidth to obtain high ARPU; and breaking the boundaries between the TV and the PC to increase the number of broadband users.

With the increasing demand for video quality, users are not satisfied with standard-definition (SD) videos. More and more high-definition (HD) videos are being provided by Digital TV, IPTV, Web TV, and mobile TV services and being supported by an increasing number of video sources and video terminals. In America, Verizon IPTV offers 150 HD channels. According to surveys, more than half of the TV terminals around the world support HD services.

In addition to HD requirements, users prefer watching programs at a time and place of their convenience. IPTV is not restricted by program schedules; instead, IPTV allows users to enjoy various audio and video programs at any time. Offering an interactive and customized TV-viewing experience, IPTV is gaining popularity. A survey by In-Stat shows that time-shifting, HDTV programs, and video on demand (VOD) are the most popular features among IPTV users.

Video basics and challenges to HDTV bearer networks

Video basics

IPTV performs video compression of received satellite signals, converts the compressed packets to IP packets through IP streaming, and delivers the IP packets through IP networks, thereby fully utilizing the reach and high transmission efficiency

of IP networks. The current mainstream video compression

standards are MPEG-2 and H.264. MPEG-2 is the dominant standard among digital TV and legacy IPTV, whereas the newly emerging standard H.264 is currently supported by some IPTV vendors. H.264 is more sophisticated than MPEG-2 and its compression efficiency is 1.5 to 2 times the latter’s. Thus, more vendors use H.264 video compression, especially for HD videos. Compressed by H.264, SDTV streaming is about 2Mbps, 1080i streaming ranges f rom 6Mbps to 8Mbps, and 1080p streaming ranges from 12Mbps to 16Mbps.

Both MPEG-2 and H.264 use the group of pictures (GoP) structure to compress videos. The GoP comprises I-frames, B-frames, and P-frames. An I-frame, or intra-coded frame, is a baseline frame that contains complete image information. It is the first frame in a GoP. Every GoP has only one I-frame, and cannot be displayed if this I-frame is lost. The I-frame contains the most data traffic and takes up more space than other frames.

In addition, the MPEG-TS encapsulation information of a compressed video packet is also very important. If IPTV data is encapsulated wi th in an MPEG-TS packet , the program specific information (PSI), including the program association table (PAT) and program map table (PMT) information, is used to de-multiplex video, audio, and other data in the transport stream (TS). A set top box (STB) starts decoding only after obtaining PSI and an I-frame.

Challenges to HDTV bearer networks

Videos bring four major challenges to the bearer networks, as they require large bandwidth, low packet loss/corruption ratio, channel change latency of less than a second, and quick fault locating.

Large bandwidth requirements. As VOD services (including TSTV services) grow, more bandwidth is consumed on bearer networks. While SDTV requires 2M, HDTV, super HD and ultra HD require 8M, 50M and 200M respectively.

Sensitive to packet loss and corruption. Due to high video compression, loss of one packet may cause mosaic or other errors in video images, and thus downgrade the user experience. Different video coding formats have different packet loss ratio (PLR) requirements according to DSL-Forum TR-126. The higher the transport stream bit rate, the smaller the required PLR. HD requires that the PLR should be smaller than 1.0E-07.

4

I

Main Topic

DEC 2010 . ISSUE 58

Channel change latency. Analog TV does not use compression technologies; images are immediately displayed on the terminal when they are received. The channel change time is usually less than a second. In the case of IPTV, the terminal device will not start decoding until it obtains the key information such as PAT, PMT, and an I-frame. As a result, the channel change latency is unstable, and the quality of user experience cannot be ensured. If without any special measure, the latency could reach 5 seconds, and it will seriously impact the user experience.

C o m p l i c a t e d o p e r a t i o n a n d maintenance. Video operation and maintenance involves O&M of IPTV, metropolitan area network (MAN), and access network. When an IPTV service has problems, the fault must be effectively and quickly located and cleared to guarantee the user experience.

HDTV bearer network solution

Huawei provides an innovative video solution called iVideo, which comprises i V S E a n d i R S M , t o a d d re s s t h e challenges to the video bearer network.

The integrated Value-added Service Enhancement (iVSE) solution uses built-in video cards of network devices to implement low-cost fast channel change (FCC) and application layer re-transmission (RET) and ensure high-quality video experience.

The in-line Real Stream Monitoring (iRSM) solution delivers low-cost and effective video operation and maintenance. The N2510 monitors the quality of video streaming on the last-mile network. The U2520 monitors the video streaming quality of the MAN. The Media Quality Monitor Center (MQMC) monitors the video streaming quality at the IPTV head-end and STBs. The MQMC cooperates with the U2520 and N2510 to deliver a unified end-to-end video operation and maintenance solution.

sends a packet retransmission request through RTCP (based on RFC 4585) to the iVSE card. Upon receiving this request, the iVSE card finds the packet according to the requested SN value, and retransmits the packet to the STB. The STB inserts this packet into the proper position in the cache and sends the packet to the decoder, thus fixing the video stream.

The RET mechanism suppor ts single-card multi-level deployment. If packet loss occurs between two video cards, a lower-level video card can serve as an RET client and request the upper-level video card to retransmit the lost packets. If the IPTV system supports the RET function, the video card considers the IPTV system as the RET server and sends a retransmission request to the IPTV system. In this case, the upper-level video card can instruct, through multicast, the STBs not to send retransmission requests. This prevents a large number of retransmission requests from being sent to the video card. Upon receiving the retransmitted packet, the video card can multicast this packet to the STBs.

As required by the RET mechanism, video packets should be encapsulated within RTP, and STBs should support retransmission. Since most legacy IPTV systems output video streams without RTP encapsulation, the IPTV systems need to be upgraded or installed with an independent RTP encapsulation device to meet the RET requirement.

For STBs to support RET, Huawei provides two solutions. One is to load SDK software to the STBs simply by binding the API interfaces. Currently, all the STBs Huawei offers for IPTV systems are integrated with this function. The other one is to cooperate with a third-party company. For example, in the case of UAE-based operator Etisalat, STBs from four vendors already exist on the telco’s network. To introduce a new IPTV system and deliver FCC and RET functions, Etisalat employed a third-party company named Soft@home to integrate the IPTV system and FCC

5

Solving large bandwidth requirements

Huawei has adopted two approaches to cut down bandwidth usage. One is to deploy multicast points as close to users as possible and leverage multicast technology. The other is to deploy content delivery networks (CDNs) for VOD and TSTV unicast services, where the content is delivered as close to users as possible and storage used to save bandwidth. iVSE uses built-in video cache cards to expand CDNs to the user side, thus resulting in savings in both cost and bandwidth, and an improved user experience.

Video cache cards can be deployed in different positions to achieve optimum balance between card-deployment cost and bandwidth saving. Huawei provides flexible video cache card deployment solutions that are suitable for various scenarios and applications, with support from Huawei’s series of routers and OLT.

Reducing packet loss/corruption ratio: RET

Videos are very sensitive to packet l o s s a n d c o r r u p t i o n . Fo r I P T V, there are various end-to-end quality factors that may cause serious packet loss or corruption, l ike aging and in t e r f e rence o f l a s t mi l e coppe r cables, family networking, and MAN congestion. Huawei has the automatic r e t r a n s m i s s i o n r e q u e s t ( R E T ) mechanism at the application layer to help reduce the packet loss/corruption ratio and remove the video mosaic.

The mechanism reveals whether a s t ream loses any packet v ia the continuity of sequence number (SN) value in the RTP header, as the SN value increases by one when one more IP packet is added to the same stream. Specifically, Huawei iVSE caches every video stream for one or two seconds on the video cache card. The STB receives, caches, and checks the video packets. Upon detecting packet loss, the STB


Huawei Communicate

DEC 2010 . ISSUE 58

and RET functions with the existing STBs.

Solving channel change requirements: FCC

The channel change time is greatly affected by the waiting time of PSI (including PAT and PMT) and I-frame. Based on the standard RTCP protocol, the FCC mechanism works by decreasing the waiting time for PSI and I-frame and thus cutting the channel change time to within a second.

When transmitting IPTV video streams, the network device uses the built-in video card to cache the video stream of each channel for a certain time interval. The cached information includes two or three GoPs and PSI. When the user changes the channel, the STB sends a unicast channel change request based on extended RTCP to the video card. The video card then sends the cached PSI of the new channel to the STB. This effectively reduces the PSI waiting time. Then the video card unicasts the content of the new channel (starting with the I-frame) to the STB. The STB immediately decodes the content. The entire channel change time is therefore less than a second.

The unicast streams are sent out at a faster speed than the normal multicast speed, to keep pace with the subsequent multicast streams. When the video card decides that the unicast streams are keeping pace with the normal multicast streams, it sends RTCP signaling to instruct the STB to join the multicast group of the new channel. Meanwhile, the video card slows down the sending speed of the unicast streams by half of the normal speed and stops sending unicast streams. The unicast streams and the multicast streams connect to each other, even as the user is oblivious of this process.

Huawei iVSE uses a special method to decide when to stop sending unicast streams, and when to connect the unicast streams to the multicast streams without packet loss. When the STB receives multicast packets, the STB sends the SN value of the first multicast packet to the video card. Then the video card stops sending unicast packets after sending out the packet with the SN value. This method ensures that the unicast packets connect to the multicast packets without packet loss.

FCC consumes large network bandwidth, especially when many users change channels simultaneously. Fast unicast also has higher requirements for last mile bandwidth. For example,

some vendors require that the last mile bandwidth must be 1.3 times the normal bandwidth. This is a huge challenge to the access network bandwidth resources. If the operator provides 8Mbps HD video bandwidth and 10Mbps access bandwidth, FCC cannot be deployed.

To address the bandwidth challenge, Huawei uses OLT, the CX600, and the NE series routers to provide video cards for FCC. Operators can deploy the video cards at suitable positions as needed. In addition, Huawei provides an intelligent method to discard unimportant video packets during FCC to save bandwidth without affecting user experience. With this method, the last mile network bandwidth can be just 1.1 times the normal bandwidth, and MAN bandwidth is cut by 30%.

Like RET, FCC requires that video streams be encapsulated within RTP, and the STBs support the FCC process. Huawei provides SDK software for upgrading of STBs. And the IPTV systems and STBs offered by Huawei can cooperate with iVSE to support FCC. Huawei also promotes the FCC standard process in the IETF. A draft in this regard has already been published. When an FCC standard is published, STB vendors only need to comply with this standard to support FCC.

In addition, to implement FCC and RET, STBs must know the IP address of every video card that corresponds to each channel. When iVSE cooperates with Huawei IPTV systems, STBs obtain the IP addresses of iVSE cards through the IPTV system configuration. When iVSE cooperates with other IPTV systems, STBs contact the iVSE center provided by Huawei to obtain the IP addresses of video cards for delivering FCC/RET. The iVSE center is preconfigured with channel information and the IP addresses of the corresponding channel cards.

Solving operation and maintenance requirements: iRSM

IPTV operation and maintenance is very complicated, When a problem occurs, it is hard to locate the fault. Moreover, a small error on the network or head-end can have serious impact on user experience. To avoid such situation, operators must quickly locate and clear the fault to ensure high-quality user experience.

Huawei provides iRSM to quickly detect and locate faults. The iRSM uses the built-in software monitoring module to deliver low-cost and effective video operation and maintenance, uses Mean

6

Main Topic

DEC 2010 . ISSUE 58

Opinion Score-V (MOS-V) and Real Media Rate (RMR) detection to quickly detect faults, and uses media delivery index (MDI)/RTP detection to quickly locate faults. All monitoring information is delivered to the U2520 and is displayed on the U2520 to facilitate IPTV operation and maintenance.

Fault detection

MOS-V and RMR are deployed at the egress of video stream to detect the quality of video streams output from the IPTV head-end. MOS-V determines the quality of the video streams and shows the output quality of the head-end, but cannot determine whether the output rate is proper. If the output IPTV video streams are of good quality but the burst rate exceeds the upper limit of the bearer network, the output video streams are also determined as unqualified. The RMR detection shows the rate changes of every video stream in real time, and helps detecting whether a rate burst occurs in the output IPTV code streams, and whether the burst rate exceeds the threshold predefined according to the network bandwidth and lining capacity.

Fault location

The media delivery index (MDI) is an algorithm that uses the MPEG-TS encapsulation information to measure the transmission quality of IPTV video streams over IP networks, and it applies to all IPTV video streams that are encapsulated within MPEG-TS packets.

It comprises two parameters: the media loss rate (MLR) and delay factor (DF). MLR indicates the rate at which packets of the tested video streams are discarded during transmission and thus the ideal MLR during the transmission of IP video streams is 0. DF indicates the delay and jitter of the tested video streams in the unit of millisecond. When the time of video contents contained in the buffer of network devices and decoders is equal to or greater than the DF value of tested video streams, the video display quality does not decline.

The MDI can easily and quickly obtain the packet-loss ratio and latencies of every video stream based on the SN value in RTP header information. The Timestamp field of the RTP packet can be compared with the local time to show the unidirectional latency of the arriving packet (depending on the network time synchronization).

End-to-end service quality monitoring

Huawei video bearer network solution can also closely work with Huawei IPTV system monitoring solution. The media quality monitor center (MQMC) monitors the video quality of the IPTV head-end and STBs, the U2520 monitors the video quality of each network node, and the N2510 monitors the video quality of the last mile network. The N2510, U2520, and MQMC can communicate with one another, and the MQMC shows the end-to-end video quality.

The end-to-end service quality monitoring (SQM) solution has the following advantages:

Real end-to-end: The network device uses a built-in SQM module to monitor the network layer quality. The video server and STB use built-in probes to monitor the service layer quality. Thus, the real end-to-end service quality monitoring from the IPTV head-end to the end users is implemented.

Unified management platform: The MQMC seamlessly connects to the U2520, and functions as a unified management platform for network layer monitoring and service layer monitoring. This management platform delivers precise and unified monitoring indices to facilitate user management and fault locating.

Reduced TCO: The built-in probes in the head-end and the STB provide the most precise and real IPTV quality monitoring indices. The built-in SQM module in a network device saves space and network interfaces, and reduces the maintenance cost. The built-in SQM module reduces CAPEX and OPEX in aspects of deployment, maintenance, and operation.

Huawei’s iVSE + iRSM solution provides FCC, RET, and end-to-end SQM functions, serves as a full-dimensional integrated video solution for service providers, reduces their CAPEX and OPEX, and ensures the high-quality video experience. This solution has been utilized in constructions of more than 100 backbone networks and 600 MANs, and is deployed by the world’s leading operators, including China Telecom, China Mobile, China Unicom, France Telecom, Deutsche Telekom, British Telecom, Telefonica, SingTel, and Etisalat.

iVideo Helping IPTV hit primetime


7

Thanks for your reading, the electronic version and subscription information are available at www.huawei.com/communicate. Questions and suggestions may be directed to the editor concerned.

Editor: Zhu Wenli [email protected]

Huawei Communicate

DEC 2010 . ISSUE 58 8

How to tackle backbone challenges

in the post-Moore era

n the new era, operators are under huge pressure to expand backbone capacity. Another challenge is greater impact resulting from

network failures given the expansion. On top of it all, continuous expansion may increasingly complicate network O&M. How should operators take on these challenges?

Surging traffic brings about expansion pressure

Broadband services such as video have sparked a surge in user demand for Internet bandwidth, necessitating backbone capacity expansion that is often expensive. Furthermore, ARPU perMbps continues to fall and some Internet applications are siphoning off part of the profits from telecom services, which in turn discourages operators from seriously investing in capacity expansion.

In r e spons e t o th i s mount ing expansion pressure, the fol lowing options may be feasible: improving node capacity, optimizing network architecture and collaboratively planning network layers.

Increasing node capacity

Amid China’s Internet boom, Internet traffic has trebled every 18 months. In comparison, routers, which act as Internet hubs, have struggled to keep up, taking two years to expand in capacity from 10G to 40G and seven years from 40G to 100G. With Internet traffic growth outpacing the Moore’s Law-defined router capacity, it is obvious that backbone networks are now in the post-Moore era.

By Cai Tao & Wang Kai

I Raising the port rate and equipment capacity is the most common and direct way of network expansion. 40Gbps is now commercia l ly avai lable , but as t ra f f ic cont inues to grow, 100Gbps will be prevalent. Increased port rates require larger-capacity equipment, which means Tb routers and OTN equipment will emerge as a result. Nevertheless, this expansion approach will pass huge capacity and cost pressure on to core nodes. In expanding backbone capacity, therefore, operators should consider optimizing network architecture in addition to improving node capacity.

Optimizing network architecture

T h e t r a d i t i o n a l hierarchical network is a cost-effective networking model when network traffic is low: Routers can perform traf f ic convergence through statistical multiplexing and deliver the scalability required for full meshing. As network traffic grows, core routers must be expanded frequently, which may hinder network growth. On the backbone, more than 50% of the traffic

Main Topic

DEC 2010 . ISSUE 58

is in transit. Forwarded by routers many times, the traffic occupies expensive router ports, and in this case, the hierarchical networking model is no longer cost-effective.

When traffic between two routers approaches link bandwidth, router-based traffic convergence becomes useless and direct router connection applies instead. When intervening convergence routers are replaced, the number of network layers is reduced. The backbone network becomes flat when only one hop is left between two routers.

In a hierarchical network, services are forwarded hop by hop. Thus, each packet is forwarded in the way car traffic at a crossing is controlled by traffic lights. In this model, performance is higher when traffic is low. In contrast, direct links between routers at the optical layer are like flyovers without traffic lights, through which large-capacity services can be dispatched swiftly to deliver strong performance despite heavy traffic.

Benefits from directly-connected routers are many. They include: simplifying the network structure; easing capacity expansion pressure; saving internetworking costs; improving network scalability and reducing network delays and jitters while improving QoS.

China Telecom is a good case in point. On its 163 backbone, the operator has gradually elevated the status of its high-traffic access nodes, such as the node in Hunan Province, to quasi-core nodes. In 2009, the original nine core nodes increased to 15 quasi-core and core nodes. The number is expected to grow in the coming years so that the network

9

How to tackle backbone challenges in the post-Moore era

structure can flatten.

Multilayer planning

Growing traffic requires networks to become flat. But what should be done to make this happen effectively?

The actual flattening of the 163 backbone is based on observation and experience. The industry is currently looking to multilayer planning for deploying a flat network more accurately and efficiently. While deploying or optimizing a network, operators will create direct links between routers according to results derived from multilayer planning tools.

Traditionally, networks are planned layer by layer, while router and transmission networks are planned separately. Due to limited information exchange, full-mesh planning is not the most cost-efficient. Operators believe that the use of expensive routers and ports in a network should be minimized because routers are costly and not all traffic has to be handled by routers. Actually, direct links between edge routers can be created to reduce the need for core router ports and cut equipment costs.

Multilayer planning has been developed to plan router and transmission networks at the same time, allowing resources to be appropriately allocated within the networks. Small-granularity services are dispatched by core routers to improve dispatching efficiency through statistical multiplexing, while large-granularity ones are transmitted over the optical layer. As such, overall network efficiency is improved. During network flattening, the

Fig.1 Backbone service matrix

100%

year

95%

90%

85%

80%

75%

70%

65%

60%

55%

50%

98.4%

87.7%

80.0%

74.3%

69.8%66.6%

63.9%61.6%

1st 2nd 3rd 4th 5th 6th 7th 8th

Huawei Communicate

DEC 2010 . ISSUE 58 10

multilayer planning tool works more accurately and efficiently than empirical methods and supports network evolution. Moreover, the tool can effectively analyze the overall TCO and simplify network expansion through incremental planning.

Let us look at a service matrix (data for the first year) shown in Fig. 1. The matrix grows 60% annually. The X-axis represents data for future years, while the Y-axis indicates the percentages of costs for the flat networking relative to hierarchical networking. Data in the figure shows that the flat networking costs are lower than those for hierarchical networking and that the cost difference widens as traffic grows. In Year 8, the costs incurred by flat networking represent only 61.62% of those for hierarchical networking. Moreover, power consumption and equipment space also fall sharply under the flat networking model. Based on the results from the multilayer planning tool, numerous scenarios are ideally linked up between edge routers with sub-wavelength signals of N x 1 GB or N x 2.5 GB, further enhancing network efficiency, though this necessitates additional communication between routers and OTN Internet ports.

As network planning is cyclical and data services are bursty, dynamic network planning is required. Dynamic planning necessitates traffic detection, multilayer PCE and bypass servers. Due to complex standardization, this type of planning is currently not feasible. Combined with traffic detection and NMS operations, semi-automatic network flattening may be a practical option for the moment.

Extensive network failures undermine customer loyalty

At present, 40G technology is commercially viable. As one single 40G link carries thousands of services, a link failure may affect many services and seriously compromise customer loyalty. Operators have to increase backbone reliability as a response.

As router recovery and protection technology advances, routers can, in theory, protect services during network downtime. In actual network rollouts, a path for rapid recovery and protection is hard to come by due to limited transmission resources. Moreover, transmission link failures may trigger a massive number of router alarms and route flaps, causing adverse impact on routers.

ASON technology can remove multipoint failures on the transport network through fast rerouting at the optical layer. Due to the lack of

support from routers, however, protection may fail or multiple protections may occur. GMPLS UNI interfaces and routers combine to provide collaborative protection for a multilayer network. This approach can speed up service protection and reduce unwanted network resources; it can also enhance backbone reliability through link groups that share risks in multilayer network planning.

Difficulty with O&M and clearance of mass alarms

Let us assume that an operator needs to handle over 6,000 alarms relating to wavelength division on a daily basis. At the router layer, the number will multiply to exceed 20,000, which may cover up the root alarms and make it impossible to identify the root cause in time. This is because one alarm at the transport layer often generates more than 10 times as many at the IP layer. As the IP and transport layers are managed by different departments, the relationship between alarms at these two layers becomes unclear. Through collaborative O&M at the IP and optical layers and a unified database, the bearer and alarm correlations between the two layers can be easily maintained to mask a huge number of derived alarms and swiftly identify the root alarms. This helps shorten the troubleshooting time from hours to minutes.

Through a combination of unified NMS and GMPLS UNI, the operator can reduce service rollout time from months to days, and are able to better respond to customer requirements and secure more market share.

In an era when traffic outpaces Moore’s Law, the operator is challenged to adapt by increasing single-node capacity and optimizing the network architecture. To increase network efficiency and reduce network costs, the collaborative IP/optical multilayer planning tool is highly instrumental and effective in implementing the flat network structure that is becoming a trend. Multilayer collaborative protection and O&M help operators considerably improve network reliability and O&M efficiency and cut O&M costs. In short, collaboration at the IP and optical layers in terms of traffic, protection, and O&M represents the best response to challenges facing backbone networks in the post-Moore era.

Editor: Chen Yuhong [email protected]


Main Topic

DEC 2010 . ISSUE 5811

Preparing network platforms for new data centers

Preparing network platforms

for new data centers

By Li Bing

in f ra s t ruc ture s , which re su l t s in inaccessible islands geared to individual projects. Equally, sharing is impossible among the huge number of storage systems that directly connect to devices and servers.

Finally, traditional DCs are expensive, energy-hungry, and wasteful. These problems are exacerbated as current demands continue to force them to expand; for example, it is estimated that annual spending on power will exceed the actual construction costs of a traditional DC network within three years.

Moreover, the financial crisis has made enterprises particularly sensitive to IT costs, while increased social responsibility has di scouraged excess ive carbon footprints. The time is thus ripe for a new wave of data centers.

New data centers enabled by cloud

Gartner ranks cloud computing as the top nascent technology for 2010 and DC restructuring as the fifth. Cloud computing enables data consolidation and enhances flexibility, simplicity and efficiency for enterprises. It does this by consolidating and virtualizing b a s i c c o m p u t i n g , n e t w o rk , a n d storage devices and resources under a unified infrastructure. Enterprises can dynamically allocate DC resources to different applications to boost operational efficiency. Research firm IDC has found that this trend is broadly supported by CIOs, which lays the foundation for constructing new-generation DCs based on cloud computing over the next few years.

Today's information age has propelled data centers (DCs) to the forefront of the national infrastructure. They are equal to transportation systems and energy resources in terms of importance, and form the heart of commercial operations. Positioning DCs within a strong and centrally managed infrastructure, however, has its own set of challenges.

Challenges facing traditional DCs

irst, system resource utilization is consistently poor – hovering below 30% on average – as traditional DCs were designed

for peak performance with no downtime. With the deployment of physical servers, storage devices, and other network devices for peak times, most enterprise DCs remain sluggish, inflexible, and unable to achieve the dynamic data allocation required by modern business. Second, traditional DCs exist within an

isolated structure that impedes both resource-sharing and

information-sharing. Applications are

closely coupled w i t h t h e i r

underlying

F

Huawei Communicate

DEC 2010 . ISSUE 58

The “cloud” enables DCs to provide both data and application services by serving as a supply center for various IT applications, including storage, office software, and databases. Virtualization t e c h n o l o g y i s c r u c i a l t o c l o u d computing as it can boost the flexibility of computing and resource allocation to better serve c loud computing. Virtualized and cloud-computing-based DCs are also much cheaper and easier to deploy, maintain, and manage, and significantly reduce energy consumption as well.

The network platform is the key

Network platforms must adapt to the new wave of DCs. Similar to a highway mesh, network platforms connect various DC devices, such as servers and storage devices. These highways are the channels through which data is read, written and transferred between servers and storage devices. Any congestion or breakdown will severely compromise DC efficiency, and can even cause complete collapse.

The new DC positions the network platform as the foundation for cloud computing and storage. To evolve a data center to a cloud-based system, the first and most important step is to renovate the network platform. A simple and scalable platform will facilitate computing and storage virtualization, optimize load-balancing, and maximize efficiency. Otherwise, operational efficiency and resource utilization will still remain low even with cloud computing. Three issues must be addressed to build such a network platform.

Adapting and optimizing the computing and storage model

New and traditional data centers differ widely in how they read and transfer data. The new DC integrates servers and storage devices so that reading, writing and synchronizing data among them is a constant and continuous process that

wide reliability is achieved by ensuring seamless switchover during a network failure, and by offering online support to enable capacity expansion and upgrades without interrupting services for enterprises.

Addressing both current and future needs

It i s e s t ima t ed tha t , f o r l a r g e enterprises, the new DCs will render traditional DCs obsolete within a few years. As a leading datacom equipment provider, Huawei has a strong history of innovating and developing DC t e chno log i e s . Suppor t ing c l oud computing and virtualization, Huawei’s network platform solution is based on its S12700 series of core switches. These employ CLOS non-blocking architecture to forward massive amounts of data at wire speed, and form the core of the new generation of DCs.

For more cost- sens i t ive SMEs, t radi t ional DCs wi l l cont inue to dominate for some time. Based on this, Huawei has also developed a network platform solution based on its S9300 core switch series. These boast the highest-density 10GB ports in the industry, and support wire speed forwarding. Not only can the S9300 series accommodate the huge growth in data traffic, but its eco-friendly design uses just 33% of the power required for similar products, thus helping customers install a green, traditional DC.

Currently, Huawei’s full lineup of datacom devices – including routers, switches, and firewalls – operate in more than 120 countries. Its related products and solutions are widely utilized by leading operators and enterprises, including China Unicom, Russia’s MegaFon, Baidu.com, and Bank of China.

12

Editor: Xu Ping [email protected]

enables rapid data access for users.Data access speed is adversely affected

as network complexity increases. The new DC network platform reduces the number of layers from three to two, and increases devices’ cache capacities. As servers and disk arrays access data based on MAC addresses, a more complex DC increases the size of MAC address tables, which causes bottlenecks. Therefore, larger caches and MAC address tables are essential for new DC switches.

Keeping up with traffic spikes

With the growing popularity of fixed broadband, 3G and Wi-Fi access, China’s broadband users topped 360 million by June 2010. P2P and video traffic is soaring, and Internet backbone traffic triples every 18 months. This huge volume of traffic is bound for DC networks and servers; thus, it is common to have DCs with 10,000 servers, and 10GB network adaptors have now become the industry mainstream. However, future developments are likely to cause bottlenecks in the forwarding performance, port density, and network device capacity of DCs. Devices must come equipped with CLOS non-blocking switching fabric and high-density 10GB ports to forward mass services at wire speed.

24/7 operations

Statistics show that USD1 million is lost if a DC shuts down for an hour, and that 70% of DC failures can be attributed to the network platform. The power, communicat ion, and manufacturing sectors are particularly affected by failures, and urgently require a stable platform.

A network platform has two levels of reliability to ensure that businesses can run 24/7: the device and the network. The basic requirements for device reliability can be met by separating the main control and service processing layers, and by providing redundancy for all key components. In turn, network-

The electronic version and subscription information are available at www.huawei.com/communicate. Questions and suggestions may be directed to the editor concerned.

Main Topic

DEC 2010 . ISSUE 58

Global businessglobal network

By Wu Fei

13

Global business, global network

ultinational companies looking to enter new markets face unique communication requirements. With geography and network environments varying considerably from country to country, branches

and subsidiaries of the company must precisely convey localized requirements to head offices and coordinate cross-border employees if they wish to establish a cohesive multinational framework. And all of this requires a ubiquitous enterprise network.

As a global leader in the telecom field, Huawei sees its products and services present in more than 100 countries, touching the lives of one third of the world’s inhabitants. Huawei’s global success could not have been possible without its enterprise network, which can provide a template for companies with similar multinational ambitions.

Data center network: Efficient and reliableState-of-the-art data centers (DCs) underpin Huawei’s base

of operations in Shenzhen. The primary device covers more than 4000m2, and powers 1,000 servers and 2,000 types of IT systems. Its Layer 2, flat and virtualized, utilizing a cutting-edge network design based on cloud computing, minimizes costs and optimizes server and storage capabilities.

Flat: The network comprises an access layer and a convergence layer. Its GE/10GE interfaces enable high-speed server access to guarantee bandwidth, improve network bearing efficiency, minimize network layers and fault points, and slash operation and maintenance (O&M) costs.

Virtualized: Loop-Free Reliable (LFR) technology virtualizes the complex network architecture into a few devices. Doing so greatly simplifies network topology, configuration, and management, eliminates network loops, boosts network reliability and maximizes O&M efficiency.

Layer 2: Huawei has evolved its DC architecture from L2+L3 into a streamlined L2 structure. This supports the in-service migration of virtual machines, enhances application reliability, and facilitates LAN and SAN convergence. Fiber Channel over Ethernet (FCoE) technology further simplifies network architecture by

M

A case study of Huawei’s enterprise network

Huawei Communicate

DEC 2010 . ISSUE 58

management, intelligent alarms, and automated alarm installation and power-off alarm unify remote management through a holistic solution for globally deployed equipment.

Engineers at the Shenzhen head office can manage equipment remotely via software to locate faults in regional networks, and guide local staff through the problem-solving process. Statistically, more than 85% of network faults can be resolved by the remote management solution; 10% involve replacing network equipment, and less than 5% require an on-site engineer. The benefits afforded by this solution are enormous in terms of time, O&M costs and efficiency.

Reaching every customer

Companies can only extend services to covered areas. Huawei’s ubiquitous, secure, and efficient enterprise network embodies multiple systems, including ERP, PDM, CRM, order management, Notes, UC, voice, and telepresence. These are coherently integrated to raise capability across the board, including in R&D, manufacturing, delivery, and service provision. Moreover, Huawei’s enterprise network solidifies the path of sustainable development, dulls the weapons of its competitors, and fosters customer satisfaction and thus loyalty.

Huawei is ready to provide its ubiquitous network to companies with a global strategy, spanning and integrating the access, bearer, and service layers. The ubiquitous network not only supports a full range of access methods, but its hierarchical and modular security system engenders a lean security solution. Its bearing technologies include a core switch that provides the largest capacity in the industry, while network bearing quality is guaranteed by 200 milliseconds protective switching, iVse video service optimization, and collaborative multicast control. The virtualized DC incorporates a router cluster with the largest available switching capacity, which combines with the LFR design to integrate services and ensure secure data access and transfer.

With the further integration of world economy, globalization is an inevitable step for many enterprises. Huawei anticipates strong partnerships with businesses aiming for that. And we hope to share over 20 years of experience in enterprise network research, construction, and maintenance to g ive f l ight to enterpr i se’s globalization aims.

14

realizing network-wide L2 switching.In 2009, IT and storage utilization surged in Huawei by

more than 50% and 60% respectively. However, network performance was stronger and more stable than ever before, with the advanced DC network easily supporting Huawei’s rapid commercial growth. In fact, the network shouldered a 30% increase in operations without requiring a single additional server or storage device. No serious faults were recorded over the whole year, and the total down time in 2009 was less than 5 minutes.

Ubiquitous global WAN

To support its global operations, Huawei’s worldwide enterprise network serves more than 90,000 employees and 200,000 terminals across more than 100 branches, 22 regional divisions, 17 R&D centers, and 36 training centers.

Naturally, people wonder how Huawei realizes such a fluid global connection using its central DC system in Shenzhen, especially in undeveloped regions that lack a robust infrastructure. By precisely matching access technologies with local network resources, Huawei’s WAN interweaves a full range of dedicated line resources, including MSTP, SDH, ATM, MPLS VPN, IPSec, SSL VPN, satellite links, 3G, E1, DDN, and FR. This allows each of its branches to freely select operation sites based on its business needs, and not on local resource limitations.

To elevate WAN QoE, Huawei began upgrading its international dedicated lines to the ones with ensured QoS in 2010 to guarantee voice and videoconferencing services. To enhance the QoS of dedicated line providers and accelerate fault repair in lines, Huawei shifted from leasing dedicated lines to leasing services from dedicated line providers. This model improves and expands the services offered by line providers, increasing efficiency for Huawei.

Remote management: Fast and cost effective

Featuring more than 15,000 switches, 500 routers, and 200 leased lines, the complexity of Huawei’s global network renders it impossible to visualize. However, its remote management solution frames network-wide O&M within a fast, tight, and inexpensive structure that ensures service bearing and ubiquitous access.

Drawing on industry best practices, Huawei has evolved its own advanced network management architecture by incorporating end-to-end network management solutions for SLA monitoring, service performance, network visualization, and network O&M. Visualized network

Editor: Li Xuefeng [email protected]


Industry Focus

DEC 2010 . ISSUE 58

The smart grid is coming

ccording to Wikipedia, a smart grid delivers electricity from suppliers to consumers using digital technology with

two-way communication to control appliances at consumers’ homes, in order to save energy, reduce cost and

Smart communications, smarter grid

increase reliability and transparency.Countries around the world have

different requirements when it comes to smart grids. Some developed countries have heavily invested in their smart grid projects to secure national energy supply, protect the environment, and boost the economy.

For example, the U.S. government has proposed a Unified National Smart

Grid that links all the nation’s local electrical networks to solve the problem of power outages. Europe initiated its SuperSmart Grid (SSG) plan, which connects Europe with northern Africa, the Middle East, and the CIS. The system will integrate wind and solar power, and smart grid capabilities into a comprehensive network. Japan will build its smart grid by focusing on

With the smart grid emerging worldwide, a smart communications system is needed to ensure its reliability and efficiency. The State Grid Corporation of China (SGCC), which has set itself the target of establishing a unified strong smart grid by 2020, offers some insights into the building of such a system.


A

By Bi Jianzhong & Zhang Qiong

15

DEC 2010 . ISSUE 58

Huawei Communicate

16

renewable energy like solar power. China’s smart grid will highlight ultra

high voltage (UHV), ultra-long-haul transmission, clean energy, automated power distribution, and bidirectional interaction with consumers. To be economical, efficient, and reliable, the smart grid requires an advanced communicat ions sys tem for rea l -time monitoring and dispatching, for example, the wind power generated in Inner Mongolia, the solar power generated in Tibet, and electricity transmitted from West to East China.

A communications network for the grid generally consists of a power distr ibution data network and an integrated data network. Future power communications networks wil l be expanded based on these two networks,

forming a strong and comprehensive infrastructure for the smart grid.

Ubiquitous interconnection

An automated power system mainly covers automatic power generation, t r a n s m i s s i o n , t r a n s f o r m a t i o n , distribution, and usage. All these fields should also be covered by the power line communications system.

Power generation: A smart grid would require that all large electric generators within a country are inter-connected, and thus a l a rge data network is required to ensure reliable power distribution. Using automatic

generating control (AGC), the inter-connected generators can generate electricity based the demand, lowering costs, while ensuring stable performance of the grid.

The smart grid can also help to integrate hydroelectric, thermal, nuclear, wind and solar power together, realizing a comprehensive and stable grid. Yet all these need to be backed up by a strong communications network for power distribution.

Power transmission: Transmission a n d t r a n s f o r m a t i o n o f p o w e r works closely with the power l ine communications system. The sensors collect running data from knife switches, transformers, and electrical lines (for example, current, voltage, power, frequency, switch status, and phase), and then send it back to the control center.

Based on the real-time data from the communications system, operational personnel can use the Supervisory Control and Data Acquisition (SCADA) s y s t em to r e a l i z e f unc t i on s l i k e telesignals, telemetry, remote control, and remote regulation. Connecting various kinds of power equipment, the communications system is of great importance to the grid.

Power distribution: Automatic power distribution requires a reliable communications system that integrates wireless and wireline services. This system should ensure that all automatic power dis tr ibut ion terminals can interact with one another and allow power distribution switches to respond properly, thus minimizing the impact of power outages in terms of area and duration.

Power usage: A ubiquitous wireless and wireline network would facilitate communica t ion be tween power -consuming devices and power-generating equipment, ensuring power is generated based on demand and used effectively, and thus enabling a stable and cost-effective power grid. Using WiMAX and passive optical network (PON) with various split ratios, Huawei can provide the optimum solution for power

Industry Focus

DEC 2010 . ISSUE 58

distribution and usage. The integrated data network for

power sys tem i s used to monitor product ion, weather, unattended transformer stations, office automation, a n d m o r e . B o a s t i n g a s e r v i c e management system, the integrated data network is a ubiquitous network that covers power generation plants, distribution centers, power supply units, service centers, and future battery charging stations and communities with smart power supply.

Wi th th e mu l t i p ro toco l l ab e l switching (MPLS) technology, Huawei provides virtual private network (VPN) solutions to power distribution centers and power supply units, helping them set up a highly reliable integrated data network with sound QoS. The solutions also help enrich service offerings, including video, voice and data services.

Strong data network

The smart grid should be strong and reliable enough to accommodate power usage that varies according to time and season. The load of a grid can fluctuate, for example, due to short circuits and the startup or shutdown of heavy-load equipment. Electricity frequency and voltage will change as a result, and the control center needs the information through the integrated data network for proper processing, so as to avoid power shortages and possible economic loss. Thus, a reliable data communications network is a must for a strong smart grid.

The backbone network of SGCC’s power distribution system has adopted Huawei’s NE series routers. Featuring high reliability, the routers separate the forwarding plane and the control plane, plus offer standby backup for key hardware components. The network uses a hierarchical architecture and has standby nodes for the national power distribution center and other key branches, while remote backup is also adopted for the key nodes. In a

mesh or half mesh networking model, the network has no single link or node, which ensures high reliability of the network services.

With bidirect ional forwarding detection (BFD), the routing convergence of the state gr id communicat ions network is reduced from the legacy 3-5 seconds to 50 milliseconds, realizing quick data transmission. In addition, nonstop routing (NSR) allows uninterrupted routing during equipment upgrades, which ensures network reliability and has no impact on the real-time application of the power system.

A strong power grid requires real-time power distribution, and this in turn requires that the data communications system perform well to prevent delay, jitter, and packet loss. The hierarchical QoS matches a queue to each power application to ensure that the real-time VPN service has a higher priority than the non-real time one. Fast fault detection also helps real-time operation. The Huawei NE routers support BFD fault detection, and can detect a fault in 10 milliseconds and report it to the network management system (NMS) for troubleshooting.

Smart grid, bright future

With the fast growth of Chinese economy, the power demand will keep increasing. In addition, the power grid will require higher scalability to meet new requirements, for example, transmitting power from Western to Eastern China, introducing ultra high voltage (UHV) lines and transformer s t a t i o n s , a n d i n t e g r a t i n g c l e a n energies such as wind and solar power. In response, the communicat ions equipment should have a large capacity and flexible configuration.

A s m a r t g r i d s h o u l d s u p p o r t bidirectional interaction. China boasts 400 million homes, a large number of SMEs and public facil it ies that consume power. To realize functions like intelligent meter reading, dynamic

electricity measurement, optimized resource allocation, and online electricity sales, it requires a smart data network with a massive number of IP addresses. In addition, a smart community would require automotive power distribution. Reaching people’s home appliances requires a large number of IP addresses, and IPv4 is falling behind in meeting such requirements.

Though they have a great potential for future development, electric cars bring new challenges to the power grids. Building battery charging stations and facilities will further extend the power-line communication network to urban and rural areas. As a result, a massive number of IP addresses are needed to support power distribution, metering, service centers, and asset management.

Both the power distribution data ne twork and the in tegra ted data network should be capable of supporting evolution from IPv4 to IPv6. As IPv4 and IPv6 will coexist for a long time, networks should ensure interoperability between the two entities.

Huawei has been a leader in IPv6 technology; its homemade IPv6 chipset supports both IPv4 and IPv6 protocols, plus large-scale and high-quality NAT64 and port network address translation (PNAT).

All these help ensure compatibility between IPv4 and IPv6 applications, and the future evolution of the power communication network.

SGCC has partnered with Huawei to build its power distribution data network and integrated data network, laying a sound foundation for future smart gr id. Besides real-t ime and efficient communications, the data ne twork has h igh f l ex ib i l i ty and scalability to support a strong power grid. SGCC is poised to build a cost-effective, secure, stable, clean and eco-friendly smart grid. Editor: Xu Peng [email protected]


17


DEC 2010 . ISSUE 58

Huawei Communicate

Network reliability is crucial

egacy enterprise Ethernet had fewer ports and services, and IT applications were largely independent of networks .

Enterprises simply required a basic network that was able to connect to ports and devices across a single node, and connect devices over a single link. This type of network was immune to the broadcast storm caused by network

Enterprise Ethernet

gets a sharper edge Highly reliable, efficient, and manageable Ethernet-based enterprise networks and data centers provide vital support to the growing number of business-based IT applications, including those relating to finance, administration, production, and websites.

By Pan Haotao & Zhang Fei

18

loops, yet its single node structure can hardly provide reliable backup when the network fails.

With technological progress and society’s increasing sophistication, enterprises increasingly rely on IT and information networks to conduct business, and in this area, inherited enterprise networks are proving to be inadequate. Today, multi-node, multi-link, and ring-based networking models have become the mainstream modes to enhance reliability; however, these networks come with loops, especially

L

Industry Focus

DEC 2010 . ISSUE 58

Fig. 1 LFR Ethernet networking structure Fig. 2 Stacked access switches

Aggregation

Access Switches Switch Stack

Access

ServersServers Servers

LACP

those that cover large campuses and data centers.

Ethe rne t ’s b roadca s t n e twork architecture is susceptible to broadcast storms. Statistics reveal that campus and data center networks experience an average of 10 to 20 problems, more than half of which seriously affect network performance. While spanning tree protocol (STP) and enhanced STP can eliminate broadcast storms, these protocols negatively affect network configuration and O&M, and are unsuitable for large networks. With loops, networks with STP are unreliable, unmanageable, and suffer low QoE.

Exis t ing network construct ion prioritizes the prevention of loops and broadcast storms to deliver a highly available, efficient, and manageable network. To realize such a network, Loop-Free Reliable (LFR) Ethernet embodies the optimum solution.

LFR Ethernet giving an edge

Underpinned by strong expertise in All-IP and Ethernet switches, Huawei has developed an LFR Ethernet solution that enhances network rel iabil ity,

shortens troubleshooting time, and increa se s bandwidth ut i l i za t ion . T h e s o l u t i o n i n t e g r a t e s s w i t c h clusters, stacking, and link-bundling technologies, and is as easy to deploy as a Layer 2 network. Compared with a network featuring redundant links, LFR greatly simplifies network O&M.

As Fig. 1 shows, the LFR Ethernet removes loops and the STP. The terminal server or PC connects with two access switches through two bundled links to enhance access reliability. Two (or more) access switches are stacked, and aggregation switches are arranged within a cluster switching system (CSS). Multiple access and aggregation links are then bundled to increase network bandwidth and reliability, and loops are eradicated between the access and aggregation layers. Switches are networked in the same way from the aggregation to core layer.

Fig . 1 i l lus t ra tes that the LFR network resembles a tree with core nodes at the root, and network traffic normally flows from the leaves to the root nodes. If an access switch fails, another in the stack automatically takes over and forwards all the traffic without affecting the aggregation devices. Equally, if an aggregation switch fails, another in

the cluster automatically responds to forward traffic without influencing the access and core devices. Core switch failure also invokes the same procedure.

The reliability, bandwidth utilization, and maintenance afforded by this design are superior to dual-link and dual-node. Maintaining an LFR Ethernet is as simple as maintaining a single-link, single-device network. Unlike a conventional Layer 3 routing network, the LFR Ethernet is free from complex routing protocols and IP addresses.

Given the crucial role of switch-clustering and stacking within this network structure, leading vendors generally apply stacking technology to box access switches, and this is now a common feature in Intranet applications.

To meet the growing demand for network bandwidth, the capability to cluster aggregation and core devices is essential. The Huawei LFR solution utilizes cluster technology to support a n o n - b l o c k i n g s w i t c h b e t we e n clustered frames, and the main control boards of the clustered devices are directly connected. This enables the non-blocking switch to function. It also eliminates the need to perform secondary switching, which is otherwise

Enterprise Ethernet gets a sharper edge

19

DEC 2010 . ISSUE 58

Huawei Communicate

switch are bundled into one entity, which simplifies the network structure, as Fig. 3 shows.

A network usually has much fewer aggregation and core nodes than access nodes. When planning to restructure networks, enterprises are usually flexible in terms of time and budgets. The process can be completed as an all-at-once project, or in phases, starting with access layer.

These three steps simplify legacy link protection switching by performing switching between two connected devices over a bundled link, which in turn eliminates the need for a complex protocol to control link monitoring and switching. Additionally, switching between devices in a single clustered node replaces conventional switching between redundant network nodes; this significantly enhances E2E service reliability. Moreover, the solution r educe s t h e numbe r o f phy s i c a l nodes by over 50% to yield a much simpler network structure and reduce management requirements. Editor: Li Xuefeng [email protected]

20


Fig. 3 Clustered aggregation/core

LACP

Aggregation

Aggregation

Cluster

Stack Stack

AccessAccess

required for a cluster with interface boards. This dramatically improves the reliability, efficiency and QoE of the cluster system, and curtails CAPEX given that slots are not occupied. Under the local traffic forwarding model, a device forwards traffic to a directly connected upper-layer device, which boosts efficiency without generating useless traffic.

Three steps to network restructuring

Enterprises can transform existing campus and data center networks into LFR networks by gradually reducing Layer 2 loops and constructing a large and robust LFR Layer 2 network.

Step 1 : P l an f e a s ib l e ne twork topology that avoids s ingle point failures.

On key nodes , two devices are deployed to realize redundant backup and thus prevent network and services from being affected by single device failures. Access switches are connected to servers, and redundancy is applied to aggregation switches, core switches, and network links. The network links should also be dual-homed to two devices. This

step eliminates single point failures, but results in a massive number of Layer 2 loops.

Step 2: Restructure the access layer to optimize bandwidth utilization in servers, simplify network structure, and eradicate loops with LFR.

Large numbers of access devices create a heavy workload when restructuring a network, but a gradual approach can minimize the impact on services. Generally, individual users connect to one switch, which requires no restructuring and allows the focus to be on the campus and data center servers.

Fig. 2 shows that the servers generally connect to two access switches through two adaptors, and work in active/standby mode. It is preferable to stack two adjacent access switches, and bundle the server access links into one virtual link. Two network adaptors can then share the server load to fully utilize bandwidth.

Step 3: Restructure the core and aggregation layers to simplify network structure.

Two adjacent aggregation switches c luster into one logic device that transforms triangular and box loops into a tree under the root node to eliminate the loop problem. Multiple upstream links for a single access or aggregation

Industry Focus

DEC 2010 . ISSUE 58

ith China’s economy g r o w i n g r a p i d l y , financial services have begun permeating every

aspect of society, making ubiquitous finance an inevitable trend. Ubiquitous finance, in turn, requires a ubiquitous network that connects bank service terminals such as outlets, branches, self-service banks, and POSs with data centers, featuring wide and high-density coverage without any blind spots.

Bottlenecks facing financial businesses

At present, each of China’s provinces has its own ubiquitous financial network to enable ubiquitous coverage. This is in the form of a tier-2 interconnected financial network that consists of three parts: a tier-2 backbone network, a tier-1 branch LAN, and an intra-city access network. It was designed to provide always-on connectivity using efficient, secure, and reliable high-bandwidth

This is particularly true for real-time services like deposits, videoconferencing, and voice services, which require high-quality bearing resources.

Third, as bank outlets are widely d i s t r i b u t e d , i t i s g e t t i n g m o re difficult to access outlets and manage the network. Outlets are the most important service channels of a bank, as they reflect its core competence. Stat is t ics show that even users of online banking services require services from traditional bank outlets at the same t ime. With rapid economic development, banks are setting up more and more outlets to offer increasingly complex services. Moreover, as outlets are widely scattered, access technologies vary widely from outlet to outlet. These increase the complexity of network management. To cope with the rapid development of outlets, the network must be capable of ubiquitous access and remote management.

Lastly, traditional private networks urgently require an upgrade. In order to connect branches, sub-branches, and

By Wu Fei

21

bearing capabilities, so as to ensure banks were able to provide quality services 24/7. However, the traditional tier-2 network has not been able to keep up with the surge in development of new banking services in recent years, and this is due to four main reasons:

First, the legacy network lacks the bandwidth and reliability that is required by businesses. Nowadays, new services l ike auditing system, videoconferencing, video surveillance, E-learning, and unified communication are being increasingly widely used. In particular, videoconferencing and video-surveillance services call for higher bandwidth and stronger network performance in areas like fault recovery, which the existing network is incapable of achieving.

Second, the legacy network lacks multi-service bearing capability to guarantee for key services. As bank services mushroom, multi-service bearing capability is a must for the network, as it guarantees delivery of key services over convergent IP network.

New services call for new financial networks

W

DEC 2010 . ISSUE 58

Huawei Communicate

22

business outlets to their data centers, banks used to lease a considerable number of ATM and SDH-based private lines from operators. As technology advances, telecom operators have moved away from massive private network construction, which means these private lines will soon be phased out. For this reason, many banks are beginning or planning to replace their private lines to sustain the expansion and development of their banking services in the long run.

Building a green, ubiquitous network

To help financial businesses address these challenges and build provincial financial networks with ubiquitous coverage, Huawei has unveiled a tier-2 financial interconnected network solut ion, leveraging i t s profound understanding of the telecom industry and deep insight into banking networks over the past two decades. Comprising a tier-2 backbone, a branch LAN, and intra-city access, the solution aims to provide banks with a provincial financial network platform that is always on, efficient, secure, reliable, and capable of providing superior financial services wherever needed.

A green pipeline: Tier-2 backbone solution

The tier-2 backbone network realizes interconnection between tier-2 and tier-1 bank branches. It leverages a range of technologies including IP, MPLS, SDH, MSTP, and WDM, to offer a ubiquitous high-bandwidth, secure, and reliable pipeline. It provides multi-service bearing capabilities with higher bearing efficiency using VPN and MPLS HQoS technologies; through visualized O&M network management and end-to-end energy-saving technologies, the solution also improves network management efficiency while consuming less energy. Wi th the ne w so lu t i on , hu rd l e s

Among the many businesses that are moving to upgrade their private lines, Agricultural Bank of China is the one that has taken the lead by upgrading its private lines to MSTP private lines, which are currently widely used in its outlets.

The bank ranks top among the four major state-owned banks in China in terms of the most extensive coverage of outlets and most branches. Due to expanding business and surging bandwidth demand from its outlets, the bank was under pressure to upgrade its legacy private line to an MSTP-based line that provided higher bandwidth and simplified interconnectivity.

Huawei’s high-end switches and routers, with a proven track record of reliable performance and stable operation during a trial run by the bank, were praised by ABC for their compatibility, switching, and Layer 3 routing. Successfully passing various tests by the bank’s head office and tier-1 branches, Huawei won the bank’s bid to assist with upgrading its tier-2 backbone and outlet access network.

Featuring large capacity and wire speed forwarding, high density, and strong scalability as well as powerful se r v ice capabi l i t i e s , the so lut ion fully met the bank’s core business requirements, and is helping to drive its rapid growth.

“In addition to offering high-quality products, Huawei is committed to providing customized quality services and solutions for enterprises,” says Li Weishi, head of Enterprise Network Product Line of Datacom at Huawei. Backed by more than two decades of experience in the telecom field, Huawei is striving with great efforts to help financial businesses serve clients wherever needed by offering them a green, competitive, and ubiquitous tier-2 customized platform.


plaguing the traditional tier-2 backbone network such as bandwidth shortage, unreliability, low multi-service bearing capacity, and challenges posed by the replacement and upgrading of private lines, are all cleared.

Convergent network: Branch LAN solution

A branch LAN offers convergence and acces s to user t e rmina l s and server clusters inside the branch office building. The solution provides an optimized combination of the following technologies: trunking, link aggregation, QoS, NAC, POE/POE+, f irewall , load balancing, Netstream, convergent management, and muting. Financial businesses can utilize this solution to build a quiet and eco-friendly LAN that integrates convergent structure, access, bearing, service provisioning, and management capabilities with a compact structure, enhanced bandwidth and reliability and reduced carbon footprint.

One-stop management: Intra-city access solution

Huawei’s intra-city access solution o f f e r s one - s top in t ra -c i ty out l e t and branch network construction, management , and ma in t enance . Equipping financial businesses with one-stop access, power supply, security, and remote network management capabilities, this solution helps improve access bandwidth and security for outlets significantly, and simplifies the network structure for outlet and branch networks. Its remote management feature can considerably improve the O&M efficiency of outlet networks and shorten service downtime caused by network failures. As such, the solution helps financial businesses rise to the challenges presented by the growing outlet base and service portfolio, making ubiquitous banking possible.

Making financial services ubiquitous


Industry Focus

DEC 2010 . ISSUE 58

Extranet takes China’s e-government nationwide

By Bi Jianzhong & Fang Su

A standardized, nationwide, and fully connected e-government extranet is providing China with a solid platform for intensive e-government applications and optimized administrative efficiency.

i n c e e - g ov e r n m e n t m a d e i t s d e b u t i n C h i n a s o m e twenty years ago, many local government departments have

deployed their own e-government infrastructure. However, the isolated nature of local authorities has largely prevented applications from being d ep loyed by more th an a s i ng l e depar tment or region. A unif ied, cohesive, and national extranet is therefore necessary to achieve the efficiency promised by e-government.

In December 2009, China’s State Information Center completed Phase I of the National e-government Extranet Project , which connects 400,000 terminals and the e-government systems of 70 central departments, and more than 10,000 local ones. The Extranet places all these units under the direct administrative control of the central government. Covering most parts of the country, this project represents the largest e-government Extranet in China both in terms of organization and geography.

In t e r n a l l y, t h e e - g ove r n m e n t Extranet forms a reliable and secure administrative platform. Externally, it bridges web portals and e-government application systems, and facilitates the role of the government as a unified public service provider. Four factors are crucial to constructing an effective e-government Extranet : secur i ty, interoperability, quality of transmission,

S and network management.

Controllable, visible, and credible

Ensuring the integrity and security of government data is of paramount importance, given that failure to do so could undermine national security. To achieve this, Huawei has adopted a three-tier concept for e-government Extranet solution: controllable, visible, and credible.

Controllable: User access to resources is limited by prior authorization, and lean traffic management is realized through NetStream and traffic cleansing.

Events that threaten information security are detected and prevented, and unauthorized terminals are isolated f rom the e -government Ext ranet through the Access Control List (ACL). Controllability also covers bandwidth management, traffic delivery, and data prioritization. These features combine to realize an efficient and fluid top-down and edge-to-edge traffic stream.

Visible: Traffic is comprehensively analyzed based on either application or user, and network-wide trends are clearly displayed through a graphical NMS. Complete system visibility allows administrators to detect and combat traff ic abnormalit ies and security protocol violations in real time.

Credible: Authentication control


23

DEC 2010 . ISSUE 58

Huawei Communicate

is unified to prevent hackers from intercepting, changing, or sending data using forged or stolen IDs. IP Security (IPsec), Secure Sockets Layer (SSL), and powerful encryption algorithms fully safeguard data during transfer.

Top-down and edge-to-edge

The e-government Extranet underpins the IT application systems of networked government units. The mature, flexible and scalable MPLS VPN features an elegant top-down and edge-to-edge organizational frame. China’s five administrative levels – national, provincial, prefecture, county, and town – are linked top down to achieve a coherent and transparent chain of command. In turn, the edge-to-edge structure laterally joins administrative units at the same level to unify the nationwide provision of one-stop civic management services.

Flexible: To enable prec i se and smooth data movements, the solution is designed to separate e-government applications of different levels, and connect those of the same level. Central government, local authorities, and State Council(China’s chief administrative authority) agencies rea l ize this s t ructure by applying the Huawei MPLS VPN over their intranets. Supporting inter-site point-to-point, full-mesh, and star networking, the Huawei solution can deploy vertical, horizontal, Layer 2, and Layer 3 VPNs.

Scalable: Previously, the fragmented deployment of the e-government project at various levels created a system of vertical efficiency and horizontal inefficiency that satisfactorily connected the central

government with local authorities top-down, but isolated local departments at the same level. The MPLS tunnels connect all government departments based on a scalable data-carrying platform with a multilayer label stack. With each layer of labels supporting up to one million connections, the MPLS fully integrates the e-government Extranet, creating solid connections both vertically and horizontally.

Mature: With nationally significant applications at stake, the e-government Extranet must employ a network technology that is proven in the field. Current MPLS VPN technology is mature, and this has been evidenced through several years of application in the mainstream networks of operators and enterprises. Underpinned by extensive experience in massive deployment scenarios and MPLS VPN management, Huawei can easi ly support the eff ic ient running of applications over the e-government Extranet.

Smooth transfer optimizes user experience

The e-government Extranet carries a range of services and applications for State Council agencies, including data transmission, videoconferencing, and video surveillance. However, these place high requirements on the bearer network; for example, end-to-end and point-to-point failback must be completed within 200 milliseconds and 50 milliseconds respectively. Breaching these times compromises the integrity of VoIP and interrupts

24

Industry Focus

DEC 2010 . ISSUE 58

video streaming, which impairs data service QoS and videoconferencing QoE.

Troubleshooting and failback

Huawei’s end-to-end troubleshooting and rapid failback mechanism for the e-government Extranet is widely applied to the backbone networks of various-level MANs to safeguard services and applications. The BFD protocol accelerates all routing protocols, Ethernet OAM, VPN fast rerouting (FRR), TE-FRR, VLL-FRR, and detection and failback. The mechanism applies these features to network faults occurring in any scenario and topology, and meets the 200-millisecond and 50-millisecond failback requirements for emergency communications, videoconferencing, IP calls, and key data services.

Video optimization

Videoconferencing has evolved from standard to high definition, though limited bandwidth forces ISPs to use high compression ratios for data and video streams. Losing even a small amount of streaming data on the bearer network during a videoconference can result in mosaic display, asynchronous voice and images, image stagnation, frozen frames, and black screens. To prevent this, Huawei’s network equipment intelligently identifies video traffic and stores key data I-frames in video streams encoded with MPEG-2 or H.264, retransmitting I-frames if any are lost accidentally. This ensures the integrity of the video stream, and significantly improves videoconferencing QoE.

Enhanced QoS

Different departments and IT applications have differing requirements for network delays, packet loss, and jitter. For example, an emergency communications system is of higher priority than other systems, video and voice services have priority over data services, and key data services are more important than ordinary data services.

Huawei’s Hierarchical QoS (H-QoS) technology easily solves transmission quality issues for various applications. Supporting 128,000 application streams, it independently assures bandwidth for each application, guarantees statistical division multiplexing (SDM) for multiple application streams, and ensures high-quality transfer across the emergency command system. For example, in response to the 2008 earthquake in Sichuan, the

State Council’s Office of Emergency Management could rapidly transfer data and video images to officials at the Emergency Command Office and to Premier Wen Jiabao in real time, thus providing crucial network support for disaster relief.

Fast deployment and easy management

The e-government Extranet incorporates multiple users and applications, thus ease-of-use (EOU) and simple maintenance are vitally important. Visible and automated management is necessary to facilitate rapid application deployment and fault detection, and unify the management of all network resources.

Visualization: Network planning is streamlined and efficient, which simplifies and accelerates configuring data; analyzing traffic; detecting anomalies; locating, isolating, and fixing faults; and outputting reports.

Automation: Automated network management includes an intelligent fault management system that analyzes massive amounts of fault alarms, identifies how they correlate, allows higher-level alarms to repress lower-level alarms, and automates configuration and upgrades for newly deployed equipment and terminals to realize plug-and-play functionality. Usability is improved and users can identify and rectify faults quickly and easily.

Unified management: The NMS can be deployed in modules to manage data, user access, and optical network equipment. Modules include the following systems: application management, traffic analysis, Internet behavior management, desktop security, and security authentication. These can be deployed as required to suit a specific scenario. Supporting third-party network interfaces, Huawei’s unified NMS can interwork with other mainstream management systems to unify the management mechanism.

The e-government Extranet enables local government networks to connect with the central government network and run smoothly. The solution creates a basic environment that integrates government resources at all levels, creates a fluid information-sharing platform, and coordinates applications.

25

Editor: Chen Yuhong [email protected]



DEC 2010 . ISSUE 58

Huawei Communicate

26

Synergizing IP and OTN transport networks

All technology solutions revolve around the network and its applications. By addressing evo lu t i on a r y n eeds , cOT N an d MC-LB technologies are spurring network convergence and creating comprehensive user value.

By Yan Qinghua & Bai Lu

Bearer Network

DEC 2010 . ISSUE 58

LB takes advantage of multiple-route transmission of data traffic to dramatically increase network reliability. In case there is a port or link fault, routing convergence will not be needed and a single-point fault can be isolated without triggering potential risks, significantly boosting network reliability.

Thanks to MC-LB technology, when reasonable network architecture is adopted, the effective transmission of data traffic on the upper IP network will not be affected by any fault in the bottom-layer optical network, and route shocking will not occur on the IP network either. MC-LB can tremendously increase network reliability without the need to deploy extra redundant resources, and achieve real IP/optical integration.

Higher networking efficiency

In a router network, the ideal networking method is one that can ensure direct connections between heavy-traffic nodes. This is because of the presence of many transit routers, which not only dramatically reduce the profitable bandwidth of the routers and the ROI, but also seriously deteriorate the QoS, making it hard to guarantee network performance.

It is very difficult for existing technologies to realize direct connections between nodes in a large or medium-sized IP network, mainly due to two limiting factors. One is the limited number of router ports; and the other is limited optical network resources.

The typical backbone router, for example, can ensure the mesh connection of 32 nodes at most (assuming that uplink and downlink ports are 50% respectively) at a speed of 10Gbps. With cOTN and MC-LB technologies, the number of mesh-connected nodes can be increased by up to eight times to reach 256 depending on the application without the need to add any port and fiber resource to the optical network. These two technologies can markedly improve network efficiency.

The networking efficiency of MC-LB is also reflected in balanced IP network traffic, effectively utilized network bandwidth and lowered CAPEX. Generally, network traffic distribution is highly imbalanced in different directions for geographical and other reasons. Traffic is heavy in some directions causing bandwidth congestion, whereas it is very small in other directions resulting in idle bandwidth. MC-LB solves this situation by balancing traffic and improving network efficiency. It uses the idle bandwidth to offload the traffic from the congested directions.

Increased ROI

Synergizing IP and OTN transport networks

27

ncreasing network convergence has sped up network architecture evolution, giving rise to cOTN (Channelized OTN) and MC-LB (Multi-channel Load Balance) technologies.

Based on IP+OTN network architecture, both can be used to creatively synergize IP and optical transport networks while raising network performance across the system.

Ideal convergence technologies

The colored interface in a router can only bring the WDM properties of the optical layer to the router interface, providing limited value for users. In comparison, cOTN and MC-LB technologies really integrate IP and optical networks through the collaboration of IP switching and optical cross-connection – a synergy that improves network reliability, efficiency, and ROI.

MC-LB solution allows data traffic in one direction to be transmitted from multiple physical ports and enables one physical port to transmit data traffic in multiple directions.

The OTN is a new-generation network standard for meeting the high IP bandwidth requirements. Combined with ODU-flex technology, cOTN channelizes the interface defined by OTN, effectively supports MC-LB, and works together with MC-LB to perfectly build a bridge between the IP network and the optical network.

Improved network reliability

In a large-scale IP network, risks such as routing black holes are not only very dangerous, but also can easily cause wide-area shocking across the entire network. Because these problems are hidden, troubleshooting is extremely difficult and the network can be affected for several hours or even a whole day. In 2007, this type of fault occurred on an operator’s network, lasting more than 10 hours and affecting a very large geographic area.

Such risks are inherent with the IP routing mechanism. The more complex a network, the harder it is for the network to avoid them. They are often triggered basically in the same way: A single-point fault (such as a port failure or link down) leads to node routing convergence, causing severe risks such as routing black holes.

In handling network traffic, the operator would be wise not to put all its eggs in one basket. MC-

I

DEC 2010 . ISSUE 58

Huawei Communicate

Network convergence from business perspective

Telecommunications have changed from a highly profitable industry to a highly competitive one with shrinking profitabil ity. The ROI in network operation is now a major concern and a top priority for operators.

Change for customer success

Rapid business growth has left many operators with excessive network layers. Due to poor network architecture, network expansion often results in an increased proportion of non-profitable bandwidth. Therefore, it has become a top priority for operators to optimize their networks by flattening the structure, reducing the number of transit ports, and increasing the proportion of profitable bandwidth.

The solution with cOTN and MC-LB is well-suited for network restructuring and can solve problems such as increased connection directions and the resulting traffic decrease in each direction. As the traffic analysis for an operator’s backbone network shows, compared with ordinary solutions, the solution can slash the number of router ports and transmission equipment by 30-50%, resulting in considerable CAPEX savings.

Many world-leading operators are engaged in lean operations that are focused on high-value services like VPN, while providing Internet and other services as well. What they have in common is that they run different applications on a single physical network. However, they still need to design different architectures for different services.

With their unique features, cOTN and MC-LB serve this need well by realizing the separation of the logic network from the physical network and carrying different services over different paths. This can dramatically lower network TCO while enhancing network maintainability.

T h e c o n n e c t i o n b e t we e n t w o routers in a network requires the same physical ports. In a real-world network, different network layers require different transmission rates for router ports. The nodes at the access and aggregation layers require many GE and 10GE ports, whereas the backbone core layer requires high-rate 40G POS/OTN ports.

The high-rate and low-rate cards of the same router are often mixed for connection requirements. This causes a series of problems: Configuring high-rate cards for access layer nodes is a waste of investment; configuring low-rate cards for backbone layer nodes reduces slot efficiency, and preparing all the high-rate and low-rate spare cards increases maintenance costs.

cOTN and MC-LB technologies realize the connection between routers with different physical ports. The technologies enable the access layer’s routers to direct ly connect to the 40G OTN por ts of the backbone layer’s routers through the GE ports, dramatically increasing router connection flexibility. Unifying the ports configured for the nodes of different network layers reduces the maintenance complexity and spare-parts cost to improve the ROI.

End-to-end OAM

A backbone network often spans several thousand kilometers, and equipment-level operation, administration, and maintenance (OAM) is far from being able to meet the needs of end-to-end applications.

As backbone IP and optical networks involve appl icat ions carr ied over long distances and large amounts of equipment of varied types, it is important to locate link faults and provide effective protect ion. The MC-LB solut ion extends, through cOTN technology, the advantages of the optical layer’s OAM to the IP devices. In this way, end-to-end application management and control can be easily achieved. Locating faults through the collaboration of multiple departments and troubleshooting them by sections is a thing of the past.

28


Interoperability and ease of deployment

In the FMC era, network convergence and service integration are improving interoperabi l i ty between network equipment from different vendors and require technological openness and interface standardization. In the application of a new technology, network compatibility and equipment interoperability must be addressed so as to protect network investment and guarantee a smooth technological transition.

By channelizing router ports, cOTN & MC-LB enable the interoperability between routers and the optical network, without requiring any new feature for optical equipments. Any optical equipment that complies with OTN standard is interoperable with the routers through cOTN & MC-LB. In this way, cOTN & MC-LB ensure the interoperability of optical NEs.

The flexibility in port configuration ensures the interoperability between new equipment and traditional equipment from other vendors. In connecting routers, MC-LB technology can best demonstrate its value. Furthermore, routers that incorporate this technology are fully interoperable with those that do not.

W i t h o u t d e p e n d i n g o n s u c h complex protocols as GMPLS, MC-LB achieves synergy between IP and optical equipment and creates future-proof networks. Once GMPLS matures, it can be deployed on these networks to improve their flexibility and operability.

An IP network is an intelligent network driven by routers. Any change in the routing protocol can have a big impact on the existing network and make deployment difficult. When implemented through physical and data link layers, MC-LB technology will not affect the routing protocol and makes for an easy deployment on the existing IP network without even changing the routing policy.

Editor: Xu Peng [email protected]

Bearer Network

DEC 2010 . ISSUE 58

New transport network requirements

he exp los ive growth o f ultra-broadband services has nur tured cont inual development of technologies

for the backbone bearer network. At the IP layer, cluster routers have enhanced the performance and switching capacity of backbone routers. At the optical layer, the optical transport network has an ultra large capacity and long distance transmission capabilities, plus intelligent functions.

Backbone networks are imposing new requirements on the transport network, including:

40G/100G high speed transmission: With increased traffic, the transport networks require 40G capacity per wavelength and should be compatible with 10G and future 100G. In 2009, leading vendors demonstrated their 100G transport and routing devices, but mass commercial deployment is not expected until later than 2012. 40G has become a must and 10G, 40G and 100G devices are expected to co-exist for a long time.

Tbit large-capacity cross-connection: The transport network is required to provide large-capacity cross connections to fully connect core nodes and establish direct routing between regional nodes. A Tbit OTN cross connection would help to bypass the traffic on IP layers.

Intelligent control plane: Key to a transport network, the ASON/GMPLS intelligent control plane can help realize end-to-end service provisioning through large-capacity cross connections, while enhancing the reliability of backbone networks. It can easily handle multi-point failures of transmission links. With

a unified control plane, the optical layer and the IP layer can realize interaction and synergy, creating a high performance, low cost backbone network.

Advantages of IP over OTN With the ASON/GMPLS control

p lane , OTN can in t roduce c ros s connect dispatching of wavelengths and sub-wavelengths to the point-to-point WDM system, enhancing the routing and switching capabilities and forming an inter-networked WDM system. OTN has the following advantages:

Better transport and higher efficiency

ITU-T has we l l -de f ined OTN standards for the mapping and bearing of Ethernet traffic. The latest standards have added ODU0 to bear GE traffic, ODU2e to 10GE LAN, and ODU4 to 100GE. The future ODUflex can map

T

IP and OTN synergy creates enhanced backbone networks

29

the traffic of different transmission rates into the OTN frame. All these have enhanced optical network capability in terms of high-speed, multi-service transmission.

The OTN enhances efficiency and lowers transmission costs. Currently, a 40G wavelength costs less than four 10G wavelengths. Using OTN to converge four 10G channels into one 40G channel is cost-effective. Also, through ODU0, ODU1 and ODU2, the OTN can map traffic over the GE level to ODU3 and ODU4, and then converge into a 40G or 100G wavelength, another enhancement that lowers transmission costs and boosts efficiency.

OTN can realize efficient traffic dispatching and end-to-end service provisioning. Based on a large-capacity cross connection and the ASON/GMPLS control plane, OTN can provide on-demand flexible bandwidth at any time.

OTN has the capability for full-service cross connection and traffic grooming, while offering non-blocking swi tching for granules inc luding O D U 0 , O D U 1 , O U D 2 , O D U 3 and wavelength, realizing the “any to any” wavelength and sub-wavelength connections in various optical fiber topologies. Compared to WDM with point-to-point networking, OTN has shown great progress with features like mesh networking.

By adding direct routes to edge nodes with a larger traffic load or with an IP offload solution, traffic can be diverted to the optical layer, reducing the pressure on core routers while strengthening the transmission efficiency of backbone networks.

More reliable transmission

The IP layer can sense services and efficiently process packet services; while the optical layer can provide abundant bandwidth to transparently transport massive amounts of traffic over thousands of kilometers. Integrating the IP and optical layers can definitely help operators build a high-performance backbone network at a lower cost.

By Liu Hongli

IP and OTN synergy creates enhanced backbone networks

DEC 2010 . ISSUE 58

Huawei Communicate

Optical networks should be able to reliably carry massive amounts of traffic. An intelligent optical network can be formed by introducing the ASON/GMPLS control plane to the OTN, which can also effectively solve multi-point failures on the backbone network. Statistics show that a WDM system extending 600km can reach 99.99% re l iabi l i ty, yet for an OTN mesh network equipped with the ASON/GMPLS control plane, reliability can be increased 10 times to 99.999%.

Wavelength-based service provisioning

Traffic on the backbone network mainly comes from enterprise customers instead of individuals. BT’s statistics show tha t ind iv idua l consumer s generate only 24% of network traffic, with the rest coming from enterprises and wholesale services, using one or multiple wavelengths. Traditionally, these wavelengths are provided by WDM sy s t ems and sw i t ch ing i s realized manually through patch cords on network nodes, making it hard to ensure reliability and deployment time. In comparison, OTN not only enables an operable optical network, but also wavelength-based private services.

Since 2007, mainstream vendors have introduced various OTN product series, with cross connection capability up to Tbit and most are equipped with an ASON/GMPLS intelligent control plane. The ASON-based ONT has been widely applied worldwide, and backbone networks are transforming from IP over WDM to IP over OTN.

Backbone network with IP/OTN synergy

IP over OTN is key to backbone networks. Currently IP over OTN generally uses a static optimization approach, requiring manual operations for network planning and configuration. A better approach is to use real-time

60 seconds in some cases.For a backbone network with IP

over OTN, the optical layer focus is on protecting physical layer devices, while IP layers are mainly protecting network nodes, ports and logical links. The co-protection of optical and IP layers enhance network reliability and efficiency. Thanks to the large number of direct routing paths on the optical layer, an IP network can reduce the number of hops to only one in some cases, simplifying the QoS mechanism and minimizing the impact of network traffic.

Network management synergy

Proper network management synergy realizes unified alarms, troubleshooting, and one-stop E2E service configuration.

Alarms from optical layers can trigger more alarms on the router layer. The synergy management of IP and OTN layers can help to locate and troubleshoot multi-layer faults, realizing unified alarm processing. Through correlation analysis, alarms from optical layers are suppressed and irrelevant alarms are f i l tered. Engineers can then focus on the root-cause alarms for fast troubleshooting and lessening O&M pressure.

With a unified network management system, the PCE and GMPLS UNI, IP and OTN can surpass the separate management and configuration model, realizing E2E service provisioning.

With IP and OTN synergy, backbone networks can remove traffic bottlenecks with unified traffic management, plus the co-protection function can enhance network reliability and performance. The network management synergy also assists in troubleshooting and alarm filtering, which enables E2E service provisioning. Practices show that for a typical backbone network, IP and OTN integration can save over 40% of CAPEX, OPEX and equipment room space. Editor: Liu Zhonglin [email protected]

optimization, which requires synergy between IP and the OTN. Huawei has launched its SingleBackbone solution, realizing synergy for traffic distribution, co-protection, and network management for both the IP and OTN layers.

Traffic distribution synergy

IP/OTN synergy for traffic distribution can help to enhance network performance and lower expansion pressure. If the traffic between any two routers exceeds the set value, the router can request bandwidth through the UNI interface. Receiving the request from IP layer, transport network will establish a direct optical routing between these two routers after routing and wavelength assignment (RWA) calculation. As a result, service providers do not need to expand the capacity as traffic is diverted through the OTN.

An IP interface for routers is four to five times more expensive than an OTN interface. As the optical layer has helped distribute traffic for routers, the number of hops and IP interfaces are reduced and so is the CAPEX for service providers. One leading operator reduced 40% of its CAPEX with the IP/OTN synergy solution, simply by bypassing the traffic from routers to the optical layers.

Co-protection for services

Bearing the services for tens of millions of users, a backbone network must be highly stable. In a Chinese operator’s backbone network, the number of line faults makes up 58% of all faults, and time required to attend to them makes up 76% of all time spent troubleshooting. The restoration technology employed by routers is mainly used for network node failures and a fiber failure will cause a large number of label switched path (LSP) failures, endangering the QoS.

Sprint’s research in the USA, shows that during the route convergence caused by line faults, IP traffic loops are generated, which cause delay, jitter and packet loss. Services are greatly affected, and the duration may last for 10 or even


30

Bearer Network

DEC 2010 . ISSUE 58

for IP and optical bearer networksIP and optical bearer networks may have its advantages, but these networks also require a different mode of protection. How does single-layer protection differ from synergetic protection, and which is the most appropriate protection solution? By Yan Qinghua

31

Synergetic protection

Synergetic protection for IP and optical bearer networks

DEC 2010 . ISSUE 58

Huawei Communicate

a logic network that carries data traffic.Unfortunately, the optical network is powerless

to solve a fault on the IP layer; if your car breaks down, for example, you need another car, not another road. However, if a fault occurs in the optical network and the road is blocked, you can take a detour and choose another route to your destination. The IP+OTN bearer network operates much like a network of roads, allowing IP packets to search for another route and take a detour if the way ahead – the physical layer – is blocked. Moreover, the work routes and protection routes must not overlap in order to fully protect traffic, and avoid one blockage from creating an impasse.

Multiple routes sharing a path form a Shared Risk Link Group (SRLG). If all routes fall under the same SRLG, network recovery is impossible if the shared path becomes faulty. In a single-layer network, either the IP or optical network can analyze its own network topology to simply and effectively avoid the pitfalls of an SRLG.

However, in a multi-layer network, separate network layers prevent synergetic analysis as the upper-layer cannot usually sense the topology of the lower-layer network. This makes SRLG risks likely, and seriously compromises the efficacy of a protection solution. Thus, a multi-layer network requires unified planning and synergetic protection to be reliable, efficient, and cost-effective.

Synergetic protection can be analyzed from three angles:

Assessing and avoiding network risks

A protection solution is limited by physical topological structure when an IP network cannot sense the topology of the optical network. If the physical topological information is sent to the router, the IP network can analyze the SRLG for the work route and protection route at the physical layer. A transmission network with high connectivity allows the IP network to search for the protection route to avoid a blocked SRLG. Conversely, low connectivity may mask the protection route, and in this case, risky links in IP network can be predicted.

Static SRLG requires the router and optical network to be manually configured by the network administrator, creating a heavy workload and a high risk of error. After updating the physical links, the topological information cannot be subsequently updated in time. Introducing a dynamic SRLG

Pros and cons of single-layer protection

P and optical networks are located on different layers, and employ different protection and recovery technologies. Accordingly, their inherent advantages and

disadvantages also differ.An optical network implements redundancy and

backup through physical media or physical channels to protect ports, nodes, and links. While ASON can help prioritize services based on an SLA and enhance channel efficiency in optical networks, channel occupancy is still exclusive. Extra costs are incurred as more physical channel resources are necessary to boost network reliability. Nevertheless, the physical layer responds quickly; switching is rapid; and, when coupled with adequate physical resources, the simple structure guarantees transmission reliability.

Unlike optical networks, resources are not exclusively occupied in IP networks, which are based on statistic multiplexing. Thus, multiple applications and users can share a single port or path. If a fault occurs, data traffic can choose alternative routes under a network-wide 1:N protection mechanism. In contrast, the 1:1 or 1:N protection offered by an optical network is limited to a single section. Thus, IP networks have a clear edge in terms of bandwidth utilization.

IP networks can also distinguish services and set appropriate QoS parameters. High value customers or services with high quality requirements can receive tier-1 protection for their data, while lower-level protection can be applied to services with less stringent requirements on network performance. This better utilizes network bandwidth based on demand, minimizes CAPEX, and maximizes QoE.

However, protection response and switching times are slow, as an IP network must calculate routes across the entire network. For example, the “butterfly effect” is easily triggered if a link becomes faulty in a network when bearing a heavy traffic load. One congested point leads to many others, compromising services network-wide.

Complementary advantages of synergetic protection

Different network layers have different functions. The optical network is the physical layer that serves as the bearer channel for the upper-layer network. Specifically, IP devices overlay the optical layer with

32

I

Bearer Network

DEC 2010 . ISSUE 58

enables the optical network equipment to provide the updated information of the router’s physical links through the UNI interface in real time. After recalculation, the router can intelligently analyze potential network risks.

The synergetic protection solution effectively eliminates threats to services by precisely judging the network’s overall reliability, and immediately responding to changes in, for example, network structure.

Simplifying route calculation for rapid service recovery

The IP network’s recovery speed is slower than that of a transmission network as the router must calculate numerous IP routes across the entire network. The calculation speed slows relative to network size and the number of links.

To solve this, the synergetic protection data derived from SRLG analysis enables the router to exclude the logic link routes of an SRLG on which a physical link is faulty. This simplifies route calculation and accelerates route convergence. In a scenario involving a huge number of routes and logic links on an SRLG, this solution is vital.

Combing SRLG route restriction and other IP protection technologies can accelerate an IP network’s recovery time. For example, a router with FRR can simultaneously analyze the SRLG’s restricted routes and calculate new ones if a port fault occurs, slashing the search time for route recovery.

Cutting costs and boosting protection

The objective of a network protection solution is to protect services. In an optical network, an effective solution provides reliable transmission channels; in an IP network it creates backup service routes. However, high reliability is inevitably accompanied by high costs.

Synergetic protection curtails expenditure by avoiding the repeated cross-layer protection measures that raise CAPEX. As the optical network already protects links, the upper-layer IP network does not need to do so, and can just focus on protecting devices. In turn, protecting services on an optical network can be prohibitively expensive due to the exclusiveness and large granularity of channels. In this case, applying IP network protection measures is more effective. Synergetic protection identifies the pros and cons of each mechanism, and exploits the advantages of the different network layers to optimize protection based on network structure and services.

Synergetic protection also raises protection efficiency by preventing repeat switching. Without the solution, a physical layer fault would activate the protection solutions of both layers, which not only delays service recovery, but also triggers oscillations in the IP network and risks the butterfly effect. Synergetic protection enables the IP network to detect the protection solution of the physical layer network, eliminate repeated switching,

Service Failure

UNI Link Problem NNI Link Problem

Optical Port Problem NNI Line Problem

No SRLG

Dynamic Protection

Optical Computing

Optical SwitchRoute Switch

Preplan Protection

Optical Switch

No Protection

Route Switch

No Switch

SRLGPort Backup No Port Backup

Optical Switch

No Link Backup

Route Switch

Link Backup

Link Protection

Fig. 1 Synergetic protection’s decision-making tree

Cooperation Mechanism

Synergetic protection for IP and optical bearer networks

33

DEC 2010 . ISSUE 58

Huawei Communicate

accelerate service recovery, and maintain the stability of the IP network.

Intelligent deployment keeps things simple

UNI and NNI interfaces connect the network devices on the two layers of an IP+OTN bearer network. UNI interfaces exist between the router and optical network, and NNI interfaces are deployed between optical networks. Synergetic protection safeguards the UNI links, NNI links, equipment por ts , board cards , and nodes by understanding the potential faults of each. While the blind spot of a single-layer network is removed by another network layer, the complexity of multi-layer, overlapped protection necessitates a single or composite solution based on the specific scenario and optimal ratio of cost and effectiveness.

The two layers share synergetic protection information and update network topology data to facilitate the solution. As shown in Fig.1, the decision-making tree quickly determines the most effective protection mechanism by assessing the best advantages to

implement - the optical network’s fast response and high bandwidth, or the IP network’s low cost and high efficiency. For faults that both layers can recover, the bottom layer’s protection solution is activated by default.

To determine the specific protective measures, network faults are divided into two types: UNI link and NNI link. For example, the solution implements FRR or VRRP on the IP layer if a UNI fault occurs, as in this case the optical layer protection obviously does not work. Equally, an NNI fault between optical network devices activates optical layer protection. For a fault on a port, the 1+1 or 1:1 port protection mechanism is implemented. Alternatively, if port protection is not deployed on a faulty port, NNI link protection can step in to search for a protection route through the control plane.

NNI link faults are subdivided into two types based on whether or not a fault occurs on an SRLG link. If it does, then backup devices cannot avoid the fault point, and successful switching is impossible. The only course of action in this scenario is to intelligently determine if the link is available to avoid pointless switching, and thus prevent the fault

from spreading across the network. I f th e NNI l ink f au l t h a s no t

occurred on the SRLG, the solution finds an available backup link. This process is also subdivided based on whether the optical network has a pre-set protection link. If it does, switching can occur directly through the optical layer’s link, and no intervention is required from the IP network. If not, dynamic route recovery on the optical layer is applied.

Usually, a suitable recovery route can be found via the dynamic route recovery undertaken by the optical network, though this is not guaranteed if other factors – such as insufficient redundant bandwidth – block the protection routes. If the pre-set time for the optical network’s protection measures expires, the IP network steps in. Thus, synergetic protection reduces waiting time and troubleshooting in the dark; the solution maximizes recovery speeds, optimizes efficiency, and minimizes the impact on services.Editor: Liu Zhonglin [email protected]


In a multi-layer network, separate network layers prevent synergetic analysis as the upper-layer cannot usually sense the topology of the lower-layer network. Thus, a multi-layer network requires unified planning and synergetic protection to be reliable, efficient, and cost-effective.

34

Bearer Network

DEC 2010 . ISSUE 58

100G stands tall with a strong backboneThe current hype and excitement about the 100G transport era has the telecom world agog with anticipation, and full of questions. When will it really come? How will the networks look? And how can we build networks to meet the high traffic requirements of this era?

By Shen Anle

When will it come?

ith the development of IP-based multimedia services, especially video services, network traffic has been continuously rising. As

a result, 40G and 100G have emerged as the key technologies capable of supporting the growth in network bandwidth. After a lengthy period of hype, 40G technologies have finally begun being deployed. According to Ovum, the compound annual growth rate (CAGR) of 40G WDM interfaces will exceed 70% from 2008 to 2014.

More than seven years elapsed between the launch of the first 40G prototype and its large-scale commercialization, and there were three major

35

reasons behind this long time to market. Firstly, the Internet bubble burst, which delayed the development of new technologies. Secondly, 40G technology is significantly more complex than 10G and can’t simply inherit its technological features. Finally, the lack of industry standards – especially for 40G and modules – has resulted in multiple 40G technologies and lack of economies of scale for 40G technologies.

Having learnt the hard way with 40G, the industry hasn’t repeated the same mistakes with the development of 100G standards, which has so far been moving smoothly. The standards will be completed by the three major standard organizations – IEEE, ITU-T, and OIF – in the second half of 2010, paving the

W

100G stands tall with a strong backbone

DEC 2010 . ISSUE 58

Huawei Communicate

way for large scale 100G application. In addition, 100G can inherit the features of 40G technologies, and some operators are even considering bypassing 40G and directly building 100G networks, given the network traffic spike and the rapid development of 100G technologies and standards.

However, a closer look at the maturity of 100G in terms of standards, technology and cost effectiveness shows us that the optimism needs to be reined in a little. The technological standards for 100GE were completed in the middle of 2010. However, the standards for 100G optical modules and 100G dense WDM (DWDM) require more time. Technologically, the entire 100G industry chain is still new; after the standards are finalized, the industry chain, which comprises chips, components, and systems, still needs time to stabilize and mature.

We have to be careful not to repeat the legacy mistakes that led to an immature 40G value chain. As the performance of 100G technologies currently compares unfavorably with 40G, the application scope of 100G technologies is limited. In addition, 100G will be expensive over the relative long term. These factors are combining to influence the schedule for 100G commercialization.

A l though the t ime to marke t fo r 100G technologies will be much shorter than for 40G, a long process of technological and market development is necessary before 100G goes commercial, meaning that it will coexist with 40G for some time. The industry generally views 2012 as the most likely time for 100G commercialization.

The basics of 100G

A complete E2E commercial 100G solution for backbone networks comprises two major technologies: the 100G router and 100G transport.

100G router technology

The 100GE line processing unit (LPU) and 100G forwarding engine are among the major technologies for routers to process signals at 100G.

A 100GE LPU must support service flexibility, high performance, and energy-saving capabilities through cutting-edge chip technology. Huawei has developed the Solar 2.0 PFE2A chip with proprietary macro instructions for packet processing (MIP) technology. Specifically designed for IP, MPLS, and ETH applications, it integrates the

36

advantage of a network processor to support flexible services and advantages of high performance, and low power consumption in an ASIC, thus overcoming the network processor’s inferior performance and the ASIC’s inability to take on new services. Underpinned by a unique energy-saving mechanism and an integrated peripheral design, the Huawei chip is the optimum tool for the 100GE LPU.

Based on the Solar chip and the IEEE 802.3ba standard, Huawei offers two types of 100GE LPUs for various application scenarios: the 1 x 100GE and the 10 x 10GE.

100G transport technology

The 100G transport technologies are mainly 100G code modulation, forward error correction (FEC), and 100G line transport technologies. Advanced code modulation technology is vital to realizing ultra long-haul and high-capacity WDM transport. Capitalizing on its years of technological accumulation and expert team, Huawei has developed multiple advanced code modulation technologies, such as sDQPSK, oPDM-DQPSK, and ePDM-QPSK.

sDQPSK technology uses polarization control to reduce the non-linear effects in a high-speed DWDM system, enabling the system to transport signals over a distance of at least 1,200km. By implementing advanced algorithms and hardware, the oPDM-DQPSK technology facilitates rapid optical polarization tracking and helps transport up to 80 wavelengths of signals at 100G. The innovative features of ePDM-QPSK technology include coherent detection, a high-speed analog digital converter (ADC), and a high-speed digital signal processor (DSP).

Based on advanced algorithms, DSP enables polarization tracking, the recovery of phase, clock, and data information, dispersion compensation and polar i za t ion mode d i sper s ion (PMD) compensation. With high tolerance to dispersion and PMD, ePDM-QPSK technology can transport up to 80 wavelengths of signals at 100G over 1,500km.

High-gain forward error correction (FEC) technology is another crucial aspect of ultra-long-haul transport. To eliminate the impairment of noise to optical signals, a 100G system requires higher-gain FEC than the existing transport systems. Huawei has independently developed a high gain and energy-saving FEC algorithm to

Bearer Network

DEC 2010 . ISSUE 58

eradicate the impairment to optical signals and ensure extra-long-haul transport.

To verify the feasibility of 100G transport technologies, Huawei has conducted a series of field signal transport experiments for 100G with mainstream operators.

In December 2009, Huawei successfully tested signal transmission in Spain – without electronics relay via 100G technology – over a haul of 1,000km on a live network. The realization of a hybrid 10G, 40G, and 100G transport mechanism proved that the system can support the huge capacity of 8Tbps over a single fiber after seamlessly upgrading from 10G to 40G/100G.

In April 2010, Huawei’s field network test in the UK confirmed hybrid 10G, 40G, and 100G transport capabilities, and successfully transmitted HDTV/IPTV services in real time.

Huawei recently achieved hybrid 10G, 40G, and 100G transport in Germany. Notably, with its advanced 100G technologies, Huawei successfully tested optical transport beyond the limit long haul of 2,000km on a live network.

These tests not only fully demonstrated that Huawei’s 100G transport technology was advanced and reliable, but also gave Hauwei valuable experience in large-scale commercial operation of 100G technology. Huawei will continue to conduct 100G tests with globally leading operators to continue to ensure stable, reliable, and advanced 100G technologies.

Challenges aheadBasic requirements for ultra large capacity

The cont inua l and rap id increase in IP traffic is pushing the bandwidth level between certain backbone routers to 100Gbps, or higher. Additionally, more 10G ports degrade 10G line bundling performance and common router efficiency. In the future, much will be expected of 100G ports.

Router switching capacity is also another pressing issue. Statistics reveal that router capacity increases 2.2 times every 18 months, slightly higher than Moore’s law, while network traffic doubles every 12 months. Obviously, network traffic outpaces router capacity, which necessitates cluster routers to serve as a partial solution. While cluster routers help mitigate the capacity issue, issues of

cost and power consumption still pose challenges to backbone network, and thus a superior long-term solution is demanded.

Given the continuous growth in network traffic and climbing costs of fiber deployment, optimally using each fiber to transport traffic is one of the core functions of transport networks. The solutions available include reducing the wavelength interval to increase wavelength numbers, applying L band, and enhancing the single-channel rate. Currently, passive optical components are designed based on 50GHz interval, which applies to many networks. With the increasing application of reconfigurable optical add-drop multiplexer (ROADM), an emerging trend involves the entire optical layer forming a network.

This makes the wavelength interval a key index to the performance of the entire network, for which reducing the wavelength interval is not the currently optimal solution. Furthermore, the L band solution is not cost effective due to the small industry scale of L band components. Therefore, 100G technologies with a 50GHz interval stand out as the optimum solution for transport network construction.

Flattening trend of network traffic

With the diversification of network applications and the maturation of the network facilities of the major Internet content providers (ICPs), network traffic is increasingly distributed and flattened. Typical examples are Internet data centers (IDC) and point-to-point (P2P) applications.

Conventionally, small IDCs are not often used and IDC capacity has to be periodically expanded, which wastes bandwidth resources in the metropolitan and backbone networks. Cloud computing has evolved IDCs from distributed data centers to cloud computing centers. Multiple physical IDCs can connect to form one virtual unified IDC to meet the increasing expansion requirements of a large ICP. IDC traffic within the same province can be integrated, which greatly increases interface utilization and further flattens network traffic.

To accommodate flattened network traffic, the network structure must be flattened at the router and transport network layers. A router network can change from a multi-level structure to a flattened structure by adding core nodes and removing convergence levels. Cluster routers are the key technology to facilitate this type of change.

37

100G stands tall with a strong backbone

DEC 2010 . ISSUE 58

Huawei Communicate


More core router nodes and less convergence levels require additional links between router ports and a mesh network to serve as the transport network. In turn, traditional DWDM equipment must evolve from a P2P system to an OTN or WDM system with networking capabilities.

Complex maintenance

The increase in network nodes and the flat network structure complicate the network at both the router and transport layers and make it hard to maintain and mange.

Firstly, the IP and transport layers seldom exchange information. Alarm information on the two layers is not correlated, so a lower layer alarm will generate many upper layer alarms, complicating fault location and maintenance.

Secondly, the independence of the two layers obviates an optimized network structure. In the case of protecting the IP layer alone, many redundant IP addresses are required, and still the effectiveness of doing so is not guaranteed as the IP layer is unaware of the information on the lower transport layer. In the scenario of two-layer protection, network efficiency falls, and the two layers become redundant.

Coordinating IP and transport

The challenges facing the backbone network can be partially solved by deploying 100G router interface technologies, cluster routers, the 100G transport network line, and networking WDM. However, these technologies do not fully meet the needs of the long-term network evolution of a 100G network.

In particular, the problems with router capacity, network efficiency on the IP and transport layers, and network complexity call for coordinated IP and transport layers. Currently, the industry has proposed various options, three examples of which are listed below.

OTN/cOTN

This solution employs channelized OTN (cOTN) technology at the router layer and interface technologies of OTN and VLAN/MPLS aware at the transport network layer. These technologies cooperate to provide an optimized IP transport network.

As part of the solution, the router layer continues to utilize existing router deployment technologies or employ cOTN technology to further enhance networking efficiency. The transport network layer incorporates the OTN and ROADM technologies to improve switching capability among either fixed 1G, 2.5G, 10G, 40G, and 100G bandwidths or variable bandwidths. The two layers form a unified and efficient bearer network through coordinated traffic flow management, hierarchical planning, and coordinated protection.

WDM interface + ROADM

This solution utilizes WDM interface and ROADM technology at the transport network layer. However, ROADM nodes cannot readjust sub-wavelengths and are consequently unable to transport data over multiple bandwidths in a backbone network. Moreover, WDM interfaces are not standardized, meaning that mainstream application is impossible. High speed 40G and 100G signals are more sensitive to physical impairment than the existing low-speed signals. In the future, large networks will still require relays and wavelength converters, and using routers for both operations is quite costly.

MPLS-TP

This solution introduces multi-protocol label switching (MPLS) as a simpler alternative to IP switching by the router. The MPLS transport profile (MPLS-TP) enhances OAM capabilities and utilizes ROADM at the optical layer. The MPLS mechanism improves network utilization, but can degrade QoS; balancing QoS and the cost of power is tricky. Moreover, MPLS-TP standardization is progressing relatively slowly, restricting the solution’s development.

In conclusion, IP and OTN coordination will dominate backbone network construction in the 100G era. Specifically, the OTN/cOTN will be the favored interaction mode of the IP and OTN layers. Huawei has fully explored each aspect of 100G’s core technologies and network construction. Huawei is ready to offer a full range of E2E solutions for backbone networks in the 100G era.

38


How to Operate

DEC 2010 . ISSUE 58

MSC Pool yields attractive ROIMSC Pool is increasingly popular with operators due to its numerous advantages, such as load balancing and enhanced network reliability. A closer look reveals it possesses a range of benefits that can contribute to an attractive ROI rate.

By Deng Weifeng & Wu Zhaojun

Advantages of MSC Pool ears of technological evolution and commercial application have shown MSC Pool to be a solution ensuring reliable voice services. MSC Pool networking

copes with surges in traffic, greatly reducing the threats caused by unbalanced load on NEs. It also implements disaster recovery at the VMSC layer to guarantee stable and seamless services.

Increase resource utilization

The resource-sharing mechanism of the MSC Pool balances loads among network elements (NEs) to mitigate threats from imbalanced traffic loads and surges that may arise based on time and

location. Traditional systems require each network office to configure equipment capacity based on the maximum traffic model, which is highly inefficient.

Conversely, the MSC Pool network sets i t s capacity in line with the total traffic volume to be

supported, allowing user t ra f f ic to

b e e v e n l y

distributed among MSC nodes. Redundancy capacity configuration costs are saved, and NE resource utilization is enhanced.

MSC Pool technology can also facilitate more extensive service coverage, while inter-MSC location updates and handover operations can be converted into intra-MSC operations. This cuts the number of handovers and signaling overheads incurred on the MSC and HLR, and saves transmission bandwidth and device resources. The subsequent freed up capacities can be allocated to process traffic to realize additional performance gains.

Improve system reliability

Mechanisms for sharing resources, balancing loads, and flexibly distributing traffic are the foundation for reliable disaster recovery. Automated load-balancing across the MSC Pool achieves real-time redundancy protection by switching service requests from a faulty MSC to another in the network pool.

The MSC Pool utilizes network resources more efficiently to curtail CAPEX and OPEX, and its innovative technologies, including geographical redundancy and immediate mobile terminating call recovery, boost network reliability tenfold to 99.9999%.

ROI analysisThe advanced MSC Pool network architecture

and optimized traffic models combine to minimize investment, which can be demonstrated through an ROI analysis. This analytical method evaluates key indices such as the investment yield, the cost recovery period, and quantitative benefits. For the MSC Pool, the assessment process primarily focuses on CAPEX, OPEX, and revenue, and then looks at

Y

39

MSC Pool yields attractive ROI

DEC 2010 . ISSUE 58

Huawei Communicate

i s deployed. For Operator A, this translated into a first year revenue increase of USD70,000.

Indirect benefits

Reduced churn and higher brand equity: The Huawei analysis revealed that the MSC Pool helps effectively avoid 12% of network faults. Enhanced reliability increases customer loyalty, the operator’s image, and indirect revenue gains.

Reduced maintenance costs: The greater efficiency of the MSC Pool’s unified O&M system reduced costs for Operator A by USD70,000 in the first year of deployment.

The Huawei MSC Pool solution has been commercially deployed in 40 scenarios across the globe. ROI analysis testifies that the Huawei MSC Pool is an effective solution for raising revenue; minimizing CAPEX, OPEX and O&M; reducing customer churn; and boosting network performance. Huawei has the experience, capabilities, and solutions to form long-term partnerships that maximize returns for operators.

The solution yields a range of direct and indirect benefits:

Direct benefits

Enhanced ne twork re l i ab i l i ty : The reliability rises from 99.999% to 99.9999%; the shorter service interruption times have in turn boosted returns. In the first year of deployment, the Huawei MSC Pool st imulated a USD20,000 revenue increase for Operator A.

More calls completed: The solution has reduced inter-MSC handovers and location updates within the MSC Pool, and has s ignif icantly raised network KPIs, such as call completion rate, which directly increases MOU. In its first year, the improvement in call completion rates yielded an extra USD2.76 million for Operator A.

Reduced investment: Load-sharing easily copes with peak traffic hours by efficiently utilizing network-wide resources, rather than just those of individual sites. This also means that the network can be expanded much more cost-efficiently. Message traffic within the pool has decreased, slashing the bandwidth needed and lowering transmission costs. Lab tests reveal that interface C/D bandwidth can be cut by over 20% after an MSC Pool

financial outputs, such as the break-even and pay-back periods, and EBITDA.

C A P E X a n d O P E X c a n b e obtained from financial statements and the management data released by maintenance departments. Quantitative revenue requires in-depth analysis as it relates to the solution features, service deployment, business models, user behaviors, macro economic conditions, and competition. Revenue analysis is the core of ROI analysis, and thus only an objective and precise analysis of revenues can guide proper investment decisions.

Case studyIn 2010, Huawei performed an ROI

analysis on the Huawei MSC Pool of Operator A based on six fiscal years of projected data. In its first year, the MSC Pool will: 1) increase revenues for Operator A by over USD2.76 million; 2) cut maintenance costs by USD70,000; 3) save USD70,000 in C/D interface bandwidth costs; 4) help to avoid 12% of network problems.

Fig. 1 shows the projected trend of Operator A’s MSC Pool up to 2015. Low CAPEX and OPEX coupled with increased revenues as a result of the solution will ensure that the investment costs are rapidly recouped.

Fig. 1 Investment return of Operator A’s MSC Pool

10.000

8.000

6.000

4.000

2.000

0.000

-2.000

-4.000

CAPEX Revenue OPEX

2010 2011 2012 2013 2014

FCF Cum. FCF


40


How to Operate

DEC 2010 . ISSUE 58

Taming the tidal wave of HSI traffic

41

Taming the tidal wave of HSI trafficThe Chinese ruler Yu the Great (c. 2200BCE) is remembered to this day for successfully taming diverting torrential waters to distant rivers, then to distant seas. Huawei draws inspiration from this strategy to help operators channel the tidal wave of traffic arising from today’s high speed Internet (HSI) services, and thus reduce the strain on networks, lower service costs, and raise QoE for subscribers.

By Wu Gang

Blocking cannot hold back the tide

2P is the major culprit when it comes to devouring bandwidth and lowering QoS guarantees for non-P2P traffic. Operators

have generally employed two methods to mitigate the upsurge in P2P traffic.

The first is to expand bandwidth capacity and unify bandwidth control. This was f i r s t adopted when P2P applications emerged, but operators soon found that the newly added capacity was quickly swamped by the explosion in traffic. They then had to resort to curbing P2P traffic by disconnecting ports, limiting connections, and sending across interference packets. Some US congressmen even proposed a bill to restrict P2P technologies.

The second method arrived with the maturity of DPI technology, as this could more accurately control P2P traffic, and thus Internet egress traffic. Operators could incorporate P2P bandwidth traffic into other service packages, or privately limit P2P traffic to preclude legal disputes and dampen the demand for Internet egress bandwidth.

P2P applications have extended their role from simply downloading large files to accommodating an array of multimedia applications, such as live P2P broadcasts and P2P voice services. Watching the news, movies, and live broadcasts via the Internet continues to gain in popularity. Limiting egress bandwidth in the face of this trend will inevitably degrade QoE, cause complaints, and possibly generate customer churn. Blocking traffic is not the answer; a new approach must be taken.

PChanneling is the new way

Internet traffic follows Moore’s Law. Although capacity of routers and switches has kept pace, the price of ISP egress bandwidth has not, and operators’ profits have suffered accordingly. The Huawei HSI service solution is designed to provide higher bandwidth at lower costs. Rather than curbing subscribers’ outgoing traffic, the solution channels traffic within the MAN, cutting the need for Internet egress bandwidth. This method not only raises QoE, but also lowers Internet access charges.

The solution comprises four high-performance components that can be separately or jointly deployed: (1) The ME60 (BRAS); (2) SIG9800 (DPI); (3) Internet Cache (Internet cache); and (4) RM9000 (policy server).

Cutting chargeable traffic

Internet cache technology is currently the most advanced technology for alleviating network bandwidth pressures, as it reduces the chargeable traffic between ISPs’ networks, which cuts networking tariff charges considerably.

By rearranging peers and caching P2P resources, the Huawei Internet cache redistributes the bulk of P2P traffic locally, and dynamically identifies and caches popular videos and files to ease the subscriber demand for inter-network traffic. In addition to diminishing Internet egress traffic, the Internet cache greatly enhances user experience by increasing the cached content within the MAN.

Internet cache adopts an offline architecture to increase networking flexibility without impacting network

DEC 2010 . ISSUE 58

Huawei Communicate

42

topology. With a loosely coupled server cluster, operators can scale up system capacity in line with the growth in network traffic without affecting existing services.

This solution has been deployed by several operators, including China Mobile, with positive results.

Higher bandwidthWith the Internet cache providing an intra-network

cache and distributing P2P traffic locally, MAN traffic has been skyrocketing. To deal with this, operators usually turn to high-performance BRAS to help cut the per-bit cost of MAN traffic.

The Huawei ME60 can support up to 256,000 subscribers, and provide 16 slots, each of which houses a 40GB BRAS board. Its highly integrated design helps reduce per-bit and per-subscriber costs. Meanwhile, the BRAS board can be integrated into the UPE, and thus increases P2P traffic efficiency as it operates closer to end users than a centralized BRAS board.

As the only device supporting hot HSI service backup, the Huawei ME60 allows a standby BRAS to take over services if the active BRAS fails, without disrupting services or requiring users to reconnect.

Optimized channeling of HSI traffic

The SIG9800 gives operators greater control over P2P traffic as it combines with the BRAS and Internet cache to channel HSI traffic into value-added services in multiple ways. These are as follows:

(1) Differentiated charging: Operators can deploy the Internet cache and other content servers within the MAN, and thus price MAN and Internet traffic differently. For example, as part of the strategy to draw subscribers to ISP networks, 8M bandwidth is allocated for ISP MAN network access, and only 1M is provided for Internet access.

(2) Time-based P2P traffic control. Operators can restrict P2P bandwidth during peak evening hours and lift this limit during off-peak daytime hours to fully utilize network resources and avoid congestion.

(3) A fair usage policy can be implemented to limit monthly subscriber traffic or P2P traffic. When traffic goes over the limit, access bandwidth can be lowered.

Global best practiceReducing maintenance costs with guaranteed user experience

Operator M offers innovative mobile services and fixed bandwidth services. Due to fierce competition, M decided to provide bandwidth of up to 20Mbps to boost QoE,

attract more subscribers, and increase competitiveness. During network construction, Operator M found

that increasing bandwidth up to 25M incurred a huge annual fee payable to its Internet backbone provider (IBP). A lower bandwidth of 20M required more BRASs, and thus more equipment room space and support facilities. Despite the increased bandwidth, the operator generated no extra profits.

Comparison and analysis led Operator M to Huawei’s full HSI service solution. The Huawei Internet cache continues to deliver higher bandwidth for subscribers and improve QoE at no added cost. Its powerful ME60 has driven down maintenance costs by halving the number of required BRAS. Finally, the rich package plans and services offered by the RM9000 and SIG9800 have boosted Operator M’s subscriber base, and stimulated further business growth.

Operator M was especially impressed that Huawei was able to fully understand its networking and business goals, and customize a complete series of HSI solutions that have produced optimal gains.

Sustaining subscriber growth and easing the network burden

Operator C’s profits suffered over the long term as a result of extremely expensive Internet egress traffic dominated by P2P applications, such as Thunder and PPStream. As subscriber growth outpaced available Internet egress bandwidth, QoE deteriorated sharply for users. As a result, Operator C planned to reroute P2P traffic locally to relieve the traffic burden on its Internet egress.

Before deploying Huawei’s Internet cache system, P2P applications consumed 19.35% of the Internet egress bandwidth. After the Thunder subsystem was subsumed into the Huawei Internet cache, and its traffic redirected locally within the WAN, the occupation of Internet egress bandwidth dropped sharply to 2.94%, which significantly raised QoE for Thunder users. The released bandwidth was then taken up by PPStream applications, which in turn drove up PPStream traffic and raised QoE for PPStream subscribers.

After implementing the PPStream subsystem in the Internet cache solution, PPStream traffic dropped from 40.08% to 21.82%, further improving QoE for PPStream subscribers and releasing bandwidth for other applications. This has substantially improved QoE, and is able to accommodate much greater subscriber numbers, easing the pressure of capacity expansion for operator C.



How to Operate

DEC 2010 . ISSUE 58

U2000: A fresh IP O&M experience

By Shen Hong & Suo Weiwei

The U2000 is a network management system (NMS) spearheading integrated management and visualized O&M in the industry. Through unified management of multiple technologies, the system eradicates bottlenecks such as complex IP network technologies and invisible service paths, significantly streamlining and simplifying IP network O&M.

large numbers, which make it necessary to swiftly identify the root cause and affected service, and provide a wide range of intelligent detection methods.

A large number of IP services, which requires operators to implement service-oriented performance management and provide accurate trend analyses so that O&M engineers can act proactively instead of waiting passively.

U2000, opening a new world of visualized O&M

Although an IP network supports a wide variety of O&M methods, only a service-oriented and visualized method serves the purpose of s impli fying network O&M.

The U2000 pioneers integrated management and visualized O&M in the industry. By conducting unified management of multiple technologies such as IP, MSTP, OTN, WDM, and PON, it enables visualized management of service-based features such as topology, faults, performance, and configuration. In doing so, the system eradicates bottlenecks such as complex IP network technologies and invisible service paths, significantly streamlining and simplifying IP network O&M.

Visualized service provisioning

IP service routes are notoriously intricate and changeable, complicating the IP network and its deployment.

Oriented to service provisioning, the U2000 offers a broad range of service templates that span access, WAN, and backbone technologies to support end-to-end service provisioning. As the system

T

43

U2000: A fresh IP O&M experience

h e t r e n d o f n e t w o r k convergence will inevitably alter the traditional way of performing O&M. As

domain-based network management falls behind this trend, integrated network management has become a compelling choice for efficient O&M.

In tandem with the transition to All-IP networking, world-leading operators such as AT&T, Verizon, Vodafone, Telefonica, BT, and China Mobile have planned to transition to All-IP bearer networks. The transition is accompanied by role adjustments for engineers plus changes in network technology, a rch i t ec ture , and equ ipment . IP technology complexity makes it much tougher job to manage an IP network than a non-IP network. Enabling non-IP engineers to perform IP network O&M is a t remendous cha l l enge operators currently face.

New challenges to network O&M

O&M main ly invo lve s s e r v i ce provi s ioning, faul t locat ion, and network performance management. Operators must address the following O&M challenges posed by an All-IP network:

All-IP network convergence, which requires the quick creation of end-to-end services, distribution of these services in batches, and a short time to market;

IP technology complexity, which necessitates delivering SDH-like O&M experience and rapidly bringing non-IP O&M engineers up to speed;

Difficulty in locating IP network faults and the presence of IP alarms in

DEC 2010 . ISSUE 58

Huawei Communicate

provides SDH-like operation, O&M engineer can easily start a parameter-based calculation by clicking the visible source and destination devices. It takes 15 seconds to provision one point-to-point service – making at least 85% more efficient than command line-based configuration. In addition, the U2000 system visualizes the bearer relations between channels, making the network clearly visible.

Accurate troubleshooting

The IP network generates a vast number of alarms. Statistics from an operator’s bearer network revealed 130,000 alarms in 130 days, the majority of which turned out to be “invalid”. For example, removing an optical fiber triggers a root alarm (link down) and a huge number of related alarms, including IP alarms at various layers such as OSPF down. Actually, related alarms (invalid) disappear automatically if the root alarm (valid) is cleared. In response, the U2000 provides a rich set of filtering mechanisms that focus on root alarms such as alarm relevance analysis, instantaneous alarm suppression, and filtering of engineering alarms. In typical scenarios, these mechanisms can screen over 85% of the invalid alarms.

In addit ion to NE-based alarm checking, the U2000 can identify the services affected by an NE alarm and then visually check a complete service on the network service topology to automatically and swiftly find the faulty section or layer. This allows users to ascertain the cause of the service fault s u c h a s i n c o r r e c t configurat ion, thus substantially increasing the ef f ic iency of fault location.

Diversified performance monitoring

Bes ides equipment a n d n e t w o rk - b a s e d

traditional performance index collection fo r ne twork s e r v i ce moni tor ing , the U2000 del ivers service-based performance management, helping operators accurately monitor service status across the network in real time through accurate service performance statistics and trend analysis. In the case of unusual network performance, the system enables O&M engineers to identify network running status and bottlenecks in advance for proactive O&M as it provides capabilities such as network-wide resource performance deterioration statistics and threshold alarms.

Also, the U2000 system is fully c o m p a t i b l e w i t h s e r v i c e - b a s e d northbound interfaces that provide capabilities such as service provisioning, p e r f o r m a n c e m a n a g e m e n t , a n d inventory configuration. Thus, the system can be readily integrated into the operator’s existing OSS.

Fresh global experience

Functioning as well as multiple t radi t ional network management systems combined, the U2000 system is an effective way of helping the operator reduce TCO. I t s v i s u a l i z e d O&M feature a l s o a l l ows

44

the operator to considerably improve IP network O&M efficiency. Moreover, it allows O&M engineers to easily perform IP service provisioning on the topology and view information such as service running status and traffic on the service topology to understand network services in real time.

So far the U2000 system has been deployed on networks of l eading operators such as Vodafone, Deutsche Te l e k o m , Fr a n c e Te l e c o m , B T, Telefonica, Swisscom, China Mobile, and China Unicom. Highly regarded for its pioneering features such as integrated management and visualized O&M, the system is definitely an ideal tool for addressing operators’ O&M needs arising from the transition to IP networking.

Editor: Pan Tao [email protected]


How to Operate

DEC 2010 . ISSUE 58

O&M and equipment imbalance

Lean O&M

he rapid improvements i n ne twork equ ipment integration and mature IP technologies have enabled

operators to reduce CAPEX and OPEX to some degree. Nevertheless, current IP technologies have generated a fresh crop of O&M challenges that prevent t rad i t iona l KPIs f rom accurate ly eva lua t ing ne twork qua l i t y. The deployment of a visible, controllable, and manageable IP network requires o p e r a t o r s t o r a i s e t h e i r O & M capabilities. The parallel pressure of the continuous decline in ARPU has forced operators to slim down their operations to ensure sustainable development.

Associated O&M

The range of solutions designed to

improve core network O&M usually focus on a single area, such as service history recording, signaling monitoring, and quality control for the IP bearer network. They are rarely associated with other solutions that have been designed to serve different suppliers’ equipment and other networks, such as the wireless or bearer networks. Consequently, legacy technologies have failed to provide a viable end-to-end O&M solution.

Supporting platformIntense competition necessitates

an open, high-performance support platform that provides a huge storage capacity, simplifies service deployments, and enab l e s p re c i s e con t ro l and management functionality. Such a platform should also allow operators to conduct target marketing by collecting and analyzing user location, behaviors, and preferences. The platform’s other functions need to include signaling monitoring, network cleansing, and the identification of malicious users and fraudulent activities on the network.

45

SmartCare: Your intelligent O&M partner

There is no doubting that core network equipment has advanced considerably in recent years. However, O&M has not kept pace, and has instead become a complex and expensive burden for operators. This imbalance has driven the industry to identify ways of boosting O&M efficiency, improving network quality, and implementing a precise network management mechanism. Only by doing so can operators hope to lower OPEX, guarantee customer satisfaction, and sustain revenue growth.

By Yu Bo

Your intelligent O&M partner

T

SmartCare

DEC 2010 . ISSUE 58

Huawei Communicate

46

Finally, controlling network traffic and QoS control via the platform can facilitate precise network planning and optimization, and enable accurate traffic pattern predictions.

SmartCare realizes precise O&M

SmartCare embodies a network O&M solution that introduces the concept of intelligent care. Unlike the previous O&M solutions that focus on optimizing network equipment performance, storage capacity, and technology, SmartCare instead realizes precise O&M throughout the entire network to constantly improve QoE, and, ultimately, raise brand equity and drive up revenue.

SmartCare logically consists of three layers: data collection, data sharing, and application. The data collection layer collects existing network data, including signaling information, equipment logs, service records, and network information. The data sharing layer stores this collected information, and provides upper-layer applications with interfaces to query and analyze data. In turn, the application layer processes service applications for target marketing, network planning and optimization,

and troubleshooting.

Associated O&M for the core and wireless networks

Today, operators can only separately obtain traffic statistics’ indices for equipment from the core and wireless networks, which fail to truly reflect network performance and user experience. For example, the call success rate (CSR) of the core network may indicate good network performance, but not reveal the actually poor performance at the wireless network side that results in degraded QoE and potential customer churn.

With SmartCare solution, equipment logs and service records can provide and automatically associate the end-to-end signaling data of the wireless and core networks. Operators can obtain associated end-to-end data about the signaling quality, CSR, call duration, and call drop rate to better understand user experience, and thus take maintenance measures accordingly.

Associated O&M for the core and IP bearer networks

The IP bearer and core networks are closely

Application Layer

TroubleshootingPublic SecurityNetwork Planning & Optimization

Precise Marketing

Statistics & Analysis

Basic Service

Unified Database Data Sharing Layer

Data CollectionLayer

Backdoor Operation Control

Voice Call Back Control

Nuisance Call Control

Spam Message Control

Network Monitoring

Network Planning & Optimization

Fault Location

Quality Control

BI (Management

Analysis)

SI (Subscriber

Behavior Analysis)

Data Mining

External Signaling Collection

Equipment Logs Service Records Bearer Network QoS Collection

Voice MOS Collection

Fig.1 Architecture of the SmartCare solution – a precise network O&M solution

How to Operate

DEC 2010 . ISSUE 58

related. The core network processes service signaling, while the IP bearer network transmits service data and media streams. Surveys reveal that over 80% of the complaints about core network performance are actually caused by IP bearer network faults. The flexibility of bearer networking complicates the assessment of the impact of IP bearer network faults on core network services. Equally, as the core and IP bearer networks are typically maintained by different departments, it usually takes several days – and sometimes even months – to resolve a core network problem caused by an IP bearer network fault.

SmartCare allows the data sharing layer to store the real-time data about the IP bearer network, such as network topology, equipment, and bandwidth data. This provides operators with visually represented information for easy fault monitoring and drives up the associated O&M efficiency of the core and IP bearer networks.

Associated O&M on different vendors’ equipment

Most networks contain equipment supplied by at least two vendors, the complex O&M of which often frustrate operators, because the vendor-provided O&M solutions only cover their own products and do not access network-wide data. While traditional signaling monitoring equipment can collect network data on a distributed basis, the resulting data can only be used to locate faults.

SmartCare enables operators to monitor signaling over IP, TDM, and ATM links. As all signaling data can be stored in a common database, operators are able to obtain signaling data for the entire network. Analysis of this data can realize certain functions that improve O&M across the entire network. These functions include outputting network-wide KPI statistics, locating faults, and synchronous fault tracing.

Obtaining and managing network-wide data

The application layer is the top layer of SmartCare and incorporates a wide range of applications like management analysis, subscriber behavior analysis, network management and optimization, public security, and troubleshooting. These applications require massive data support from the existing network. For example, for nuisance call barring, user activities across the entire network must be monitored

to identify malicious users and set restrictions against them. By vastly simplifying data mining, SmartCare makes it possible for operators to obtain this type of service information for all network users.

Becoming smarterEnd-to-end O&M on different vendors’ equipment

The SmartCare concept derives from the Circuit Switched (CS) domain. The solution serves not only Huawei core networks but also entire end-to-end networks (including wireless, core, and IP bearer networks) that use non-Huawei network equipment. SmartCare can also cover the packet switched (PS) domain, IP multimedia subsystem (IMS), cloud computing network, and content delivery network (CDN).

The sharing layer evolving toward the cloud computing platform

In China and India, the petabyte (PB) is used to measure data. With 106GB equaling 1 petabyte, the volume of modern data is a serious challenge for traditional data warehouses and relational databases. Given its unique distributed computing and expandability, cloud computing can almost certainly guide the future development of the sharing layer. Therefore, SmartCare must incorporate a cloud computing platform to enable massive data storage.

Building the application layer block by block

The SmartCare architecture comprises relatively simple and full-featured data collection and sharing layers that provide operators with a solid platform through which to implement these functions. Once the SmartCare solution is deployed, operators can access nearly all the network-wide data in real time and then determine how it would be best utilized. The data stored at the sharing layer can be used by operators and third parties to develop applications. Using the same platform and interfaces, these applications can be easily developed and integrated, yielding greater profits for operators.

Editor: Xue Hua [email protected]


SmartCare: Your intelligent O&M partner

4747

NOV 2010 . ISSUE 58

Huawei Communicate

4848

Leading Edge

NOV 2010 . ISSUE 58 49

HUAWEI COMMUNICATEof GPON and 10G GPON over the current OLT platform, passing all of the required tests for successful co-existence of GPON and 10G GPON in an established ODN. Yi Xiang,

Documents