Satellite Network Design · •Two different data rates are important when sizing a TDMA network… • IP Rate • IP Rate is the actual IP throughput including IP headers and data

1

Satellite Network Design

Robert Girard

Customer Solutions Engineering

2

Agenda • Which Network is Best for My Network?

• Service Level Discussion

• Satellite Network Technology Options

• Total Cost of Ownership (OPEX Vs CAPEX)

• Procurement & Logistics

• Space Segment, Satellite Equipment

• Registration with local authorities

• Logistics around equipment shipping and installation

• Satellite Economics – Beyond the Cost per MHz

• Procuring a Satellite Service

• Cost Structure of a Satellite Network

• Comparison of Different Satellite Networks

• Not All Satellites Perform Equally

3














4

Which Network is Best for My Network? Service Level Elements

• The most commonly discussed Service Level elements are those at the IP Layer:

• Committed Information Rate (CIR)

• Burstable Information Rate (BIR)

• Oversubscription Ratio (??)

• Quality of Service (??)

• The common Service Level Elements can be met only if the following are ensured:

• Latency

• Bit Error Rate (BER)

• Availability

• What is a “1Mbps Service”??

• 1Mbps dedicated outbound, 1Mpbs dedicated inbound??

• 1Mbps dedicated outbound, 256kbps dedicated inbound??

• 1Mbps shared (5:1) outbound with 256kbps CIR, 256kbps shared (5:1) inbound with 32kbps CIR??

5

Which Network is Best for My Network?

Service Level Elements

… sufficient margins must be included to ensure link integrity …

•Latency

•Bit Error Rate (BER)

•Availability

Optimizing The Satellite Link

6










7

Which Network is Best for My Network? Satellite Network Technology Options

• Satellite

• Frequency Band

• Hub Antenna Size

• Hub Location

• Remote Antenna Size

• Remote Locations

• Service Level

• Latency, Jitter, etc.

• Availability, Downtime, etc.

• Voice Traffic

• Number of VoIP Lines

• % Usage on Average

• % Usage Maximum

• Data Traffic

• CIR

• BIR

• Oversubscription Ratio

• Video Traffic

• Quality

Necessary Inputs to Determine Proper Solution

8


Satellite Network Technology Options

• Hub-based shared mechanism (Statistical Multiplexing)

• Timing references, burst guard band etc to remotes creates overhead • TDM / TDMA

• DVB, DVB-S2

• “IP Packet Switching over an MCPC Carrier”

• Single Channel per Carrier (SCPC) • Non-contended Capacity per site

• All “bursts” are traffic, one after another not overhead

• TDM/MF-SCPC

9


Satellite Network Technology Options

Advantage Disadvantage

Dedicated bandwidth for each

remote inbound

Each remote requires its own

space segment

Provides superior Quality of

Service for mission critical

applications

Expensive OPEX if each remote

bandwidth is not fully utilized

Low Latency and Low Jitter SCPC modems typically more

expensive than TDMA modems

Best transmission method for

real-time applications, voice,

data, video, broadcast, etc.

Fixed data rates on the inbound

links

TDM/MF-SCPC Model

10

Which Network is Best for My Network? Satellite Network Technology Options

Advantage Disadvantage

Sharing of satellite bandwidth High Latency and Increased

Jitter

Lower overall OPEX compared

to dedicated pipes

Demanding remotes can burden

the system

Good for low data rate

applications

Fragmentation of packets. Less

effective for voice and video

Low cost remotes Expensive hub equipment

Large population of users All remotes must be designed

around worst case link

TDM/MF-TDMA Model

11














12

Which Network is Best for My Network? Total Cost of Ownership (OPEX Vs CAPEX)

• Operating Expenses (OPEX)

• Satellite space segment

• Teleport operations

• Licensing

• Capital Expenses (CAPEX)

• Hardware

• Antenna, RF, HPA, Converters, Modem etc.

• Routers Switching equipment etc.

• Logistical Challenges

• Transport, Duties, Taxes etc

Operating

Expenses

Capital

Expenses

Network Operations + Depreciation

Total Cost of Ownership

Operations &

Maintenance

Transmission OPEX

Power

Spares/Support

Training

Site

Rental

Network

Equipment

Site

Equipment

Civil

Works

NRO

Transmission

Equipment

13


• Satellite Segment

• New Satellites bring better peformance

• Better Modulation and Forward Error Correction Techniques

• Teleport operations

• Remote login options

• VNO Options

• Licensing

• Liberalisation of Markets reduces the licensing fees

OPEX COSTS

14


• Two models of hub hardware cost assignment

• Per project

• Over multiple projects

• Hub platform costs can introduce barriers to entry

• Entry Level Hubs introduced to reduce barrier

• Virtual Network Operator (VNO) concept

• Growth costs must also be considered

• Function of how many end customers a “starter kit” can support

• Equipment re-use

• Some platforms use the same hardware for hub and remote

CAPEX (Hub Hardware)

15


Total Cost of Ownership (OPEX Vs CAPEX)

• Remote costs can be large part of the total cost of a network

• Portion of TCO grows with the size of the network

• Indoor Kit

• Low-cost TDMA/DVB-RCS Indoor Units (IDUs) have dropped in price to $1,000

• SCPC modems $6,000+

• Outdoor Kit

• Antenna and ODU sizing based on either shared carrier size or dedicated carrier size

• Logistical Challenges

• Shipping and installation in remote area

CAPEX (Remote Hardware)

16



• Two different data rates are important when sizing a TDMA network…

• IP Rate

• IP Rate is the actual IP throughput including IP headers and data at Layer 3 of the OSI model

• Represents actual LAN traffic on both remote and hub LANs

• Information Rate

• Information Rate is the actual Layer 2 information, including TDMA framing overhead, sent over the satellite

• Link budgets must account for Information Rate, not IP Rate

• Different TDMA platforms have different IP Rate / Information Rate ratios

• Depends on TDMA satellite access method

• aloha, slotted aloha, deterministic, selective, etc.

17



Data

Mbps

Layer 2 O/H

FEC

Modulation

• Data

• Actual Information • Layer 2 O/H

• Control

• Addressing *

• Guard Band *

• Reference Bursts *

• Modulation

• Number of Bits/Hz

• FEC

• Added Redundancy

Mbps vs. MHz

18


Data

Mbps

Layer 2 O/H

FEC

Modulation …Service REVENUE based upon

IP Rate in Mbps…

…Main OPEX cost based upon

total MHz required…

An end user’s service level requirements and

per-site price points will determine what

realistic margins may be achievable…

Mbps vs. MHz Revenue and Cost Bases

19


Spectral Efficiency vs. Eb/No

20


• Allocated BW

• Portion of transponder

actually used

• Linear function of modulation and FEC

• Decreases with higher order mods and FECs

• “Bandwidth Limited” links have greater allocated

BW than PEB

• Power Equivalent BW

• Fraction of transponder power required to close

link

• Complicated function of hub antenna, remote

antenna and satellite specifics along with

required Eb/No

• Increases with higher order mods and FECs

• “Power Limited” links have greater PEB than

Allocated

Bandwidth vs. Power

21


Relative Bandwidth (%) – for same data rate

-110 -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110

16QAM 7/8

16QAM 3/4

8PSK 5/6

8PSK 2/3

QPSK 7/8

QPSK 3/4

QPSK 1/2

QPSK 1/2 = 100%

Bandwidth vs. Power

22


• Advances in FEC can offer ≥3-5 dB of performance over currently used methods

• 3 dB of Coding Gain can:

• Reduce required bandwidth by 50% (OPEX)

• Increase data throughput by a factor of 2 (OPEX)

• Reduce antenna size by 30% (CAPEX)

• Reduce transmitter power by a factor of 2 (CAPEX)

• Provides more link margin (Service Level)

• What to look out for is

• Latency (Translates to Service Level)

• Eb/No Required (Translates to power; CAPEX)

• Bandwidth (Translates to allocated capacity on satellite; OPEX)

Forward Error Correction

23


• Turbo Product Coding (TPC)

• Iterative decoding process produces a likelihood and confidence level measure for each bit

• Low latency (vs. TCC, Vit/RS)

• Due to the fact that there is no need to buffer for interleaving

• Turbo Product Coding

• Lower Eb/No requires less power

• Higher efficiency requires less bandwidth

• Low Density Parity Check (LDPC)

• Basis of new DVB-S2 standard

• Third-class of Turbo Code

• Turbo Product Coding (TPC)

• Turbo Convolutional Coding (TCC)

• Iteratively decoded block code

• Performs 0.7 dB – 1.2 dB better than TPC at low FEC rates (3/4 and below)

• While coding gain is greater, processing delay can be an issue

Less Power

Less BW

Viterbi / RS TPC

24



Time

Mb

ps

/ M

Hz

QPSK

8PSK

DVB-S

DVB-S2

LDPC

CinC

16/32APSK

Relative Efficiency of Alternative Schemes

25

So Which Solution is Best??

TDM / MF - SCPC TDM / MF-TDMA

26

So Which Solution is Best?? • Satellite

• Frequency Band

• Hub Antenna Size

• Hub Location

• Remote Antenna Size

• Remote Locations

• Service Level

– Latency, Jitter

– Availability, Downtime

• Voice Traffic

– Number of VoIP Lines

– % Usage on Average

– % Usage Maximum

• Data Traffic

– CIR

– BIR

– Oversubscription Ratio

• Video Traffic

– Quality

… only with this information can one

determine the best option…

27














28

Procurement & Logistics


• Procure the correct beam Spot, Zone, Hemi or Zone beam for business requirement

• Procure Satellite Equipment according the linkbudget and business requirement


• Register the frequency and equipment with local authorities


• Find logistical solution to the deployment of the remote terminals

29














30

Procuring a Satellite Service: The Full Story

• Alpha and Beta are competing Mobile Network Operators.

• They both need satellite connectivity to support their operations.

• However, they have a very different procurement approach.

Alpha Beta

1. Selects the satellite operator which has the

lowest price per MHz.

2. Then, buys the cheapest equipment

available on the market.

3. Doesn’t fully take into account future

requirements in terms of equipment,

coverage and bandwidth.

1. Designs the satellite network taking into

consideration equipment and bandwidth

requirements.

2. Performs a Total Cost of Ownership analysis

including CAPEX and OPEX.

3. Considers future requirements to select an

effective invest-as-you-grow solution.

4. Trains in-house satellite experts to keep

optimizing the network

31


What could be the consequences of the two different approaches?

• An initial “cheap” design may prove to be more expensive over time:

• Company Alpha may need to invest more than Beta to expand

the network

• With limitations of ground equipment, bandwidth-saving techniques can hardly be

implemented.

• Possible lower performance

• Service availability, satellite performance, technical support, … matter.

Always compare apples with apples!

32

Procuring a Satellite Service: The Full Story • Example with Alpha and Beta’s respective services :

Alpha Beta

Information Rate 10 x 1 Mbps 10 x 1 Mbps

Efficiency 1.6 bps/Hz 2.8 bps/Hz

Required bandwidth 6.25 MHz 3.57 MHz

Price $/MHz 2,000 $/MHz* 2,500 $/MHz*

Bandwidth MRC 12,500 $/month 8,929 $/month

Equipment investment 100,000 $* 180,000 $*

Contract length 36 months 36 months

Total Cost of Ownership 550,000 $ 501,429 $

Equipment feature Entry-level modems,

Modcod 8PSK 3/4,

Roll-off factor 40%

Advanced modems with Carrier Cancellation

Technique and ACM,

Modcod QPSK 7/8, Roll-off 25%

Satellite performance C-band global beam low EIRP C-band zone beam high EIRP

* Indicative figures

33


• In addition to saving more money despite a higher initial investment:

• Beta’s service reaches virtually 100% availability thanks to ACM (Adaptive

Coding and Modulation)

• Beta will require less bandwidth to increase the data rates, compared to

Alpha → easier growth

• Beta is more competitive in the market since its cost per Mbps is about

30% lower than Alpha’s

• With the high performance satellite used by Beta, additional low-cost sites

can be easily deployed with smaller antennas and less power.

34


• Alpha and Beta are competing Mobile Network Operators.

• They both need satellite connectivity to support their operations.

• However, they have a very different procurement approach.

Alpha Beta

1. Selects the satellite operator which has the

lowest price per MHz.

2. Then, buys the cheapest equipment

available on the market.

3. Doesn’t fully take into account future

requirements in terms of equipment,

coverage and bandwidth.

1. Designs the satellite network taking into

consideration equipment and bandwidth

requirements.

2. Performs a Total Cost of Ownership analysis

including CAPEX and OPEX.

3. Considers future requirements to select an

effective invest-as-you-grow solution.

4. Trains in-house satellite experts to keep

optimizing the network

35


What could be the consequences of the two different approaches?

• An initial “cheap” design may prove to be more expensive over time:

• Company Alpha may need to invest more than Beta to expand

the network

• With limitations of ground equipment, bandwidth-saving techniques can hardly be

implemented.

• Possible lower performance

• Service availability, satellite performance, technical support, … matter.

Always compare apples with apples!

36

Procuring a Satellite Service: The Full Story • Example with Alpha and Beta’s respective services :

Alpha Beta

Information Rate 10 x 1 Mbps 10 x 1 Mbps

Efficiency 1.6 bps/Hz 2.8 bps/Hz

Required bandwidth 6.25 MHz 3.57 MHz

Price $/MHz 2,000 $/MHz* 2,500 $/MHz*

Bandwidth MRC 12,500 $/month 8,929 $/month

Equipment investment 100,000 $* 180,000 $*

Contract length 36 months 36 months

Total Cost of Ownership 550,000 $ 501,429 $

Equipment feature Entry-level modems,

Modcod 8PSK 3/4,

Roll-off factor 40%

Advanced modems with Carrier Cancellation

Technique and ACM,

Modcod QPSK 7/8, Roll-off 25%

Satellite performance C-band global beam low EIRP C-band zone beam high EIRP

* Indicative figures

37


• In addition to saving more money despite a higher initial investment:

• Beta’s service reaches virtually 100% availability thanks to ACM (Adaptive

Coding and Modulation)

• Beta will require less bandwidth to increase the data rates, compared to

Alpha → easier growth

• Beta is more competitive in the market since its cost per Mbps is about

30% lower than Alpha’s

• With the high performance satellite used by Beta, additional low-cost sites

can be easily deployed with smaller antennas and less power.

38














39

The Cost Structure of a Satellite Network

• As seen in the previous example, several parameters must be considered

• Equipment:

• Antenna size

• BUC size

• Modem capabilities

• Network topology:

• Star or Mesh

• Satellite bandwidth:

• Performance of the satellite

• Type of coverage: global, hemispheric,

zone or spot beam

• Dedicated or shared bandwidth

• Other: license fees, installation,

maintenance, international shipment, …

The performance requirements can

significantly impact the network design

thus the overall cost.

Define them sensibly and be ready for

tradeoffs!

The good design of a network will

consider the investment and recurring

costs required to meet the requirements,

while taking into consideration scalability.

40

The Cost Structure of a Satellite Network • A link budget report typically provides information that help assessing the cost of the solution

Carrier Information

(depends on modem capabilities)

Required

Bandwidth

Required

Transponder

Power (PEB) Designed Link Availability

Antenna size and required BUC power

Satellite and transponder

information

41














42

Dedicated SCPC Links • SCPC = Single Channel Per Carrier

• Typically used for point-to-point links with dedicated bandwidth

• Example of a network:

• Hub to Remote 1: 2 Mbps

• Hub to Remote 2: 2 Mbps

• Remotes to Hub: 1 Mbps each

• All sites have a 2.4m antenna and entry-level

modems.

For 99.96% availability

6.8 MHz are required

for the total network

Efficiency is only 0.88 bps/Hz

43

Improving Efficiency with Better Equipment • Same network as previously, but with different hardware configuration

• Now, if the requirements are modified a bit:

• Using Adaptive Coding and Modulation (ACM) and Carrier Cancellation Technique

• Maximum throughput 2 Mbps /1 Mbps (95% of the time)

• Minimum throughput 1.6 Mbps /0.94 Mbps

• Required bandwidth is 2.4 MHz and efficiency is 2.5 bps/Hz

Antenna sizes Modems Bandwidth Efficiency

Initial network 2.4m Entry-level 6.8 MHz 0.88 bps/Hz

Larger antennas 3.7m Entry-level 4.8 MHz 1.25 bps/Hz

Lower roll-off factor 3.7m Advanced 4.3 MHz 1.40 bps/Hz

Carrier Cancellation

Technique

3.7m Advanced with CCT 2.9 MHz 2.07 bps/Hz

44

Improving Efficiency with Better Equipment

• As seen in previous examples, using better equipment leads to

higher efficiencies thus lower recurring costs.

• However:

• Only a Total Cost of Ownership (TCO) analysis can determine whether the

investment on hardware is worth the bandwidth savings

• The size of the network, the required bandwidth and possible savings must

be taken into consideration.

45














46

Not All Satellites Perform Equally • The efficiency that can be achieved for a given satellite link also depends on the characteristics

of the satellite and transponder:

• Power density and G/T

• Beam coverage:

• The wider the beam, the larger the service area can be, but …

• Wider beams (especially global beams) typically have lower power density

• Note that some satellite operators only have global C-band beams for services in Africa

• Elevation angle: preferably above 20 degrees

• Available capacity

• Intelsat’s EpicNG satellites represent a major step forward:

• High throughput, increased power density, flexibility, vendor-agnostic, etc.

The cost/MHz will

typically vary in function

of these parameters

47

Not All Satellites Perform Equally

What about Ka-band? 1/2

• Ka-band is more susceptible to rain attenuation

> 18 dB

Legend:

Link Margin required for

99.6% availability Ka-band

Ku-band

48

Not All Satellites Perform Equally What about Ka-band? 2/2

• Most Ka-band spot beams are smaller than Ku-band beams: more beams required to cover an area

• In an attempt to cover the globe, some operators are stretching their beams, which reduces their power and efficiency

• A few myths on Ka-band:

• “Higher frequencies provide higher throughput”

Myth: There is nothing fundamental in a frequency band which supports higher throughput

• “Ka-band is more cost-effective because it allows us of smaller antennas”

Myth: Higher frequencies result in greater path loss between the antenna and the satellite, which nullifies

the increase in antenna performance. For similar link performance, larger Ka-band terminals are required

• “Attenuation mitigating techniques can compensate Ka-band rain fade”

Myth: There is a limit of how much rain fade ACM and UPC can address and these techniques are unlikely to be fully able to

compensate for Ka-band rain fade.

• “High Throughput Satellites (HTS) are Ka-band satellites”

Myth: An HTS satellite is one that uses significant frequency reuse techniques to multiply the effective throughput capacity of the

satellite. EpicNG HTS satellites use C-, Ku- and Ka-bands.

49

Sharing Resources for Better Cost Efficiency

• Dedicated SCPC links are well-suited to:

• Connections that need to be up all the time

• Traffic patterns that do not have dynamic variations

• Sharing resources:

• Satellite bandwidth

• Modulators and demodulators hub cards

• Solutions with shared resources are preferable when:

• Network is large with a central hub: avoid having multiple modems at hub

• Traffic demand is dynamic and varies within the network

50

Sharing Resources for Better Cost Efficiency • Typical configuration:

• Central hub with one or several modulator and demodulator cards

• Remote sites equipped with modems and possibly DVB receivers

• Shared outbound carrier (from hub to remotes), typically DVB-S2

• Dedicated or shared inbound carriers (from remotes to hub):

• TDMA, MF-TDMA, dSCPC, Mx-DMA

• Most equipment manufacturers have such solutions:

• Comtech Heights, Newtec Dialog, iDirect Flex, Gilat SkyEdge, etc.

• Intelsat offers solutions based on Newtec and iDirect hub equipment

51

Satellite Brings Benefits to Your Business

• Satellite technology should not be seen as last resort. It offers:

• Unmatched reliability

• Very high availability

• Short service implementation time in remote areas

• Wide coverage

• Easy point-to-multipoint communications

• Good value for money with state-of-the-art technology

• Opportunities for growth in untapped rural markets

• And so much more …

52

Conclusion

• Do not focus solely on unit prices for satellite capacity

• Always ask:

• What is the satellite performance?

• What equipment takes full benefit of the satellite capacity?

• What network topology do I need for the service?

• What efficiency can I reach?

• What is the Total Cost of Ownership?

• What support can I get from the satellite operator?

• What revenues and benefits can I drive from this service?

53

Thank you

Questions

Satellite Network Design · •Two different data rates are important when sizing a TDMA network… • IP Rate • IP Rate is the actual IP throughput including IP headers and data

Documents