Ethernet guideline

1/154 43-14/CSH 109 32/1 Uen Rev A

MINI-LINK TN ETSIEthernet Guideline

Description

.

Copyright

Ericsson AB 2009

Disclaimer

No part of this document may be reproduced in any form without the written permission of the copyright owner.

The contents of this document are subject to revision without notice due to continued progress in methodology, design and manufacturing. Ericsson shall have no liability for any error or damage of any kind resulting from the use of this document.

1/154 43-14/CSH 109 32/1 Uen ii

1 PREFACE................................................................................................................4

2 ACRONYMS AND GLOSSARY..............................................................................5

3 BACKGROUND INFORMATION............................................................................6

4 DESIGN GUIDELINES..........................................................................................10

5 REFERENCES......................................................................................................53

1/154 43-14/CSH 109 32/1 Uen 29.04.2009 iii

MINI-LINK TN ETSI

1 PrefaceThe main objective of this document is to provide design guidelines for an Ethernet transport network, built up with MINI-LINK TN and relevant 3rd party products. The document is focusing on a stepwise how to approach for the main workflows (e.g. configure L1 service). A more product oriented and architecture focus with detailed technical descriptions is provided in the MINI-LINK TN ETSI Ethernet packet Transport document.

The network types discussed in this document are addressing typical mobile backhaul (and broadband access) topologies.

The target audience for this document is internal Ericsson personnel that seek technical information and understanding on how to build an Ethernet transport network with MINI-LINK TN.

The MINI-LINK TN capabilities referenced in this document are equal to the 4.2 release.

1/154 43-14/CSH 109 32/1 Uen Rev A 29.04.2009 Ericsson AB 2009 4 (53)Ericsson Internal

MINI-LINK TN ETSI

2 Acronyms and GlossaryEthernet related abbreviations that are extensively used in this document are listed below.

BE Best Effort

CoS Class of Service

CPE Customer Premises Equipment

E-NNI External Network-Network Interface

EPL Ethernet Private Line

FE Fast Ethernet

GE Gigabit Ethernet

HSU High Speed Unit

IME Inverse Multiplexer Ethernet

Interface type An Ethernet interface in MINI-LINK TN can be of different types (e.g. UNI, I-NNI)

I-NNI Internal Network-Network Interface

LAG Link Aggregation Group

LLF Link Loss Forwarding

MAN Metropolitan Area Network

MEF Metro Ethernet Forum

NE Net Element

NNI Network-Network Interface

OAM Operations And Maintenance

PVID Default VID used untagged frames or un-trusted interface

RL Radio Link

RSTP Rapid Spanning Tree Protocol

SIU Site Integration Unit

TCP Transport Control Protocol

UDP User Datagram Protocol

UNI User-Network Interface

VID VLAN Identifier

VLAN Virtual LAN


MINI-LINK TN ETSI

3 Background informationThis chapter provides the background information (e.g. network topology) and other required assumptions to be used in the design guideline chapter

3.1 Network topology

A mobile backhaul can be built in many different ways based on individual customer legacy network, preferences, future plans and supported applications. Still, the figure below is a typical representation of a mobile backhaul network (RAN).

Figure 1: Mobile backhaul example

The Core Network is connected to the RAN through one or multiple BSC/RNC nodes. In the upper parts of the RAN, multiple ring structures are often used. Out in the edge of the RAN the base stations are typically connected in small depth star/chain formations from the rings.


Core Network

2G:BSC3G:RNC4G/LTE:AGW

2G:RBS3G:NodeB4G/LTE:eNodeBGW

MINI-LINK TN ETSI

A mobile RAN can be logically split in a HRAN and LRAN part. The HRAN consists of high capacity network elements high up in the RAN and connects to the BSC/RNC/AGW sites. The HRAN is typically built with optical rings. The LRAN is the lower parts of the RAN and connects to the base stations/CPE. A LRAN NE has typical lower capacity than a HRAN NE and uses a combination of microwave, fiber and copper transport technologies.

The grouping of the network elements in a HRAN and LRAN is not a formal and ambiguous split of the network. It is more of a way to organize and structure the RAN.

Figure 2: HRAN/LRAN

MINI-LINK TN supports a large variety of network interfaces and is well suited for the LRAN part of a mobile backhaul network. MINI-LINK TN is typically deployed in both the end and aggregation sites in the LRAN.


Core Network

HRAN

LRAN2G:RBS3G:NodeB4G/LTE:eNodeBGW

2G:BSC3G:RNC4G/LTE:AGW

MINI-LINK TN ETSI

The figure below illustrates a mobile backhaul network with the same HRAN/LRAN logical separation as previously discussed in this chapter. This example will be used for the design guideline discussions later on in this document.

Figure 3: Mobile backhaul network example

3.2 Site topologies

The site topology will differ based on the offered connectivity. Traditionally the sites have been application unique, meaning that a site has offered connectivity for one application only (e.g. GSM RBS). The figure below illustrates traditional 2G and 3G sites with typical circuit and packet drop capacity. A 2G site would require a SIU board to support packet transport (Abis over IP). All packets coming out of the SIU board are untagged. This means that MINI-LINK TN must tag all incoming packet when connected to a SIU board. An Ericsson NodeB 3G base station, on the other hand, delivers only tagged frames.

Figure 4: Sites with 2G and 3G equipment only


NPU3B

NodeB

4-6*E1

10Mbps

NPU3B

RBS

1-2*E1

5Mbps

3G site2G site

NodeB/ eNodeB

RBS

RBS

RBS

Optical ring

RNC/ BSC/ AGW

Packet Link

Core Network

NodeB/ eNodeB

NodeB/ eNodeB

MINI-LINK TN ETSI

Recently the site topologies have changed. Today, to save cost the transport network is shared between multiple applications (i.e. Fixed Mobile Convergence) and a site typically offer different network connectivity (i.e. GSM, Node B and eNodeB base stations). This implies a transport network and site topology that must be able to aggregate multiple Ethernet LAN ports as well as interfacing traditional PDH circuits. The figure below illustrates a combined 2G/3G/4G site. The traffic from the different radio base stations are aggregated in MINI-LINK TN.

Figure 5: Site configuration with aggregation in MINI-LINK TN

3.3 Dimensioning parameters

This chapter describes the scaling and dimensioning requirements in a mobile RAN backhaul network

Parameter Value

Max chain length 20

Ring size with RSTP 20

Max aggregated (Eth/TDM) southbound links in RNC site 16

Max aggregated (Eth/TDM) southbound links in HRAN hub site 16

Max aggregated (Eth/TDM) southbound links in LRAN hub site 8

Max base stations per RNC/BSC 200


100Mbps

eNodeB (LTE)

NPU3B

RBS

NodeB

4-6*E1

10Mbps

85Mbps

5Mbps

1-2*E1

5-8*E1

2G/3G/4G site

MINI-LINK TN ETSI

4 Design guidelines

4.1 Different design views

Designing an Ethernet transport network is a challenging task and consists of multiple small and large design decisions to be made. The design decisions can be grouped together and is in this document referred to as different views (i.e. functional areas).

The figure below illustrates the different design views that are described later on in this chapter.

Figure 6: Different design views

4.2 Port characteristics

4.2.1 Definitions

The Ethernet capable ports in MINI-LINK TN have different characteristics and capabilities depending on the actual configuration set up and usage. In MINI-LINK TN an Ethernet port is associated with an interface, service and port type. The combination of these settings uniquely defines the characteristics and capabilities for a port.

Port type

Both LAN and WAN port types can be used to transport Ethernet traffic. The main difference between these two port types is smaller buffers on a LAN port. As a design principle in MINI-LINK TN it is assumed that the WAN port is the congestion point, thus minimum LAN buffers.


Security view

Protection view Con.service view

CoS view

MINI-LINK TN ETSI

Service type

MINI-LINK TN supports L1 and L2 services. More details and a thorough definition are provided later in this document.

Interface type/port role

Each Ethernet capable interface in MINI-LINK TN is configured as a specific interface type. The interface type defines the supported feature set for that interface. The following interface types are supported:

UNI

• Represents the demarcation point between the operator’s domain and connected customer devices.

• Offers different security and control mechanisms to protect the operator’s network.

• Supports apply/strip of tags

• Supports priority classification based on both L3 and L2 priority information

• Supports congestion handling (not L1 LAN port)

• Used for both L1 and L2 Ethernet connection services

UNI-Light

• Used at the network edge between operator and users/customer domains.

• Do not offer the same demarcation feature set as UNI, but supports congestion handling.

• Use both L3 and L2 priority information for TC priority queue mapping

• Used for L1 Ethernet connection services only

I-NNI

• Used internally in the operator’s network domain

• Supports congestion handling (not L1 LAN port)

• Assumes pre-tagged frames, i.e. only L2 priority information used.


MINI-LINK TN ETSI


E-NNI1

• Used in the network interface between two network operators.

• Represents the demarcation point between the two operator’s network domains.

• Offers different security and control mechanisms to protect the different operator domains.

• Supports VID/priority remapping

• Supports congestion handling


The figure below illustrates a MINI-LINK TN with different Ethernet capable interfaces. The capabilities of each interface are defined by the interface service and port type definitions.

Figure 7: Ethernet port capabilities in MINI-LINK TN

1 Not supported in current release


L2/I-NNI/WAN

L1/UNI/LANL1/I-NNI/WAN

L2/I-NNI/WAN

L2/I-NNI/WAN

MINI-LINK TN ETSI

4.2.2 Configuring interface types

4.2.2.1 UNI interface type

The following configuration alternatives are supported for the UNI interface type

• Frame admittance

a. Allow un-tagged (DEFAULT SETTING)

i. Configure tag removal for outgoing traffic

ii. Set default VID (PVID) for incoming traffic (DEFAULT PVID=1)

iii. For trusted port set “L3 trusted” (i.e. DSCP/MPLS-EXP)

iv. For un-trusted port and non-IP frames specify default priority

b. Allow priority tagged

i. Set default VID (PVID) for incoming traffic (DEFAULT PVID=1)

ii. For trusted port set “L2 trusted” (port assumed trusted)

iii. For un-trusted port specify default priority

c. Allow tagged

i. For trusted port set “L2 trusted”

ii. For un-trusted port specify default priority and default VID (PVID) for incoming traffic

2. Additional security features

a. Storm Protection (DEFAULT 100 frames/sec)

b. Port Blocking (DEFAULT OFF)

c. White List (DEFAULT OFF)

d. MAC address limit per port (DEFAULT OFF)

3. Policing/color marking of ingress flow (DEFAULT OFF) (currently not supported)

4. Congestion handling (see chapter 4.6 for more info)


MINI-LINK TN ETSI

4.2.2.2 UNI-Light interface type

The UNI-Light interface type is a customer/user external interface but is significantly simpler to configure than an UNI port. An UNI-Light interface supports congestion handling (see chapter 4.6 for more info)

4.2.2.3 I-NNI interface type

The I-NNI interface type is an internal interface and supports congestion handling (see chapter 4.6 for more info)


MINI-LINK TN ETSI

4.3 Connection service view

A MINI-LINK TN can provide both L1 and L2 Ethernet services. A L1 service offers a dedicated point to point Ethernet connection between two termination points, while a L2 service performs aggregation of multiple Ethernet flows.

An operational network will typically contain both L1 and L2 Ethernet connection services. A L2 Ethernet connection service domain is deployed in the centre of the backhaul network as a network switching core. At the edge of the network multiple L1 services (i.e. port extensions chapter 4.3.1.3) connects remote edge devices (e.g. RBS) to the switching core. In addition individual L1 services can be used to create dedicated tunnels (chapter 4.3.1.3) through the switching core. The figure below illustrates a hypothetical mobile network.

Figure 8: L1 and L2 Ethernet connection services in a Mobile RAN

Both the Port Extension/Dedicated Tunnel L1 services and L2 service are more closely described below in this chapter.


Core Network

L1 Ethernet connection service (i.e. Port Extension)

Mobile RAN

L2 Ethernet connection service

L1 Ethernet connection service (i.e. Dedicated Tunnel)

MINI-LINK TN ETSI

4.3.1 L1 Ethernet service

4.3.1.1 General

The main purpose of the L1 Ethernet service offered by MINI-LINK TN is to provide a direct connection between two termination points. A termination point is either a physical LAN port as in the Dedicated Tunnel use case or a WAN port connected to the switch as in the UNI port extension use case. The LAN ports can be electrical, optical (only GE), FE or GE while the WAN ports can be Ethernet over PDH (EoPDH), Ethernet over SDH (EoSDH) or Packet Link.

On each side of the L1 Ethernet service, the appropriate interface type (e.g. I-NNI) is selected for the termination points2. The interface type defines the functionality level supported for the termination point. Each use case described in this chapter is associated with a set of interface types. The UNI port extension use case is for instance associated with an UNI-light interface type, while I-NNI is selected for the network internal I-NNI port extension use case. The UNI-light interface type supports only a subset (i.e. congestion handling) of the traditional UNI port feature set.

An L1 Ethernet service is symmetrical in the sense that both termination points offer similar behavior. Ethernet frames will receive the same forwarding behavior independently of the direction of the flow. The figure below illustrates the basic set up.

Figure 9: General description of L1 Ethernet service

An L1 Ethernet service can be used for numerous scenarios. However, from a link utilization point of view, it is best suited for connections with constant rate flows (e.g. file transfers). Since an L1 Ethernet service doesn’t support aggregation there are no potential statistical benefits from time and day variation in multiple links. Traffic flows on a L1 service with a heavy burst nature will either lead to packet drop or an under utilized connection.

2 Please note that interface types currently cannot be selected for an L1 service


Transport connection

L1 Ethernet service

Termination points

Interface type (e.g. I-NNI)

LAN/ WAN

LAN/ WAN

Interface type (e.g. I-NNI)

MINI-LINK TN ETSI

For flows with high demands regarding a predictable throughput, jitter and packet loss, an L1 service is good alternative.

An L1 Ethernet service in MINI-LINK TN offers the same behavior as an EPL service defined by MEF.

4.3.1.2 Characteristics

An L1 Ethernet service offers the following characteristics:

• An L1 Ethernet service offers a point to point connection between two termination points. This means no switching between the endpoints.

• An L1 Ethernet service can utilize an Ethernet over PDH/SDH or Packet radio link WAN connection.

• An L1 Ethernet service offers a predictable throughput between the termination point ports. The throughput is configurable (2-620Mbps).

• An L1 Ethernet service offers a predictable delay and delay variation between the termination points ports (i.e. no aggregation points).

• An L1 Ethernet service offers low packet loss between the termination points ports (i.e. no aggregation points).

• Unicast/multicast and broadcast service frames are forwarded transparently between the termination points.

• L2CP control frames are forwarded transparently between the termination points.

• Flow control pause frames are discarded

• An L1 Ethernet service does not offer any protection of the termination ports. On the WAN connection however between the termination ports, different protection mechanisms can be used, e.g. MMU radio protection. Protection can also be provided through the connected devices at the termination points.

• Link Loss Forwarding (LLF) is supported to enable fast protection switching in the connected devices at the termination points.

• As a design principle for an L1 Ethernet service, it is assumed that the WAN link is the congestion point. Therefore in the LAN direction, the connected interfaces to the L1 Ethernet service have to be over provisioned, i.e. higher capacity than the WAN link.


MINI-LINK TN ETSI

• The LAN port in a L1 Ethernet connection service is not a congestion point, therefore the LAN port doesn’t have a buffer.

• A WAN port shall be able to buffer 100ms of traffic per priority queue/traffic class.

• Since a LAN port has no buffer, only full duplex are supported, i.e. no half duplex.

• A L1 Ethernet service utilizing a Packet Link WAN connection is limited to one radio hop, while Ethernet over PDH/SDH WAN connections can utilize multiple radio hops in a chain.

4.3.1.3 Use cases

An L1 Ethernet service in MINI-LINK TN can be used as a Port Extension to transparently connect over a radio hop a non MINI-LINK TN device on a remote site to the operator’s L2 switching domain. In the figure below, a Port Extension is used to connect device “A” to an operator’s MINI-LINK TN L2 network.

Figure 10: Port Extension use case

A Port Extension connection provides a simple and cost effective way of extending the operator’s network domain to remote sites. From an operational point of view, a Port Extension can be regarded as a generic extension cord. The interface types on each side of the L1 service are selected based on the Port Extension use case variant. The following Port Extension use cases are supported:

• UNI port extension

A network connection that connects a customer/user equipment to an operator network domain.


”A”MINI-LINK TN

L2 Network


Port Extension

Termination points

LAN

WAN

MINI-LINK TN ETSI

• I-NNI port extension

A network connection that connects two separate network clusters within the same operator network domain.

• E-NNI port extension

A network connection that connects two separate operator domains.

The use cases are visualized in the figure below.

Figure 11: Port extension with a L1 Ethernet service

The WAN connection in a Port Extension is typically one radio hop. However, multiple hops can be linked together. For the Ethernet over PDH/SDH WAN alternative the connections are cross-connected in the backplane of the interim MINI-LINK TN nodes. A Packet Link WAN connection though requires external LAN cables to connect the multiple hops. This will in fact daisy chain multiple port extensions in to one connection, i.e. a Port Extension with Packet Link supports only one hop.

An L1 Ethernet service in MINI-LINK TN can also be used to provide a transparent end to end Dedicated Tunnel through the operator’s network. The figure below illustrates the use case.

Figure 12: Dedicated Tunnel between ports A and B with a L1 Ethernet service



Dedicated Tunnel

Termination points

LAN LAN Network domain

Network domain

”B””A”

Operator 1 Network domain


User/ customer domain

I-NNI port extensionA

UNI port extension CPE

E-NNI port extension

MINI-LINK TN ETSI

UNI Port Extension

As previously stated in this chapter, an UNI Port Extension can be used to connect a CPE device at a remote site to the operators L2 network domain. The figure below illustrates the scenario.

Figure 13: UNI Port Extension use case

At the remote site, a MINI-LINK TN is connected to the CPE device with a LAN cable. The LAN port is set up as an UNI-light interface type and offer congestion handling.

At the operator network site, the UNI Port Extension is terminated in a WAN port connected to the Ethernet switch in MINI-LINK TN. The WAN port is set an UNI interface type and is the demarcation point between the customer/user domain and the operator domain.

An Ethernet frames priority/VLAN settings are modified at the UNI to adapt to the network domain for incoming traffic and customer/user domain for outgoing traffic. Since the UNI interface is in the switch at the operator site, the UNI port extension WAN link operates on the customer defined priority settings.

In a mobile backhaul network the UNI Port Extension can be used if

• the network operator offer connectivity to other mobile operator’s base stations. In this case a full demarcation point is required.

• the operator’s own base stations don’t tag the Ethernet traffic. In this case the classification and tagging feature must be done in MINI-LINK TN. This use case is common for Ericsson 2G mobile network solutions.


CPE

UNI Port Extension

Network segment

UNIUNI-light

LAN WANWAN

Operator Network domain

Customer/ user domain

L2 Ethernet service

MINI-LINK TN ETSI

I-NNI Port Extension

An I-NNI Port Extension is conceptually equal to an UNI Port Extension. However an I-NNI port extension is used internally in the operator network domain with no need for a demarcation point (i.e. UNI). Therefore, the termination points in an I-NNI Port Extension are set up as I-NNI. In the figure below a device “A” on a remote site in the operator domain is connected to the MINI-LINK TN L2 network.

Figure 14: Connecting device “A” to an operator’s network with I-NNI Port Extension

In a mobile backhaul network the I-NNI Port Extension can be used if

• the base stations tag the Ethernet traffic and are owned by the network operator.

E-NNI Port Extension3

An E-NNI Port Extension can be used as a network connection between two separate operator domains. The E-NNI interface is the demarcation point between the two operator network domains.

In the figure below the E-NNI Port Extension connects operator 1 network domain to the E-NNI interface. The E-NNI Port Extension connection uses operater1’s priority settings.

3 Not supported in current release



I-NNI Port Extension

I-NNI I-NNI

LAN WANWANA

Network segment

L2 Ethernet service

MINI-LINK TN ETSI

Figure 15: E-NNI Port Extension use case

The E-NNI Port Extension is a good alternative for connecting two separate operator network domains that doesn’t have a common site. If the operators share site facilities a direct LAN connection can replace the E-NNI Port Extension connection.

Dedicated Tunnel

The Dedicated Tunnel use case provides an end to end private connection between two network edge LAN ports in an operator’s network. The Dedicated Tunnel WAN connection can be configured with an Ethernet over PDH/SDH or Packet Link.

Figure 16: Dedicated tunnel use case


LAN

Dedicated Tunnel

UNI, UNI-Light,I-NNI

Transport networkLAN



Network segment Network segment


LAN



E-NNI Port Extension

E-NNI I-NNI

LAN WANWAN Network segment

Network segment

L2 Ethernet serviceL2 Ethernet service

UNI, UNI-Light,I-NNI

MINI-LINK TN ETSI

The interface type in the termination points can be configured as UNI, UNI-Light or I-NNI. The actual choice depends on what is connected at the termination points (i.e. LAN ports). The following design rules apply:

• UNI interface type. If full demarcation point (security parts) is required at the network edge towards an external customer. Can operate on both tagged and untagged Ethernet frames. Does not tag frames.

• UNI-Light interface type. If connected customer network is “trusted”. Can operate on both tagged and untagged Ethernet frames.

• I-NNI interface type. Used internally in the operator’s network to connect two network domains. The Dedicated Tunnel using the I-NNI interface type is a transparent tunnel and can connect any packet based network technology (e.g. MPLS) that can be carried over an Ethernet connection. Can operate on both tagged and untagged Ethernet frames

Several Dedicated Tunnel connections can be used to connect multiple not directly connected packet domains. By utilizing the L1 Ethernet service in MINI-LINK TN, it is possible in an efficient and flexible way to combine the clusters into one network. The figure below illustrates the scenario

.

Figure 17: Multiple L1 Ethernet service connections


Transport Network

L1 Ethernet service

Ethernet switch

Network segment

Network segment

Network segment Network

segment

MINI-LINK TN ETSI

4.3.1.4 HW deployment alternatives

This chapter presents different HW configuration alternatives for the L1 Ethernet services use cases previously presented.

Port Extension

From a HW deployment point of view, the different Port Extension variants are equal. The differentiation is done by configuring the termination points with appropriate interface type.

On the central site a NPU3 B is required to interact directly with the MINI-LINK TN L2 switch domain. For the remote site different alternatives are available. The preferred alternative is typically selected based on required capacity for the WAN connection. The following options are available

• WAN capacity up to 30Mbps: Eth/PDH WAN connection with NPU3

• WAN capacity up to 155Mbps: Packet Link WAN connection with Compact Node4

The figure below illustrates a port extension configuration with an Ethernet over PDH connection using NPU3 at the remote site.

Figure 18: Port Extension with Ethernet over PDH WAN connection

If multiple LAN ports are required at the remote site the NPU3 can be replaced with an ETU2 board. In this case, one L1 Ethernet service would be configured per LAN port. An alternative would be to deploy an NPU3 B at the remote site. This would in fact change the Port Extension and L1 Ethernet service to a L2 Ethernet service interface at the remote site

4 Not currently released


LAN

Port Extension

ET

U3

NPU 3 BMM

U2

B

/C/D

/H

LAN

NP

U3

MINI-LINK TN L2 network

MINI-LINK TN

MINI-LINK TN

L2 Ethernet service

A

MM

U2

B

/C/D

/H

MINI-LINK TN ETSI

Dedicated tunnel

A Dedicated Tunnel through an operator’s network can be set up using different HW configurations. The HW alternatives offer different WAN capacity and interface types.

With a NPU3 B in the termination points all interface types and WAN capacity alternatives are supported. The figure below illustrates a NPU3 B with ETU3 configuration utilizing an Ethernet over PDH WAN connection. If the LAN ports on NPU3 B are occupied with other traffic connections, the LAN ports on ETU3 can be used instead.

Figure 19: Dedicated Tunnel with NPU3 B/ETU3 and Ethernet over PDH

If higher capacity is required a Packet Link WAN connection can be used. The figure below illustrates NPU3 B/ETU3 used together with a MMU2 D modem board. Please note that the current implementation is limited to one hop only with a Packet Link WAN connection.

Figure 20: Dedicated Tunnel with NPU3 B and Packet Link


LAN

Dedicated Tunnel

MM

U2

D

MM

U2 D

LAN LANCustomer/user segment

NPU 3 B

NPU 3 B

MINI-LINK TNMINI-LINK TN

Customer/user segment

Operator’s Transport Network

Dedicated Tunnel

LAN LANCustomer/user segment

NPU 3 B

NPU 3 B

MINI-LINK TNMINI-LINK TN


ET

U3

ET

U3

MM

U2

B/C

/D/H

MM

U2

B

/C/D

/H

MINI-LINK TN ETSI

Internally in the operator’s network, if the capacity demand is below 100Mbps, the NPU3 B can be replaced with NPU3 and/or ETU2. The NPU3/ETU2 boards support Ethernet over PDH WAN connections and the I-NNI interface type only. The figure below illustrates a Dedicated Tunnel with NPU3.

Figure 21: Dedicated Tunnel using NPU3

The ETU2 board can be used if multiple Dedicated Tunnels are to be terminated in the same MINI-LINK TN chassis. In the figure below two 30Mbps Dedicated Tunnels are configured.

Figure 22: Dedicated Tunnel using NPU3 and ETU2

The ETU2, NPU3 and NPU3 B/ETU are interoperable for Ethernet over PDH connections, which mean that they can be combined freely up to 30Mbps. For 30-95Mbps Ethernet over PDH connections an ETU board must be in both sides of the Ethernet over PDH connection.


Dedicated Tunnel

LAN LAN

NP

U3

MINI-LINK TNN

PU

3MINI-LINK TN




MM

U2

B

/C/D

/H

MM

U2

B/C

/D/H

LAN

LAN

Dedicated Tunnel

LAN





NP

U3

MINI-LINK TN

MM

U2

B

/C/D

/H

NP

U3

MINI-LINK TN

MM

U2

B/C

/D/H

NP

U3

MINI-LINK TN

MM

U2

B/C

/D/H

MINI-LINK TN ETSI

4.3.1.5 Configuration of L1 Ethernet services

This chapter describes the configuration of a L1 Ethernet service

4.3.1.5.1 Port Extension use case

The following configuration is required for a Port Extension L1 Ethernet service.

Step 1: Configure physical WAN connection

1. Set up WAN connection with the required capacity.

2. Optional: Enable protection if required (e.g. IME)

Step 2: Configure MINI-LINK TN on central site

Prerequisite: switch enabled

1. Connect WAN connection to switch port

2. Configure switch with VID/port settings

3. Optional: change interface type for the WAN port (DEFAULT I-NNI). If interface type changed to UNI, see chapter 4.2.2.1 for more info)

4. Optional: change congestion handling settings for the WAN port. Connected to the switch, the WAN port will inherit the L2 settings (see chapter 4.6.2 for more info)

Step 3: Configure MINI-LINK TN on remote site

1. Connect LAN port to WAN links (PDH-IME)

2. Change interface type for the LAN port to UNI-Light or I-NNI (DEFAULT UNI).5

3. Optional: change interface type for the WAN port (DEFAULT I-NNI). 6

4. Optional: change congestion handling settings (only WAN port). Parameter settings should be equal to WAN set up on the central site (see chapter 4.6 for more info)

5 Currently interface type can not be set for an L1 service6 Currently interface type can not be set for an L1 service


MINI-LINK TN ETSI

4.3.1.5.2 Dedicated Tunnel use case

The following configuration is required for a Dedicated Tunnel L1 Ethernet service.

Step 1: Configure physical WAN connection(s)

1. Set up WAN connections with the required capacity between the termination points.

a. Ethernet over PDH/SDH

i. Cross connect the E1s through the PDH/SDH domain

b. Packet Link

i. Set up Packet Link physical connections

ii. Configure a L1 WAN-WAN connection through the switches in the interim MINI-LINK TN nodes.

2. Optional: Enable protection if required (e.g. IME)

Step 2: Configure termination points

1. Connect the LAN ports to the WAN ports on each side of the Dedicated Tunnel

2. Optional: change interface type for the LAN ports (DEFAULT UNI). If interface type UNI, see chapter 4.2.2.1 for more info)

3. Optional: change interface type for the WAN ports (DEFAULT I-NNI).

4. Optional: change congestion handling settings (only WAN ports) (see chapter 4.6 for more info)


MINI-LINK TN ETSI

4.3.2 L2 Ethernet service

4.3.2.1 General

An L2 Ethernet service enables aggregation of multiple traffic flows. Based on destination MAC address and tagging information a frame is efficiently switched between the different Ethernet interfaces connected to the switch. The aggregation of traffic flows enables statistical gains and a better utilization of the transport network.

In MINI-LINK TN the Ethernet switching is done on the central NPU. Any Ethernet capable LAN or WAN interface in MINI-LINK TN can be connected to the central switch and provide an L2 Ethernet service.

The figure bellow illustrates the scenario were multiple MINI-LINK TN WAN radio ports and local drop LAN ports are connected in a L2 switching domain.

Figure 23: Switching in L2 Ethernet service

Since an L2 Ethernet service supports aggregation of multiple links, there are potential statistical benefits from time and day variation. This means that an L2 Ethernet service with multiple flows potentially can handle traffic bursts without over-dimensioning the network infrastructure.


Switched Transport Network

MINI-LINK TN

WAN

UNI

LAN LAN

MINI-LINK TNUNI

WAN

LAN

MINI-LINK TN

WAN

L2 Ethernet service

UNI

MINI-LINK TN ETSI

4.3.2.2 Characteristics

An L2 Ethernet service offers the following characteristics:

• An L2 Ethernet service offers IEEE802.1D and IEEE802.1Q bridging of unicast/multicast and broadcast service frames between any LAN and/or WAN interface.

• L2CP control frames from connected devices at the UNI will be discarded. Tunneling of external L2CP frames are planned for future releases

• WAN links in an L2 Ethernet service can be Ethernet over PDH/SDH or Packet radio link.

• An L2 Ethernet service can potentially introduce unwanted delay, delay variation a packet loss for a flow due to traffic burst in one or more of the other flows.

• Multiple protection mechanisms can be used for the different WAN and LAN ports in an L2 Ethernet service (RSTP, LAG and RL/PDH IME).

• A LAN port connected to a L2 service has a 100kByte buffer per port.

• A WAN port connected to a L2 service has a 100ms buffer per priority queue/traffic class.


MINI-LINK TN ETSI

4.3.2.3 Use cases

An L2 Ethernet service is used were multiple flows are groomed into a higher capacity link. Aggregation points in ring and tree structures are well suited for L2 Ethernet service deployments. The figure below illustrates an L2 Ethernet service in MINI-LINK TN in a tree and ring topology.

Figure 24: L2 Ethernet service in a ring and tree topology

In the figure above, all WAN and LAN connections are configured as an L2 Ethernet service and connected to the embedded switch. The local drop/LAN port is configured as an UNI interface since in this case the RBS only propagates untagged frames (like Ericsson’s 2G RBS). If the frames are tagged in the RBS (like Ericsson’s 3G NodeB) and the RBS is owned by the operator the local drop/LAN port can be configured as the simpler I-NNI interface type. The north/southbound WAN ports are typically I-NNI interfaces (connected to another MINI-LINK TN L2 switch interface). In addition Port Extensions can be used to connect equipment on remote sites to the L2 switching domain. The figure below shows a remote site connected to a L2 Ethernet service domain with an UNI Port Extension.


UNII-NNI

I-NNI I-NNI

UNI

I-NNI

NORTH

SOUTH

MINI-LINK TN ETSI

Figure 25: UNI port extension connected to a L2 Ethernet service

Typically you will find L2 Ethernet service connections in the center of a network topology were multiple MINI-LINK TN are linked together in a L2 Ethernet service domain. At the edge of the L2 Ethernet service domain you will find multiple I-NNI/UNI port extensions (L1 Ethernet service) connecting remote sites to the L2 domain.

Figure 26: L2 Ethernet service domain with multiple port extensions


MINI-LINK TN

UNI light

WAN Switched Transport Network

MINI-LINK TNWAN

UNI

LAN LAN

MINI-LINK TNUNI

WAN

LAN

UNI

MINI-LINK TN

WAN WAN

L2 Ethernet servicePort Extension

Backbone

L2 Ethernet service domain

L1

L1 L1 L1 L1

L1

L1

L1

MINI-LINK TN ETSI

In some cases, an L2 Ethernet service can be combined with one or multiple L1 Ethernet connections in MINI-LINK TN.

Figure 27: Combined L1 and L2 Ethernet services in one MINI-LINK TN

4.3.2.4 HW deployment alternatives

An L2 Ethernet service in MINI-LINK TN is supported by the embedded switch on the NPU3 B board. The required LAN and WAN ports on the additional interface boards are connected to the switch ports with high speed connections in the backplane. The figure below illustrates the HW build up.

Figure 28: L2 Ethernet service with NPU3 B


NPU3 B ET

U3

L2 Ethernet service

MM

U2

B/C

/D/H

LAN

MM

U2

D

/HS

XU

3 B

MM

U2

E

/F

ET

U3

Ethernet over PDH

Ethernet over SDH

Packet Link

LAN

MINI-LINK TN

WA

N po

rtsL2 Ethernet service

ports

MINI-LINK TN

LA

N p

orts

L1 Ethernet service ports

MINI-LINK TN ETSI

The following dimensioning rules apply for the embedded switch in a MINI-LINK TN chassis:

• Number of switch ports: 9+1

Up to seven high speed connections + two front LAN ports + one DCN over Ethernet (management only). The high speed connections can be any WAN port and/or ETU3 LAN electrical/optical port.

• Switching capacity: 10Gbps

The Ethernet switch is a non-blocking central switch. Below are two examples of required switching capacity in maximum chassis configurations.

AMM6p C maximum configuration:

Two front LAN ports on NPU on NPU (2*1Gbps) 2Gbps

Two LAN ports on ETU3 (2*1Gbps) 2Gbps

Five Packet Link modems (5*0,4Gbps) 2Gbps

DCN over VLAN 0,1Gbps

Total switching capacity 6,1Gbps

AMM6p D maximum configuration:

Two front LAN ports on NPU on NPU (2*1Gbps) 2Gbps

Six LAN ports on ETU3 (6*1Gbps) 6Gbps

One RL-IME group 0,6Gbps

DCN over VLAN 0,1Gbps

Total switching capacity 8,7Gbps

• High speed connection to small boards: 2Gbps

Backplane connection from central switch to boards like ETU3 and SXU3 B.

• High speed connection to large boards: 1Gbps

Backplane connection from central switch to boards like MMU2


MINI-LINK TN ETSI

The embedded switch in MINI-LINK TN can be configured with a number of different WAN and LAN ports. The following dimensioning rules apply.

General

• ETU3 resource control

An ETU3 board provides additional WAN and LAN interfaces to the embedded switch through a high speed connection. Maximum speed on the high speed connection is 2Gbps. Since the combined LAN and WAN ports on ETU3 can exceed this speed limit, a resource control mechanism is supervising the aggregated capacity. When a new LAN or WAN port is added to an ETU3 board, the resource control mechanism verifies that the aggregated speed is below the limit. The port speed used for the resource control mechanism is the maximum possible speed, i.e. for a 10/100/1000Base T port the used speed value is 1000Mbps. The maximum port speed capabilities can be limited from the management system (i.e. fixed mode) to facilitate more LAN ports on ETU3.

Electrical LAN ports

• 1-2 ports: use NPU3 B front LAN ports

• 3-9 ports: use NPU3 B + ETU3 LAN ports. (max four ports per ETU3)

Optical LAN ports

• 1-6 ports: Use ETU3 SFP based LAN ports. (max two ports per ETU3)

WAN ports

• Directions

o Ethernet over PDH: 7 directions with two ETU3 boards

o Packet Link: 5 directions

• Link capacity

o up to 30Mbps: Ethernet over PDH with NPU3

o up to 95Mbps: Ethernet over PDH with ETU3/NPU3 B and/or ETU2

o up to 340Mbps: Packet Link with MMU2 D/H

o up to 600Mbps: Ethernet over SDH with SXU3 B + 4*MMU2 E/F


MINI-LINK TN ETSI

o up to 1Gbps: Packet Link with RL-IME (future)

4.3.2.5 Configuration of a L2 Ethernet service

This chapter describes the configuration of a L2 Ethernet service

Step 1: Configure the switch

1. Enable MAC (802.1D)/VLAN (802.1Q) based switching (requires license)

2. For VLAN based switching, configure VLAN per port

3. Optional: Disable MAC learning (DEFAULT ON)

4. Optional: change aging timer (DEFAULT 300 sec)

Step 2: Connect LAN/WAN ports to switch

Prerequisites: Physical connections established

1. Connect UNI ports

a. Configure the UNI interface type (see chapter 4.2.2.1 for more info)

b. Connect UNI interface to switch port 1-10

2. Connect I-NNI ports

a. Configure the I-NNI interface type (see chapter 4.2.2.2 for more info)

b. Connect I-NNI interface to switch port 1-10


MINI-LINK TN ETSI

4.3.3 Network example

The figure below illustrates a Mobile Backhaul network example with different Ethernet connection services supported by MINI-LINK TN.

Figure 29: MINI-LINK connection service example


NodeB

RBS

NodeB

RBS

NodeB

RBSOptical ring OMS

1410

OMS 1410

RNC/BSCOMS 1410

OMS 1410

OMS 1410

OMS 1410

OMS 1410

OMS 1410

Packet Link

L1 I-NNI port extension

Operator 2

L1 E-NNI port extension

L1 UNI port extension

L2 switching domain

MINI-LINK TN ETSI

4.4 Security view

The connected external customer equipment at the network edge (i.e. UNI), is a potential security threat to an operators network. To counter these threats most operators implement different security features to better control the packet flow in to the operator network. The risk level will to a high degree differ based on the connect customer types. An operator providing internet services to a large number of residential customers will typically need more effective security mechanisms than if only the operators own RBS nodes are connected in a mobile backhaul.

The MINI-LINK TN product family is hardened, from a security point of view, e.g. all unused Operating System ports are turned off. In addition, the following security mechanisms can be deployed:

Frame admittance

The frame admittance feature enables an operator to accept only specified frame types at the network edge. All other frame types are discarded. The feature can be configured individually per UNI port.

The following frame types can be specified

• Admit Q-tagged

• Admit un-tagged

• Admit priority tagged

Admit untagged frames are default setting. This is compliant with Ericsson’s SIU board configuration.

White lists

A white list can be configured individually per UNI port. The white list specifies source MAC addresses allowed to send traffic on a port.

The white list is default off.

It is recommended to configure white lists with expected source MAC addresses for all UNI ports interfacing un-trusted devices.


MINI-LINK TN ETSI

Broadcast/multicast storm protection

Broadcast and multicast filters can be configured individually per UNI port. The filters are specified in frames/sec

The purpose of this functionality is to protect the network from e.g. equipment failure or other scenarios, where a lot of traffic (broadcast and/or multicast storms) is generated at the network.

The filters are default set to 100 frames/sec.

The filters can be individually enabled/disabled per port.

Broadcast and multicast frames outside the filter limit are discarded and counted.

Port blocking

Port blocking can be used to prevent certain ports from sending packets to other ports, even if they are on the same VLAN. Each ingress port can be individual prevented from sending packet to any of the egress ports.

The port blocking is default off.

MAC address limiting per port

The MAC address limiting per port feature prevents one UNI port to flood the entire MAC address learning table in MINI-LINK TN.

The limitation can be individually enabled/disabled per port.

The function is default on for UNI ports and the limit is set to 150 entries.

Frames that violate the limiting function are discarded and counted.


MINI-LINK TN ETSI

4.5 Protection view

Connections with high capacity, serving a large number of customers are candidates for protection in an operator’s network. Thus protected connections are often “high up” in the network were the aggregated capacity and supported subscribers are regarded as critical.

Protection of a connection always comes with a cost so whether to deploy protection or not is often a cost/benefit discussion. The actual capacity and subscriber limits for when to apply protection is very much a strategic decision that is unique for each operator.

MINI-LINK TN offers two main categories of protection

• Network protection mechanisms which span multiple nodes, i.e. cluster

• Link protection mechanisms which protects one physical connection

The two protection categories are illustrated in the figure below.

Figure 30: Network and link protection


Network protection

Link protection

Core Network

RNCBSC

RBS/ Node B

MINI-LINK TN ETSI

4.5.1 Rapid Spanning Tree Protocol (RSTP)

RSTP is used to avoid loops in an Ethernet network.

Use case:

• RSTP is a networking protection mechanism that can be enabled for switch ports in an L2 Ethernet service. This means that RSTP can not be enabled on a port set up for a L1 Ethernet service.

• For a mobile backhaul network, RSTP is primarily used in a ring configuration (up to 20 nodes). However, RSTP can also be used in other topologies with potential loop connections (e.g. mesh).

The figure below shows ten MINI-LINK TN in a ring with one feeder connection. To avoid a loop and offer protection RSTP is enabled.

Figure 31: RSTP protection in MINI-LINK TN ring topology

Characteristics:

• All Ethernet capable LAN and WAN ports connected to the switch in MINI-LINK TN can be configured with RSTP in a ring topology.

• RSTP protection switching time is to a large extent topology dependent. A mesh topology would in general trigger longer RSTP switching times than a ring structure. The switching time is measured as the time period with loss of traffic.


Feeder link

MINI-LINK TN ETSI

• For a ring, typical RSTP switching time is 50-100ms for loss of connection.

Figure 32: RSTP switch time for loss of connection

• The RSTP switching time for reconnection of a lost connection is less, since there is no fault detection time in the reconnection case.

Figure 33: RSTP switch time for re-established connection

• The RSTP protection switching times are only valid for a ring with all MINI-LINK TN elements. Introducing other NE in the ring will not disrupt the RSTP functionality. However, longer switching times can be expected.

• RSTP protection switching is revertive, i.e. when a faulty connection is back in service, a RSTP recalculation will be triggered. In most cases the original set up (breaking point) in the ring will be reinstated.

• RSTP is triggered by physical layer down on active link.

Configuration steps

Step1: Select RSTP root bridge.

1. Set the feeder node in the ring to RSTP root bridge (i.e. northbound interface in the ring). The shortest path cost from each NE is then calculated towards the traffic focal point.

2. The RSTP root bridge selection is automatically done by RSTP (MAC address + NE priority)

3. To manually select the root bridge, the NE priority must be set to minimum value for the desired root bridge NE.


Link down

Link failure detected. RSTP switching

triggered

RSTP switching completed.

Packet loss

Failured Link up

Link up detected. RSTP switching

triggered

RSTP switching completed.

Packet loss

MINI-LINK TN ETSI

Figure 34: Set root bridge

Step2: Enable all RSTP ports in the ring.

1. All LAN and WAN ports in the ring has to be enabled for RSTP. The RSTP process is default turned on in NE.

2. Additional RSTP configuration per port

a. Direct connections (e.g. radio hops) shall be configured as POINT TO POINT

b. LAN ports shall be set up with AUTONEG disabled, i.e. use manual settings for duplex mode and speed

c. Set DISABLE AUTO EDGE for all ports in the ring

d. Set ADMIN EDGE for all ports not part of the ring (e.g. local drop).

Figure 35: Configure ports


RSTP root bridge

Feeder link

RSTP path cost calculation

Configure LAN/WAN ports

in the ring

Feeder link

RSTP root bridge

MINI-LINK TN ETSI

Step3: Configure the breaking point in the ring.

• The breakpoint is automatically set by RSTP based on path cost calculations to root bridge. The path cost is based on link speed.

• RSTP sets the breakpoint such that path cost up to the Root bridge is minimized for each NE in the ring

• For a ring with similar speed on all hops, the breaking point is automatically set “opposite” the Root bridge

• The breaking point can be manually altered by increasing the link cost on the desired link.

Figure 36: Set breaking point

Step4: Verify the following after the configuration is completed

• That the correct NE is selected as Root bridge

• That the correct link has the breaking point. In a ring, only one link shall be in status blocking. The rest of the links shall be in forwarding state.

Figure 37: Verify settings


RSTP root bridge

Breaking point

Feeder link

Traffic flow

Traffic flow

RSTP root bridge

Breaking point

Feeder link

RSTP path cost calculation

MINI-LINK TN ETSI

4.5.2 Link Aggregation Group (LAG)

LAG is used to bundle multiple physical ports in to one logical bundle to offer higher link capacity and protection.

Use case

• LAG is a link protection mechanism that can be configured for switch ports in a L2 Ethernet service. This means that LAG is not supported for a L1 Ethernet service.

• LAG is typically used in the LRAN/HRAN interface where a MINI-LINK TN is connected towards an OMS1410 with multiple LAN ports.

Figure 38: LAG protection in MINI-LINK TN

Characteristics

• All Ethernet capable LAN and WAN ports in MINI-LINK TN set up as a L2 connection service can be configured in a LAG.

• The LAN and WAN ports in a LAG can be distributed on different interface boards.

• One LAG can contain up to eight physical links.

• One MINI-LINK TN can support four LAGs.

• There are no hard restrictions on speed difference between the physical links in a LAG, i.e. a LAG can contain physical links with “any” throughput capacity. However, similar link speed on the LAG links is recommended to better handle the traffic load.

• The traffic load distribution in the LAG is done per flow, i.e. a flow is sent only in one physical link. Frame order must be the same since no sequence number. A flow is identified by source MAC address + destination MAC address (SA+DA).


LAN

LAG

OMS 1410OMS 1410 HRANLRAN

MINI-LINK TN ETSI

• The load balancing and link utilization is better “higher up” in the network with many flows.

• Ethernet carrier loss in one of the physical links in a LAG shall result in no more than 50ms traffic disturbance.

• LAG is revertive, i.e. when a faulty connection is back in service, the connection is automatically reinstated into the LAG with no traffic disturbance.

Configuration steps

1. Configure the first physical link to be included in LAG

2. Connect the link to a L2 switch port

3. Enable LAG for the L2 switch port

4. Define LAG name if required

5. Enable notifications of required (i.e. degraded service, lost links)

6. Configure additional physical links to be included in LAG

7. Connect the additional links to L2 switch ports

8. Include the new L2 switch ports in the LAG

4.5.3 Inverse Multiplexer Ethernet (IME)

The IME protocol is used to transport Ethernet frames over a PDH network (PDH-IME) or natively over a radio hop (RL-IME).

Use case

• When Ethernet is transported over a Packet Link or a PDH network, the IME protocol provides a flexible protection mechanism, i.e. if a link fails the remaining link will carry the traffic.

Figure 39: IME protection in MINI-LINK TN


IME

Network segment

Network segment

MINI-LINK TN ETSI

Characteristics

• Ethernet over PDH with the PDH-IME protocol can bundle both LAN and WAN, fixed and radio connections in one logical connection.

• The IME protection can be used for L1 and L2 Ethernet connection services.

• Maximum 50ms traffic disturbance if a physical link fails or is reinstated.

• A failed link is automatically reinstated in the IME group when the failing condition ends.

Configuration steps

1. Configure the physical links to be included in the IME

2. Include the physical links in an IME group

3. Manual adding or removing physical links in the IME group can disrupt the traffic for max 50 ms.

4.5.4 Link Loss Forwarding (LLF)

LLF speeds up protection switching in external equipment connected to an L1 Ethernet services.

When an L1 Ethernet connection se up with LLF fails, both termination points are taken out of service, i.e. if either one of the LAN ports or the WAN link fails both LAN ports are taken out of service. The figure below illustrates a failure on the WAN link, which with LLF results in disabled LAN ports.

Figure 40: LLF in MINI-LINK TN


LANLAN

MINI-LINK TN ETSI

4.5.5 Other protection mechanisms

In addition to the Ethernet specific protection mechanisms described previously in this chapter, legacy physical layer protection mechanisms can be used when Ethernet is mapped over a circuit based infrastructure.

1+1 radio protection

Use case

• 1+1 radio protection offers L1 protection over a radio hop.

Characteristics

• Completely agnostic to the Ethernet transport.

• Can be used for both L1 and L2 Ethernet connection services.

• Protection switching times dependent of error situation and the protection set up.

Configuration steps

1. No Ethernet specific configuration is required

SNCP

Use case

• SNCP is a networking protection mechanism that can be used in a PDH ring structure.

Characteristics

• Completely agnostic to the Ethernet transport.

• Can be used for both L1 and L2 Ethernet connection services.

• Traffic disturbance below 50ms.

Configuration steps

1. No Ethernet specific configuration is required


MINI-LINK TN ETSI

4.5.6 Multi layer protection

Protection can be enabled both on link and network layers for high availability. In the figure below, RSTP is enabled in addition to LAG/IME on the link layer

Figure 41: Multi layer protection in a ring

RSTP protection switching is triggered when the last physical connection in an IME group is failing.


RSTP OMS 1410

OMS 1410

OMS 1410

OMS 1410

OMS 1410

OMS 1410

IME

LAG

MINI-LINK TN ETSI

4.5.7 Protection example

The figure below is an example on how different protection mechanisms can be used in a network.

Figure 42: Protection example


NodeB

RBS

NodeB

RBS

NodeB

RBSOptical ring OMS

1410

OMS 1410

RNC/BSCOMS 1410

OMS 1410

OMS 1410

OMS 1410

OMS 1410

OMS 1410

Packet Link

RL-IME protection

Operator 2

LAG

RSTP

LAG

MINI-LINK TN ETSI

4.6 Class of Service view

MINI-LINK TN supports a set of Class of Service (CoS) mechanisms to handle congestion situations in the operator’s network. A congestion point is typically an interface in the network with potentially higher input than can be served by the interface. By adjusting the CoS parameters, the operator can secure a predictable propagation of Ethernet frames through the network.

As an alternative to enabling congestion handling mechanisms an operator can oversubscribe all links in the network. This will eliminate the need for any buffering and priority separation of CoS flows.

More information regarding the congestion handling mechanisms in MINI-LINK TN can be found in ref [1]

The differentiation of Ethernet flows in a congestion point is obtained by applying different priority marking on the different flows. The priority marking defines how a frame will be serviced in the buffer compared to other flows. Please note that the priority differentiation between the Ethernet flows doesn’t give any absolute quality guarantees regarding throughput and latency. The priority distinctions provide only an internal relative order for how to expedite frames in a congested port.

4.6.1 General CoS design principles

The following high level design principles should be considered:

• The same CoS settings should be used throughout the network. If different NE is configured differently the CoS behavior will be non consistent.

• Synch and important management traffic should be put on highest priority

• Real time traffic flows (e.g. voice) should be put on higher priority than any BE data traffic

• Buffers for real-time traffic flows should be kept small, i.e. a real time frame has no value after a long buffering period.

• Flow control should only be used for single priority networks, i.e. the priority differentiation is bypassed with flow control since all priorities are treated equal.

• Heavily congested links can be identified with the link utilization and buffer filling PM measurements. The PM data can provide feedback on congestion levels and if more capacity is needed.


MINI-LINK TN ETSI

4.6.2 Configuration of CoS

The following configuration can be done for the CoS mechanisms in MINI-LINK TN. The parameter settings applies to all interface types (e.g. UNI).

For a lot of network installations, the default parameter settings can be used. The table below shows the different CoS parameters and the default settings

Parameter Range Default setting

Egress buffer

# TC queues 1-8 8

Buffer size – LAN Fixed 100kByte for L2 service 100kByte

Buffer size – WAN 1-100ms buffering capacity per TC queue of full link speed. ex) 100Mbps link speed => 2,5MByte buffer size.

Adjustable per TC in 64kByte steps

200ms buffering capacity of full link speed

Buffer per TC queue 0-100% of buffer size TC0: 50%, the rest equally shared

Buffer scheduling

Strict Priority (SP) On(off) per port On

SP and WFQ/WDRR7 templates

On/off per port Off

Discarding of frames

Tail dropping On On

Timestamp based/aging 0-100ms TC0: 100ms, the rest 10ms

Color dropping8 On/off per TC Off

WRED9 On/off per TC Off

Priority classification10

Trust customer priority setting?

Not-trusted, L3 trusted or L2 trusted per port

Not trusted

7 Not currently released8 Not currently released9 Not currently released10 UNI interface type only


MINI-LINK TN ETSI

Default priority (used when not-trusted or no customer priority info.)

0-7 0

L3 classification (L3 trusted must be set)

Customer DSCP/MPLS settings to network priority settings defined in mapping table per port.

Off

L2 classification (L2 trusted must be set)

Customer PCP settings to network priority settings defined in mapping table per port.

Off

Mapping of priorities

L2 priority to TC queue (tagged frames)

IEEE802.1D, IEEE802.1Q, custom mapping per port

IEEE802.1D

L3 priority to TC queue (untagged frames in L1 service only)

Mapping table port. Off

5 References[1] - MINI-LINK TN ETSI Packet Transport Technical Reference


Ethernet guideline

Engineering

ethernet transport network

network types

network example

1l1 ethernet service

2l2 ethernet service

minilink tn capabilities

network topologya mobile

interface types