ICCP Multichassis VLAN Redundancy

Carrier Ethernet network high availability can be achieved by employing intra- and inter-chassis redundancy mechanisms. The Multichassis Link Aggregation Control Protocol (mLACP) solution addresses the latter, where a carrier wants dual-homed device (DHD) to two upstream points of attachment (PoA) for redundancy. Some carriers do not run loop prevention control protocols in their access networks, so an alternate redundancy scheme is necessary.

The implementation of mLACP supports DHD with an active/standby topology. Interchassis Communication Protocol (ICCP) Multichassis VLAN Redundancy, also known as Pseudo mLACP, provides a flexible dual-homing redundancy mechanism. It uses similar principles as mLACP. The Pseudo mLACP solution extends the mLACP functionality to support active/active PoAs deployments. This enables flexibility in network planning and efficient resource utilization.

Pseudo mLACP has the following advantages over mLACP:

  • Pseudo mLACP supports per-VLAN active/active redundancy without any load-balancing requirements on the CE.

  • Pseudo mLACP is independent of the access redundancy mechanism; therefore, it provides a network-based redundancy solution. It allows maximum flexibility for the Provider Edge (PE)-Customer Edge (CE) interoperability in terms of dual-homing redundancy and recovery.

Prerequisites for ICCP Multichassis VLAN Redundancy

  • mLACP support is required for Pseudo mLACP.

Restrictions for ICCP Multichassis VLAN Redundancy

  • Max bundle should not be configured on a Pseudo mLACP enabled port channel.

  • Pseudo mLACP does not work with most of the Layer 2 control protocols or Spanning Tree Protocol (STP) including Multiple Spanning Tree Protocol (MSTP) or VLAN Trunking Protocol (VTP).

  • When a service instance is configured under a Pseudo mLACP port channel, all the outer tag VLANs of a service instance must be a part of either a primary VLAN list or a secondary VLAN list.

  • Outer VLANs of one service instance cannot be mixed with the primary and secondary VLAN list on a Pseudo mLACP port channel.

  • Brute-Force mode configuration is not supported.

  • VLAN force-switchover configuration is applicable only for nonrevertive mode.

  • The DHD nodes must support the LACP functionality.

  • The DHD nodes must support MVRP MAC flush functionality in Pseudo mLACP topology.

Information About ICCP Multichassis VLAN Redundancy

Pseudo mLACP Multihoming Redundancy

The provider edge (PE) ports are configured in such a way that they act as if connected to a virtual device over a Multichassis link aggregation group (MC-LAG) with mLACP. Points of Attachment (PoAs) can be placed in active/active mode with manual VLAN load balancing. DHD ports are configured into two individual port channels that are physically connected to each of the PoAs. Interchassis Communication Protocol (ICCP), with new extensions is used for interchassis communication to control the failover process. Multiple VLAN Registration Protocol (MVRP) lite is used for active VLAN notification and MAC flushing toward the access side. For MAC flushing notification toward the core, MVRP lite, Multiple I-SID Registration Protocol (MIRP) lite, or LDP MAC withdraw can be used.

Pseudo mLACP provides:

  • The active/active mode redundancy of two PoAs in a redundancy group. This provides higher bandwidth utilization than mLACP and other active/standby link-level schemes. Pseudo mLACP eliminates the required wasted link bandwidth on the standby PoA.

  • Flexible access network topologies, that is, access network dual-homing and access device dual-homing.

  • Service provider control over the provisioning, role assignment, failover, and load sharing between PoAs.

  • PE node redundancy for Virtual Private Wire Service (VPWS), Virtual Private LAN Service (VPLS), and native Ethernet aggregation.

The DHD is configured with two different port channels that are connected to a single multichassis LAG (mLAG) on the PoA side. The LACP module on the PoAs receives two different port keys from the two different port channels on the DHD. The mLAG on the PoA ignores the port keys from the DHD’s LACP PDUs to form a single bundle on the PoA side.

The mLACP module provides failover and recovery notifications to Pseudo mLACP. Reversion delay is processed by the mLACP module. mLACP provides a CLI interface for Pseudo mLACP VLANs and mode configuration. mLACP supports VLAN-based active/active redundancy, in addition to PoA-level active/standby redundancy. VLAN-based active/active redundancy allows you to bundle links on both the PoAs based on the Pseudo mLACP configuration. Pseudo mLACP and mLACP port-channels can be configured together on the same pair of PoAs, and both can use the same redundancy group.

After failover, the new active PoA activates the standby VLAN list on the port-channel. However, to receive traffic on the newly active VLAN’s DHD, networks must flush their MAC address table and learn the new MAC address of the new PoA port channel interface. The existing MVRP lite support is used for DHD-side MAC flushing.

Pseudo mLACP Active/Active Support

Pseudo mLACP supports active/active redundancy without the restriction of symmetric VLAN-based load sharing in both the Provider Edge (PE) and the Customer Edge (CE).

Figure 1. Active/Active Support

Pseudo mLACP provides VLAN-based redundancy by allowing you to specify one primary interface and one secondary interface or a PoA pair for each member VLAN. The configuration determines the PoA that will be initially active for a VLAN, by using the primary and secondary VLAN lists under the Pseudo mLACP interface. Only the active PoA will forward frames for the respective VLANs. The standby PoA will be in the blocking mode (bidirectional), dropping all the frames received on the standby VLANs. The failover will occur for all the VLANs in the active/standby list and not on a per-VLAN basis. Pseudo mLACP provides per-port-channel VLAN load balancing. You can statistically configure the primary and secondary VLAN list on each of the PoAs.

The DHD nodes are configured such that each of their uplinks to a PoA operates as an individual port channel. Each interface must be configured to forward all local VLANs on all uplinks belonging to the mLAG.

The data-path forwarding scheme causes the DHD to automatically learn which PoA or interface is active for a given VLAN. This learning occurs at an individual destination MAC address level.

Failure Recovery

Pseudo mLACP uses revertive behavior (which is the default behavior) after the failure recovery to support the active/active model. You can configure a nonrevertive mode.

Reversion occurs the same way that the original failover occurs. The reversion must be initiated by the new active PoA for the given VLANs, by signaling that the PoA is relinquishing its active role for the VLAN. This is done through an ICCP Pseudo mLACP port-state TLV, which indicates that it is no longer in the active mode for the affected VLANs. Upon a TLV receipt, the recovering PoA unblocks the affected VLANs, and triggers MAC flushes toward both the access side and the core side).

mLACP reversion delay applies for Pseudo mLACP operations. However, reversion occurs only for failed-over VLANs. The forced failover mechanism based on dynamic port-priority change cannot be used for Pseudo mLACP because all the member links will remain in the bundle state. Use the mlacp reversion-delay command to configure the mLACP reversion timer. Use the mlacp load-balance force-switchover portchannel command to configure forced VLAN switchover.

Pseudo mLACP Failover Operations

The Pseudo mLACP forces a PoA failover to the standby PoA when one of the following failures occurs:


Note

mLACP failover will not be triggered if Pseudo mLACP is not configured correctly. If the mLACP failover occurs before the peer PoA is configured with Pseudo mLACP, the failover will occur as long as the peer PoA meets the mLACP failover requirements.

  • Access side link or port failure—This failure is triggered by a min-link failure. On receiving a min-link failure, all the active VLANs on the port-channel failover to the other PoA. This failover is initiated by sending a Pseudo mLACP PORT-STATE TLV message, indicating that the port state is DOWN.

  • Node failure—The surviving PoA’s Pseudo mLACP receives notification of node failure and initiates failover of all VLANs that were in standby mode on all shared mLAGs. After recovery, both POAs synchronize again.

  • PoA uplink failure—The failing PoA signals the peer about the core isolation using the Pseudo mLACP PORT-STATE TLV, indicating that the PoA is isolated. It places all the VLANs in the blocking mode.

How to Configure ICCP Multichassis VLAN Redundancy

Configuring a Port Channel for Pseudo mLACP

Perform this task to configure a port channel for Pseudo mLACP.

Before you begin


Note

The redundancy group should be configured. Redundancy group configuration for Pseudo mLACP is the same as for mLACP.

SUMMARY STEPS

  1. enable
  2. configure terminal
  3. interface port-channel number
  4. mlacp interchassis group group-id
  5. mlacp mode active-active
  6. mlacp load-balance primary vlan vlan-id
  7. mlacp load-balance secondary vlan vlan-id
  8. end

DETAILED STEPS

  Command or Action Purpose

Step 1

enable

Example:

Router> enable
Enables privileged EXEC mode.

  • Enter your password if prompted.

Step 2

configure terminal

Example:

Router# configure terminal

Enters global configuration mode.

Step 3

interface port-channel number

Example:

Router(config)# interface port-channel 1

Configures the port channel and enters interface configuration mode.

Step 4

mlacp interchassis group group-id

Example:

Router(config-if)# mlacp interchassis group 1

Specifies that the port channel is an mLACP port-channel .

Step 5

mlacp mode active-active

Example:

Router(config-if)# mlacp mode active-active

Enables pseudo mLACP operations on the PoA and allows the PoA to form an LACP bundle even if the partner receives an LACP PDU from two different port channels on a dual-homed network (DHN) or dual-homed device (DHD).

Step 6

mlacp load-balance primary vlan vlan-id

Example:

Router(config-if)# mlacp load-balance primary vlan 10,20

Configures the list of primary VLANs that will be active and inactive on the given PoA.

Step 7

mlacp load-balance secondary vlan vlan-id

Example:

Router(config-if)# mlacp load-balance secondary vlan 30,100

Configures the list of secondary VLANs that will be active and inactive on the given PoA.

Step 8

end

Example:

Router(config-if)# end

Exits interface configuration mode and returns to privileged EXEC mode.

Configuration Examples for ICCP Multichassis VLAN Redundancy

Example: Port Channel Configuration for Pseudo mLACP

The following example shows how to configure the port channel on the active and standby PoA for Pseudo mLACP.

Active PoA-POA1 

Router# configure terminal
Router(config)# interface port-channel1
Router(config-if)# mlacp interchassis group 1 
Router(config-if)# mlacp mode active-active 
Router(config-if)# mlacp load-balance primary vlan 10,20  
Router(config-if)# mlacp load-balance secondary vlan 30,100 
Router(config-if)# end

Standby PoA-POA2 

Router# configure terminal
Router(config)# interface port-channel1
Router(config-if)# mlacp interchassis group 1 
Router(config-if)# mlacp mode active-active 
Router(config-if)# mlacp load-balance primary vlan 30,100  
Router(config-if)# mlacp load-balance secondary vlan 10,20 
Router(config-if)# end

Feature Information for ICCP Multichassis VLAN Redundancy

The following table provides release information about the feature or features described in this module. This table lists only the software release that introduced support for a given feature in a given software release train. Unless noted otherwise, subsequent releases of that software release train also support that feature.

Use Cisco Feature Navigator to find information about platform support and Cisco software image support. To access Cisco Feature Navigator, go to www.cisco.com/go/cfn. An account on Cisco.com is not required.
Table 1. Feature Information for ICCP Multiichassis VLAN Redundancy

Feature Name

Releases

Feature Information

ICCP Multichassis VLAN Redundancy

Pseudo mLACP provides a flexible dual-homing redundancy mechanism. It uses similar principles as mLACP, but without the implementation of LACP between the PEs and CEs. The PE ports are configured in such a way that they act as if connected to a virtual device over an MC-LAG with mLACP. Ports can be placed in active/active mode with manual VLAN load balancing.

The following commands were introduced or modified: debug lacp, debug mvrp, mlacp load-balance, mlacp load-balance force-switchover, mlacp mode active-active, mlacp reversion-delay, show lacp.

Glossary

active attachment circuit—The link that is actively forwarding traffic between the DHD and the active PoA.

active PW—The pseudowire that is forwarding traffic on the active PoA.

BD—bridge domain.

BFD—bidirectional forwarding detection.

DHD—dual-homed device. A node that is connected to two switches over a multichassis link aggregation group for the purpose of redundancy.

DHN—dual-homed network. A network that is connected to two switches to provide redundancy.

H-VPLS—Hierarchical Virtual Private LAN Service.

ICC—Interchassis Communication Channel.

ICCP—Interchassis Communication Protocol.

ICPM—Interchassis Protocol Manager.

ICRM—Interchassis Redundancy Manager.

LACP—Link Aggregation Control Protocol.

LAG—link aggregation group.

LDP—Link Distribution Protocol.

MCEC—Multichassis EtherChannel.

mLACP—Multichassis LACP.

PoA—point of attachment. One of a pair of switches running multichassis link aggregation group with a DHD.

PW-RED—pseudowire redundancy.

standby attachment circuit—The link that is in standby mode between the DHD and the standby PoA.

standby PW—The pseudowire that is in standby mode on either an active or a standby PoA.

uPE—user-facing Provider Edge.

VPLS—Virtual Private LAN Service.

VPWS—Virtual Private Wire Service.