Introduction

Multichassis Link Aggregation (MC-LAG) is an ASR9k link aggregation between two different chassis. This document provides an overview of MLAG.

MC-LAG Overview

MC-LAG allows for a third device to connect a Bundle interface to two different ASR9K chassis.
It provides a link and device-level redundancy.
Only one device performs the forwarding (Active/Standby).
Cisco nomenclature refers to the CE at Dual-Homed-Device (DHD) and the two PEs configured with MC-LAG as the **Point of Attachment (POA).
Loss of connectivity between the POAs will lean to a split brains, where both of the devices try to become active.
The Interchassis Communication Protocol (ICCP) manages the setup and controls the Redundancy Groups between PEs.
There are two main groups of redundancy models Access Network Redundancy Model and Core Network Redundancy Model.
- Access Network Redundancy Model
  - ICCP Service Multihoming (ICCP-SM)
  - mLACP
  - Pseduo mLACP
- Core Network Redundancy Model
  - One Way Pseudowire Redundancy
  - Two Way Pseudowire Redundancy

Control used to manage the Redundancy Groups (RG) between two of the POA chassis.
It uses an LDP link between two PEs.
The PE can be directly connected or using multiple hops to connect to each other.
A PE can be a member of multiple Redundancy Groups.
The ICCP protocol operates between the active and the standby POAs, and allows the POAs to coordinate their configuration, determine which POA is active, and trigger a POA to become active.
Applications running on the two POAs (mLACP, IGMP snooping, DHCP snooping or ANCP) synchronize their state using ICCP.

Basic configuration of an MC-LAG:
- Configure the ICCP redundancy group for the POA.
- Define the redundant neighbor IP address (LDP Router-ID)