MC LAG

MC-LAG is a special case of LAG (link aggregation group) covered in IEEE 802.3ad in which the LAG ends terminates on separate chassis. To understand MC-LAG, it is eminent to understand the basic concepts of LAG first. These thoughts then can be extrapolated to fit the MC-LAG working

LAG (Link Aggregation Group) is a method of inverse multiplexing over multiple Ethernet links, and is defined by the IEEE 802.3ad standard. The standard states, “Link Aggregation allows one or more links to be aggregated together to form a Link Aggregation Group, such that a MAC client can treat the Link Aggregation Group as if it were a single link”. This layer 2 transparency is achieved by the LAG using a single MAC address for all the device’s ports in the LAG group. LAG can be configured as either static or dynamic. Dynamic LAG uses a peer-to-peer protocol for control, called the Link Aggregate Control Protocol (LACP). This LACP protocol is defined within the 802.3ad standard.LAG can be implemented in two ways. LAG N and LAG N+N. LAG N is the load sharing mode of LAG and LAG N+N provides the Worker Standby Flavour.

A basic use of LAG to “trunk” 4 Ethernet links is shown in the figure below. This is an example of LAG N where N is 4. LAG group is controlled by Actor/Master. The partner can act as a Slave or Master depending on the LACP priority number set between the two. The LAG N protocol automatically distributes and load balances the traffic across the working links within a LAG, so maximising the use of the group if Ethernet links go down or come back up, so providing improved resilience and throughput.

For a different style of resilience between 2 nodes, a complete implementation of the LACP protocol supports separate worker/standby LAG subgroups,For LAG N+N the worker links as a group will fail over to the standby links if any one or more or all of the links in the worker group fail. Note, LACP marks links as in standby mode using an “out of sync” flag.

A further extension of LAG is MC (Multi-Chassis) LAG. This provides redundant Layer 2 access connectivity that adds node-level redundancy in addition to the normal LAG link-level redundancy, allowing two or more systems to share a common LAG endpoint. The multiple end points present a single logical chassis to the start point, such that the start node does not need to be aware that MC-LAG is being used