Arista Campus Operations Specialist · Campus operations

MLAG in the campus: peer link, dual-homed access and failure behaviour

MLAG in the campus: peer link, dual-homed access and failure behaviour. Reading chapters, hands-on simulator labs, interview questions, scenarios and a quiz.

43 min read9 chapters2 labs12 quiz6 scenarios12 interview Q&A

This first module is free: read the lesson and take the quiz. Labs need an account: sign up free to run up to 3 hands-on labs.

Log inStart free trial
This reading material cannot be printed. Read it on config.nwkings.com.
01

Why MLAG: two switches acting as one for the access layer

In a campus, an access switch or a server needs two uplinks so one failure does not cut it off. Spanning tree would block one of them. Multi-chassis link aggregation (MLAG) lets two distribution switches look like one switch to the device below, so both uplinks carry traffic.

Analogy: a shop with two doors run by two staff who share one counter

Customers see one shop. If one door closes, they use the other and are served as before. The two staff keep one shared ledger so nothing is lost. MLAG peers keep one shared view of MAC addresses and ports.

What MLAG gives

  • Both uplinks forward traffic: no blocked links.
  • Switch or link failure leaves the access device connected.
  • Maintenance without outage: upgrade one peer at a time.
  • Simple for the downstream device: it sees an ordinary port channel.

Spanning tree versus MLAG

AspectSpanning tree onlyWith MLAG
Redundant linksOne blocked per loopAll active
FailoverReconvergence, secondsFast, link-level
BandwidthHalf wastedFully used
DesignSimple, wastefulNeeds a peer link and care

Where it sits in the campus

MLAG pairs are usually placed at the distribution or leaf layer, with access switches and servers dual-homed to them. Above the pair, routing links or a layer 3 uplink connect to the core or spine.

Access switchPeer APeer BCore

One port channel to two peers.

A floor access switch has one port channel with a member to each distribution peer. When Peer A is reloaded for an upgrade, the member to Peer B keeps carrying traffic and users do not notice.

Common mistake: treating MLAG as a way to avoid thinking about loops. You still need a clean design and loop protection.

The blocked uplink

A floor had two uplinks to different switches and half the bandwidth was idle because spanning tree blocked one. Moving to an MLAG pair gave full use of both links.

Lesson: use MLAG for dual-homing.

What problem does MLAG solve?

It lets a device connect to two switches with one port channel so both links are active and a switch failure causes no outage.

Key takeaways

  • MLAG makes two switches look like one.
  • All links forward; no blocked ports.
  • Upgrade one peer at a time.
  • Typically at distribution or leaf.
  • Still needs loop protection.

Upgrade to keep reading

The free plan lets you read chapter 1 of this module. Upgrade to open the remaining chapters.

  1. The parts of an MLAG pair: domain, peer link, peer address
  2. State synchronisation and how traffic flows
  3. Split-brain: what it is and how to protect against it
  4. Failover and fast convergence for access switches
  5. Verifying and reading MLAG state
  6. Troubleshooting MLAG faults
  7. Maintenance and upgrades on an MLAG pair
  8. Module summary and checklist

Join the community