Graceful Routing Shutdown for MC-LAG Software Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-Chassis Link Aggregation Group (MC-LAG) upgrades face network disruption due to simultaneous traffic loss during failover, as control plane protocols shut down during node reboots, leading to in-flight traffic loss during software updates.
Innovation Solution
A method involving a graceful shutdown of routing protocols on one network device, advertising a maximum metric to reroute traffic and allowing the other device to handle network traffic until the update is complete, ensuring minimal disruption by maintaining redundancy through MC-LAG.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If software update is performed on one node of MC-LAG, then the node can be upgraded to newer software version, but network traffic loss occurs during the update process
Solution Approach 1:
The patent applies preliminary action by performing a graceful shutdown of routing protocols before the actual software update. This preliminary step allows the system to prepare for the update by shutting down protocols in a controlled manner, advertising maximum metrics to reroute traffic away from the updating node, and ensuring that the peer node is ready to handle traffic. This prevents traffic loss by preparing the network topology in advance before the node goes offline for updating.
2Ease of manufacture
If control plane protocols are shut down during node reboot, then software update can be applied, but in-flight traffic is lost during the shutdown period
Solution Approach 1:
The patent performs preliminary actions by shutting down routing protocols gracefully before the node reboot occurs. This includes advertising maximum metrics to upstream routers to reroute traffic through the peer node, and notifying downstream devices to prepare for the topology change. By doing this in advance, the actual traffic loss window is minimized to only the brief reboot period, rather than the entire update process.
Solution Approach 2:
The patent applies preliminary anti-action by preemptively rerouting traffic away from the node before it goes offline. Through graceful shutdown procedures, the system advertises maximum metrics to upstream routers and prepares downstream devices, creating counter-actions that prevent traffic from being sent to the updating node. This eliminates in-flight traffic loss by ensuring traffic is already routed through the peer node before the node becomes unavailable.
3Reliability
If simultaneous updates are performed on both MC-LAG nodes, then redundancy is maintained, but complete network failure occurs due to traffic loss
Solution Approach 1:
The patent applies preliminary action by ensuring that one node is fully updated and operational before initiating updates on the peer node. The graceful shutdown procedure on the first node includes preparing the peer node to take over traffic handling. Only after the first node completes its update and is back online does the system proceed to update the second node. This sequential approach with proper handover ensures that at least one operational node always exists in the MC-LAG pair, preventing complete network failure.
Data Source
AI summary
Examples disclosed herein relate to performing a software update on a network device forming a MC-LAG. In an example, a software update onto a first network device and a second network device may be downloaded. The first network device and the second network device may form a MC-LAG that may provide a redundant connectivity to a network device in a network. A shutdown of routing protocols on the first network device may be performed. Each of the routing protocols may advertise a maximum metric on the first network device, leading to a recalculation of network routes by the network device. The first network device may be updated with the software update. Until the software update on the first network device is complete, network traffic on the network may be routed through the second network device.