Network Redundancy – The What, Why & How?

May 29, 2018 Enterprise Networking

Network RedundancyBen-Hur, great movie…Charleton Heston, great actor…. The Chariot Race, it is, to this day, considered one of the greatest action sequences ever made. Did you know they used 82 horses imported from Yugoslavia and 8,000 screaming extras to fill the arena? Today, most of those horses and people would be replaced by CGI. Extras are now redundant and no longer needed. That type of redundancy is slightly different than the redundancy we talk about in those cold data centers and dingy network closets. Network redundancy is, according to Techopedia, a process through which additional or alternate instances of network devices or equipment are installed within network infrastructure. It is a method for ensuring network availability in case of failure. In real talk – it’s what allows you to sleep at night, go on vacation for weeks at a time or take that Friday off to play the back nine. It’s all necessary – hence the need for network redundancy.

How do we achieve network redundancy? You think you have the answer; extra cabling and more power, right? Sure, those are ways. But lets talk about two technologies we’ve implemented quite a bit – vPC and VSS. Both of these technologies provide ways of achieving redundancy. Both have concepts in common, but both are different. Despite some of these differences, they can be in production on your network at the exact same time.

vPC: Virtual Port Channel is native to the Nexus line of switches (3k/5k/7k/9k.) We’re all familiar with normal etherchannel. But by virtualizing your etherchannel, you’re taking links that are physically connected to two different Nexus switches and connecting them to a third device thereby making it appear as a single etherchannel to this third device. What’s great about this is that the third device can be a switch, server, or any other device that supports link aggregation. By the way, it doesn’t have to be a Cisco device!

Many clients have this in place, but might not understand what they’re getting out of it.  “My Nexus switches, they’re really one switch, right?” No. Each vPC switch runs its own control plane and they both work independently of each other. Without getting too technical, consider the following about vPC:

  • vPC eliminates STP blocked ports
  • Uses multiple uplinks for more bandwidth
  • Feature fast convergence and failover
  • Each vPC peer processes half of the traffic from the access layer
  • Two switches = two control planes. Any issues stay local to that switch.

VSS: Virtual Switching System can be done with your 4500/6500 chassis (check your Supervisor Module) or the Catalyst 4500-X. The concept is to have two chassis’ configured to appear as one single logical switch. One chassis is configured as “Active” and the other is a designated “Standby”, in case you happen to be sipping margaritas in Mexico.

What does the Active chassis do? Everything….layer 2, Layer 3, SNMP, SSH, etc. eems It seems like a lot to put on one chassis, doesn’t it? Not really. The idea is to make it easy on you by only having to manage one IP and configure one switch. If the Active goes down, everything shifts to the Standby without any intervention on your part. The key is to make sure everything important is dual-connected to both chassis. Network closets, servers, firewalls, WAN links, etc. The brains between the Active and Standby chassis’ reside in the VSL link. It’s what keeps the two in sync.

The key similarity between the two designs is Multi-chassis Etherchannel (MCEC). All links being active provides for a highly available and highly resilient network. Despite the differences, the goal is the same – no downtime (in your network) so that you can enjoy a little personal downtime yourself!


Bobby Grewal

Bobby Grewal  is a Networking Engineer rockstar with years of networking know-how under his belt!