April 06, 2011

Spanning tree is dead, right?

There has been a lot of press lately around virtual switching and the fact that it’s needed to get a reliable, highly available network. I started thinking about that and the fact that I have what I consider to be a very reliable network and I don’t use virtual switching.

In fact we use a combination of rapid spanning tree, OSPF, and VRRP, which according to many networking companies doesn’t support highly available networking. That said, we make regular changes during the day with no downtime. How can this be?

As I thought about it in the context of five 9’s, it occurred to me that five 9’s is like 315 seconds a year. Rapid spanning tree reconverges in under 2 seconds, so in theory I can have over 150 switch failures and still meet five 9’s availability. Am I missing something?

I’m all for pushing for new technology and there are probably some very valid reasons for virtual switching. If it’s truly easier to configure, that’s a benefit, though I also wonder if easier to configure leads to more wrong configurations. Plus I sort of like having two independent switches. Upgrades are easier and I can let one bake for a day or so before doing the other one, just in case.

Maybe I’m just getting to be a skeptic in my old age mature years, but this sort of feels like a solution looking for a problem. What do others think? Feel free to comment…
About The Contributor:
Extreme Marketing Team

See My Other Posts