Stateful Switchover versus Virtual Router Redundancy Protocol?

A

Argher

Limp Gawd
Joined
Jun 23, 2004
Messages
374
Is anyone familiar with VRRP as opposed to Cisco's stateful switchover? I'm trying to achieve highest availability in my network without breaking the bank to quite the tune of Cisco's gear. HP offers some nice options with VRRP in their ProCurve line, however as I understand it, stateful switchover not only switches between switches, but also between individual route processors on a switch, implying a redundancy of processor on the individual switches themselves - is this the case?

Are there other options for high availability networks that I have not seen?

Thanks much.
 
No...HSRP is more or less identical to VRRP.... SSO = Stateful Switchover, a relatively new feature on Cisco gear.

I'm thinking about whether having SSO is worth the additional investment over VRRP (the standards-based version of HSRP).
 
This all depends on your application. Are you trying to engineer something from the ground up? SSO with HSRP or VRRP will just add one more layer of redundancy to your switched network. Yes, it will failover between route/switch supervisor processors without a blip due to the state table synch.

Is it worth the investment? Well... technically you can have one chassis with 2 supervisor engines in it with SSO and have a fully redundant switch network without HSRP or VRRP. But If your chassis goes down, your SOL... Then again, you could buy two chassis, 1 sup for each chassis and run HSRP/VRRP between the two and have the same type of redundancy...just much more cost involved with the second chassis, electricity, etc.... Like I said earlier, it all depends on the application.
 
In this case I believe we would be using SSO in addition to HSRP, which seems to me to be a little overkill - or do you think it truly would add another "9" to our uptime? I'm thinking not necessarily...the chances of one switch and an RP on another switch doing down at the same time seems absurdly miniscule, no?

By that token, HSRP/VRRP seems like a more failure-proof method than purely SSO, as what happens if the fabric goes bad in an SSO switch? You'd still be outta luck, no?
 
as Xphile and others have said, statefull switchover will make your HSRP or VRRP happen just a little bit faster and without killing existing established tcp sessions (hence statefull).

I think HSRP / VRRP has a few seconds of downtime, but from what I have been reading lately things like switch stacks will still attempt to forward out packets as they are learning the cam / arp etc. tables.

with HSRP/VRRP you will be redundant but your users might notice a second or two blip. SSO will eliminate the blip.

If you were my bank / trading firm / etc, I would want you to setup SSO with HSRP/VRRP for critical infrastructure such as database etc.
 
You must log in or register to reply here.
Back
Top