[c-nsp] IOS reliability

[Peter Rathlev]

On Wed, 2009-01-07 at 11:18 +0000, A.L.M.Buxey at lboro.ac.uk wrote:
> there can be the occasional bug or glitch - mainly (in
> our experience) due to new types of traffic on the network.
> our main downtimes are due to power outage (in excess of
> our UPS 6hr capability) and firmware updates.
> 
> however, you can help to negate any IOS issue by eg
> using dual supervisors in the 6500 platform, or by
> using eg s720-10G virtual router architecture. - or
> in the 3750 world, by having dual-homed stacks etc.
> 
> ie use resiliency to work around any possible IOS/platform
> issues (this is true for any chosen network vendor!)

We have duplicated chassis' everywhere, relying on StackWise/STP/HSRP/
BGP (depeding on the situation) to take care of failures. This gives us
some figures on how long time the network will take to converge, but of
course it doesn't tell us how often this will happen. :-)

> of course, this then comes down to the classic 'if you run
> the same version on both boxes they'll get hit by the same
> bug' perennial question - I'm always happy to run a slightly
> newer version alongside the old trusted....and safeharbor
> is always a good option unless you cant use it (eg new a feature
> thats not in a safeharbor release!)

We've given this some thought, especially since the failover relies only
on chassis independent protocols, but haven't gotten around to doing
something about it. But we'll consider closer in the future.

Thanks for the input.

Regards,
Peter