[c-nsp] 6500 VSS for campus L3 core?

[Oliver Garraux]

That mirrors my thinking as well.  I don't think doing VSS in purely
routed environment like a core really buys you anything.  Other than
more risk :)

And even at the distribution level, I'd be a lot more comfortable with
VPC rather than VSS given that VPC gives you kind of separate control
planes.

Oliver
-------------------------------------

Oliver Garraux
Check out my blog:  blog.garraux.net
Follow me on Twitter:  twitter.com/olivergarraux


On Wed, Feb 13, 2013 at 11:55 AM, Andy Ellsworth <andy at dar.net> wrote:
> For those of you running 6500 VSS and using the traditional 3-layer campus
> model (core/distro/access) with a layer 3 core...do you use VSS in your
> core?
>
> We are using VSS in the core and distro layers today, but the more
> $WEIRD_THINGS software bugs we encounter that affect the whole VSS pair on
> our distros, the less inclined I am to keep VSS in the core.
>
> My thinking is that at the distro layer, you at least get some significant
> benefits for your layer 2 downlinks to your access switches (MEC, no STP,
> no FHRPs). But in the core, where all of your connectivity is
> point-to-point L3 links to distros, we can get all of those same benefits
> with EC + ECMP OSPF (and we don't have anything hitting the core that would
> need FHRPs).
>
> The conclusion I've come to is that VSS at the distro layer has significant
> benefits with some downsides, while VSS for a L3 core has very few benefits
> but increased risk of software/human error failures.
>
> Has anyone else gone through this thought exercise?
>
> -Andy
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/