RE: VPC layer 2 extension between data centers

That could be one reason for two control planes in vPC vs. one in VSS. I assumed it was because there are certain failure scenarios in VSS that are catastrophic because of the single routing control plane, but vPC can recover from this because they're still separate L2 and L3 devices.

Cisco Live has a good session on this: "BRKRST-2063 - vPC and VSS Best Practice, Deployment and Operation (2012 Melbourne)"
https://ciscolive365.com/connect/sessionDetail.ww?SESSION_ID=2819&backBtn=true

Marc I hope you can make it out to Cisco Live Orlando this year too. We're gearing up for another INE party that's going to be even bigger and better than last year ;)

Brian McGahan, CCIE #8593 (R&S/SP/Security), CCDE #2013::13
bmcgahan_at_INE.com

Internetwork Expert, Inc.
http://www.INE.com

-----Original Message-----
From: marc edwards [mailto:renorider_at_gmail.com]
Sent: Tuesday, March 19, 2013 1:09 PM
To: Brian McGahan
Subject: Re: VPC layer 2 extension between data centers

I immediately thought OTV as well. drawback being OTV need$ Licen$e.

It is interesting how MEC's evolved on the two different product lines. I beleive Nexus didn't logically combine Chassis because the of the unique differences that are required for the FC fabric. Since Catalyst was never interested in unified ports, they went route that 'stacks' the switches on top.

Marc Edwards
CCIE #38259

On Tue, Mar 19, 2013 at 11:00 AM, Brian McGahan <bmcgahan_at_ine.com> wrote:
> This is a normal validated design actually: http://imgur.com/2ILkATY
>
> See "Cisco Data Center Interconnect Design and Implementation Guide"
> http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/ns94
> 9/ns304/ns975/data_center_interconnect_design_guide.pdf
>
> VSS will support the same design here as vPC, but vPC has better support for split-brain detection during certain corner-case failure scenarios. If designed properly -operative word being *if* here :) - you should not get a layer 2 loop like Marc mentioned. Of course whenever you span the layer 2 domain you also span the failure domain, so you have to make sure all your ducks are in a row with things like your root bridge placement, root guard, bpdu filters, bridge assurance, etc.
>
> It would ideally be a better design to do this with OTV vs. Back-to-Back vPC, because OTV has special enhancements specifically for the DCI, such as termination of the STP domain, layer 2 filtering capabilities, ARP caching on the edge, etc. Your limiting factor will be though that OTV isn't supported on Nexus 5K, only 7K. Another option would be to dedicate a separate platform to it, such as ASR 1002, etc. The drawback of going to N7K or ASR of course then is the cost.
>
> So long story short, yes you can do back-to-back vPC for this design and it will work.
>
>
> HTH,
>
> Brian McGahan, CCIE #8593 (R&S/SP/Security), CCDE #2013::13
> bmcgahan_at_INE.com
>
> Internetwork Expert, Inc.
> http://www.INE.com
>
>
> -----Original Message-----
> From: nobody_at_groupstudy.com [mailto:nobody_at_groupstudy.com] On Behalf
> Of marc abel
> Sent: Tuesday, March 19, 2013 11:41 AM
> To: Viet-Hung TON
> Cc: Cisco certification
> Subject: Re: VPC layer 2 extension between data centers
>
> This is technically feasible, but sounds like a truly awful idea to me.
> Rather than adding redundancy you are turning both data centers into one giant failure domain. I just saw a layer 2 loop wipe out an entire data center, with this solution you could wipe out both.
>
> My $0.02 cents.
>
> -Marc
>
>
> On Tue, Mar 19, 2013 at 10:16 AM, Viet-Hung TON <vton_at_integra.fr> wrote:
>
>> Hi all,
>>
>> Do you have any experiences about using VPC to connect between two
>> data centers in different geography? I plan using 2 nexus 5000 each
>> site fo redundancy, the connections are based on physical dark fibers
>> WDM. The latency between 2 sites is around 1.5 - 2 ms. Otherwise, the
>> VSS runs well in this case?
>>
>> Thanks for your share.
>>
>> Viet
>>
>>
>> Blogs and organic groups at http://www.ccie.net
>>
>> ______________________________**______________________________**
>> ___________
>> Subscription information may be found at:
>> http://www.groupstudy.com/** list/CCIELab.html
>> <http://www.groupstudy.com/list/CCIELab.html>
>>
>>
>>
>>
>>
>>
>>
>>
>
>
> --
> Marc Abel
> CCIE #35470
> (Routing and Switching)
>
>
> Blogs and organic groups at http://www.ccie.net
>
> ______________________________________________________________________
> _ Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
> Blogs and organic groups at http://www.ccie.net
>
> ______________________________________________________________________
> _ Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html

Blogs and organic groups at http://www.ccie.net
Received on Tue Mar 19 2013 - 13:21:04 ART