Mailing List Archive

Using S50n for distribution
I'm considering getting a pair of S50n's for use in a colo setting. I'm
liking the S50n's versus C6500 or C150 due to the following:
* we're not currently CPU bound
* we're doing about 600 Mbps upstream and growing at least 50Mbps/mo so we'll
need to upgrade to 10GE soon
* the S-series is stackable
* dual power supply capable

The biggest questions I have:
* can this handle a filtered BGP feed
* how fast do the switches boot individually and stacked
* do you interact with each switch in a stack or as one; if as one, can you
debug problems on each individual switch?

Any other observations or comments people have about switching from Cisco land
to Force10 land would be appreciated.

-Doug
Re: Using S50n for distribution [ In reply to ]
Doug Warner wrote:

> Subject: [F10-nsp] Using S50n for distribution
>
>

> The biggest questions I have:
> * can this handle a filtered BGP feed
>
Yup. It's the same BGP implementation that you see on the E- and
C-Series, so route-maps, communities, localpref, etc...

> * how fast do the switches boot individually and stacked
>
Individually: about 90 seconds. I haven't tested the stack, but it's
probably only slightly longer, as they all boot in parrallel.

> * do you interact with each switch in a stack or as one; if as one, can you
> debug problems on each individual switch?
>
Stacking creates a single, logical unit. You can still debug based on
individual 'blades' if you will, but globally affecting commands will
still affect things globally. I guess the best way to think about the
individual switches in the stack would be as you would a linecard in a
chassis.


> Any other observations or comments people have about switching from Cisco land
> to Force10 land would be appreciated.
>
I'll let others chime in as I'm sort of biased*

--
Peter Wohlers
*(Force10 SE)

_______________________________________________
force10-nsp mailing list
force10-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/force10-nsp