Well, it depends on your actual traffic volumes.
Cisco routers have performance limitations operating as bridges
or mixed routing/bridgin modes and you might hit the wall well before
155mb/s of bridged traffic.
You could look at ethernet bridging via the Fore switches, can you
plug Fast-E cards into them and let them transparently bridge between
the two sites across the existing OC-12?
You can make the same Class-C appear in two places relatively
transparent to the servers (host routes and proxy-arp/aliasing),
this is fairly trivial if the servers don't need to talk to each
other, a pain if there's a lot community action.
Ideally you'd like to subset things into nice Class-C sized clusters
than can be moved within that 10-server limit, but in addition to
whatever the moving guys say, the more boxes you move at once, the
more downtime gettting them put back together and untangling what
didn't quite go as expected.
If you have $$$ to burn in exchange for a maximally transparent move,
I'd look at a hardware based ethernet bridging solution. Some of the
folks in the corporate world may have a better perspective on large
bridged ethernet configurations in Ciscoland, NSP's usually manage to
dodge that one.
Goerge
> From cisco-nsp-request@puck.nether.net Thu Aug 24 11:34:25 2000
> Resent-Date: Thu, 24 Aug 2000 11:36:38 -0400
> Received-Date: Thu, 24 Aug 2000 11:34:07 -0400
> Reply-To: <tgorman@lucent.com>
> From: "Todd Gorman" <tgorman@lucent.com>
> To: <cisco-nsp@puck.nether.net>
> Date: Thu, 24 Aug 2000 11:33:28 -0400
> Subject: [nsp] Moving a data center
> Resent-From: cisco-nsp@puck.nether.net
> X-Mailing-List: <cisco-nsp@puck.nether.net> archive/latest/3498
> X-Loop: cisco-nsp@puck.nether.net
> Precedence: list
> Resent-Sender: cisco-nsp-request@puck.nether.net
>
> My client is attempting to plan a data center relocation. They have
> approached us for ideas, so I wanted to present it to the Cisco masses to
> take advantage of vast wealth of knowledge out there.
>
> The new data center is currently reachable via ATM over SONET. Between the
> old and the new data centers is a pair of FORE ASX200s. Each FORE ATM
> Switch has two 622Mb connections (one to each neighbor FORE ATM switch via
> the SONET node), and each Cisco 7507 has a single 155MB connection to the
> FORE ATM Switch. Basically its ethernet to ATM/SONET to ethernet. We
> cannot re-address any servers, we must maintain the current address scheme
> (routable class c's, no private addresses). The move does not have to
> happen fast, and the moving company insists it can move only 16 servers at a
> time, even though there are 100's (NT as well as UNIX). This makes us think
> we should either attempt to configure bridging between the old and new data
> center, or perhaps subdivide each subnet and use static routes until the
> entire subnet can be put back together at the new location.
>
> Does anyone have any "industry best practices", ideas, or suggestions on
> keeping this move as transparent to the users as possible? Thanks!
>
>
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:15 EDT