I don't think you can do the 1G distribution on the 4900M without converting the 10G interfaces back to dual 1G. I have heard from others on the list that this severely limits your queue sizes but. Ymmv. Beating the multi-vendor drum this is a perfect use for the juniper ex4200 series. I have been giving my cisco se a hard time because they don't have an all fiber stackable with dual power that can do 10G. The ex will give you 24 fiber ports with 2 10G per switch for about 10k list. I know this is A cisco list but it's what I'd use. I have a hunch that they do this by design to force us to buy chassis based switches.
In your situation I'd check when the sup32 goes eos/eol. You may be dodging a bullet by upgrading to the 720. Sent from my iPhone On Dec 23, 2010, at 9:05 AM, Holemans Wim <[email protected]> wrote: > We have 3 campus with on each campus a 6506-E/Sup720-10G as 'master router' > and a 6506/E-Sup32-8gbit as backup router, in a HSRP config. In each router > we also have GBIC boards to connect the different buildings. These Sup32 > routers also act as L2 concentrator for part of each campus. > > Now we are thinking about connecting both routers to each other on each > campus with a 10G connection. As the Sup32 don't have a 10G yet, we have > multiple options to do so. > We can add a 10G board to the chassis, replace the supervisor with a Sup720 > or replace the whole router with a 4900M. > If I take a look at listprices, I get 28000$ for Sup720, 20000$ for 6704 (but > these are Xenpacks), 37500$ for 6708 and 22000$ for 4900M (base + 10/100/1000 > card, dual power). > We have 65XX as routers because we had FWSM boards in them but these are > not used anymore. > Based on the price, it seems we best opt to replace the 6506-E/Sup32 with the > 4900M option (there is also a difference in maintenance cost). With Twingig > convertors this offers us a good combination of 10G and 1G SFP ports. For > 7500$ we can add a second 8 port X2 board that gives us extra 10G/SFP-ports > if needed. > > Has anyone had bad/good experience with using a 4900M as router, given the > following environment : > > - Router acts as backup router, so in 99.xxx% of the time it only > has to forward L2 traffic > > - Only static routes, no active routing protocol. > > - 40 vlans, 40 SVI's with ACLs on it > > - No IPv6 for the moment, but according to the specs, the 4900M > should handle IPv6 in hardware just fine. > > - No Qos yet, but we are planning to implement that in 2011 > > > I know we lose the netflow capability if the primary router fails, but we can > live with that. > > All comments are welcome. > > Wim Holemans > Network Services > University of Antwerp > Belgium > > _______________________________________________ > cisco-nsp mailing list [email protected] > https://puck.nether.net/mailman/listinfo/cisco-nsp > archive at http://puck.nether.net/pipermail/cisco-nsp/ > _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
