Alex - When I have seen packet loss between two providers peering across the various shared medium FDDI interconnects, it has almost always been due to one or more of the following situations : - Network A has congested pipe(s) into the L2 thing. - Network B has congested pipe(s) into the L2 thing. - Network A has congested pipe(s) to the rest of their network. - Network B has congested pipe(s) to the rest of their network. - The peering session between Network A and Network B is traversing the network of FDDI switches at the IX, as the networks are not peering across a FDDI switch that they have in common, and the links are congested. The above situations are mostly correctable, assuming : - Additional FDDI ports are available. - Additional capacity to their network is obtainable. Other things one can do : - FD-FDDI - Peering session can be moved to a FDDI switch that the two networks have in common. Don't get me wrong-- certainly one should explore means of interconnecting beyond the shared medium FDDI option. The various ATM public peering services seem a viable way of scaling public peering as it exists today... while also affording a network the opportunity to put "private [atm-vc] peering" on their marketing web pages. ;) This is my personal observation- hope it helps. - jsb
Does anyone care about trying to get packet loss over MAE-East reduced any more? Some peers have quite heavy packet loss to them from where I'm sitting, and have done for a good while. It seems to me it's not a problem with my port, and I don't think I have head of line blocking problems, which means it's either a Gigaswitch problem, or their ports are simply full. Does anyone still hassle Worldcom and peers about this or have we lost hope of ever getting it fixed (i.e. people peer privately or hope other exchange points or a replacement will come along).
-- Alex Bligh GX Networks (formerly Xara Networks)