>Is a different load balancing algorithm possible here? Perhaps flow-based >load-balancing instead of packet-based would solve the problem. Less >throughput>achieved per flow but it should balance itself out when you factor >in all the other flows. Plus no out-of-order packets.>>Justin
Hello, Unfortunately the point of this load balancing is to allow more throughput per flow. There is actually more throughput possible, however a good amount (20-30%) is not available with 2 paths, and almost no gain whatsoever is made when a third path is added. Is there no queueing mechanism that can mitigate this that we could apply on the interface facing us? C. Flav _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
