On 5 mei 2008, at 0:57, Adrian Chadd wrote:
I'd seriously be looking at making current -software- run more efficiently before counting ipv6-related power savings.
Good luck with that. Obviously there is a lot to be gained at that end, but that doesn't mean we should ignore power use in the network. One thing that could help here is to increase the average packet size. Whenever I've looked, this has always hovered around 500 bytes for internet traffic. If we can get jumboframes widely deployed, it should be doable to double that. Since most work in routers and switches is per-packet rather than per-bit, this has the potential to save a good amount of power. Now obviously this only works in practice if routers and switches actually use less power when there are fewer packets, which is not a given. It helps even more if the maximum throughput isn't based on 64- byte packets. Why do people demand that, anyway? The only thing I can think of is DoS attacks. But that can be solved by only allowing end- users to send an average packet size of 500 (or 250, or whatever) bytes. So if you have a 10 Mbps connection you don't get to send 14000 64-byte packets per second, but a maximum of 2500 packets per second. So with 64-byte packets you only get to use 1.25 Mbps. I'm guessing having a 4x10Gbps line card that "only" does 14 Mpps total rather than 14 Mpps per port would be a good deal cheaper. Obviously if you're a service provider with a customer that sends 10 Gbps worth of VoIP you can only use one of those 4 ports but somehow, I'm thinking few people use 10 Gbps worth of VoIP... Iljitsch PS. Am I the only one who is annoyed by the reduction in usable subject space by the superfluous [NANOG]?