> I think that's the catch phrase... "open" meaning, not blocked. So don't > block p2p or any other traffic, just throttle it down... WAY down...
I gave a talk about doing this with Linux + HTB a couple of years ago. I had our head-end traffic shaper doing classful queuing, giving each type of traffic a priority level, an average bandwidth level and a maximum that it could borrow from the other classes if they weren't doing anything. The borrowing concept is nice, allowing the majority of surfing/emailing customers to get what they need/want during peak hours, and the underground junk works slightly better after hours. Oh, and all the "good" P2P still works. VoIP, VPN/RDP/SSH and other latency-sensitive items got highest priority and a decent bandwidth allocation. The office got the next priority down (includes telemetry and other admin traffic) and plenty of bandwidth. Meat and potatoes apps like web surfing and email got a middle to lower priority with a big chunk of bandwidth. Unknown traffic got a low priority with some bandwidth but the option to borrow from others so that I didn't break anything. Known P2P traffic got the lowest priority and bandwidth allocation, but had enough that it wouldn't totally stop. Didn't get any complaints. This was very critical when we were bumping our heads on our DS3's, but doesn't buy me much with Gig-E circuits. So instead we monitor individual AP sites occasionally for heavy use. Just like everyone else we usually see one or two heavy users pop up every once in a while. And 9 times out of 10 it's a teenager. -- Bryan -------------------------------------------------------------------------------- WISPA Wants You! Join today! http://signup.wispa.org/ -------------------------------------------------------------------------------- WISPA Wireless List: wireless@wispa.org Subscribe/Unsubscribe: http://lists.wispa.org/mailman/listinfo/wireless Archives: http://lists.wispa.org/pipermail/wireless/
