I love follow-up. Follow-up is good: study something, then study the follow-ups, and you will learn more.
So, a few follow-ups:
- I've been fascinated by the Air France Flight 447 investigation (background here and here), so if you were equally interested, don't miss this wonderful article in this month's Popular Mechanics: What Really Happened Aboard Air France 447:
Two years after the Airbus 330 plunged into the Atlantic Ocean, Air France 447's flight-data recorders finally turned up. The revelations from the pilot transcript paint a surprising picture of chaos in the cockpit, and confusion between the pilots that led to the crash.
- I've been following Jim Gettys and his studies into the network queueing phenomenon known as "BufferBloat". If you've been paying attention to this, too, you won't want to miss this discussion in the ACM Queue column. Meanwhile, Patrick McManus, who is hard at work on the new SPDY networking protocol, has an essay on the topic in which he notes some recent published research, and worries that there is still more research needed:
A classic HTTP/1.x flow is pretty short - giving it a signal to backoff doesn't save you much - it has sent much of what it needs to send already anyhow. Unless you drop almost all of that flow from your buffers you haven't achieved much. Further, a loss event has a high chance of damaging the flow more seriously than you intended - dropping a SYN or the last packet of the data train is a packet that will have very slow retry timers, and short flows are comprised of high percentages of these kinds of packets.
Understanding TCP's behaviors is certainly complicated; I recently wrote about this at some length on the Perforce blog.
Problems such as the complexity of modern systems such as those controlling airplanes, nuclear reactors, etc., or the unexpected inter-actions of networking equipment across the planet, continue to be extremely hard. Only dedicated study of many years or decades is going to bring progress, so I'm pleased to note such progress when it occurs!
No comments:
Post a Comment