>> Brad Knowles wrote faster than he thought: >>> True anycast for protocols other than UDP? I sincerely doubt it. >>> If you have evidence for this, I'd love to see it. >> Works fine for TCP and ICMP. What protocols do you have doubts about? > I have doubts about TCP. Can you explain in more detail how it works for > TCP, especially for the case where the route from IP address A to IP > address B changes to a different machine that serves IP address B, while > in the middle of a connection? And how often do route changes occur?
in actual use, we have been surprised at how well anycast tcp works within a large isp. like O(10^5) users for multiple hours. most routing instability has a relatively small number of real causes. randy #---------------------------------------------------------------------- # To unsubscribe, send a message to <[EMAIL PROTECTED]>.
