On Jun 21, 2010, at 23:55, Joe Abley wrote: Everyone: Thanks for the replies regarding the /24 announcement from a "/20 allocated block". Yes, obviously the /20 announcement will handle the traffic, too. I'm a regular reader on NANOG and consistently impressed by the expertise on display and the speed with which it's generously handed out. :-)
> I'm interested in the idea of anycasting one of the pool.ntp.org > herd-members. Every time I've suggested such a thing I've been told > (paraphrasing) that a good (server, client) NTP session exhibits reasonable > RTT stability, this constitutes, in effect, a long-lived transaction, and > hence anycast is not a good answer unless you have confidence that the > potential for oscillations is low, or that the frequency of the oscillations > is very low (i.e. in a private network this might be a good answer, but > across the public Internet it's a poor answer). > > Has the thinking changed, or did I just misunderstand? I think the thinking on NTP [ see below ] is the same; but indeed when I wrote "possibly other UDP based services" experimenting with that was my idea, too. I believe some of the CDNs are anycast based (Cachefly?) and they did some extensive tests with very long http transactions. (And I guess do a big test daily in running the service...). However -- Much of the pool.ntp.org traffic is from SNTP clients where the NTP considerations don't apply. (In summary: SNTP = dumb client that just asks for the time now; NTP = clever server that keeps track of the time. The protocol is the same, but the usage quite different). - ask
