News for hackers. Stuff that matters.
Re:TCP Over IP Anycast – Pipe Dream or Reality? by paulsutter on Friday November 27, 2015 @04:09AM (#10636547)
We had the same experience at Quantcast. Better results with anycast within continents, and DNS to distribute a per-continent anycast addresses.

Some know-it-all busybodies like to proclaim that anycast can't work for TCP because the route can change during the session, but the reality is that:

- route changes are infrequent

- individual sessions are typically much shorter than route change intervals

- we never experienced proposed problems with sustained route flapping

- route changes are readily detected and handled as (rare) dead connections. The nicest bit is that transferred live connections can get rejected and closed immediately by the new server allowing for rapid recovery (unlike DNS changes which leaves things hanging for a long time).

We had extremely fast datacenter failover using anycast (100% of traffic moved as soon as 5 seconds). Contrasted to the minutes-to-hours nightmare delays that characterize DNS failover. (The DNS time-to-live is an advisory at best, and outright ignored by many clients)

It's too bad more hosting providers don't allow route announcements.

Re:TCP Over IP Anycast – Pipe Dream or Reality? by sspies on Friday November 27, 2015 @04:12AM (#10636555)
I can fully agree with your experience using DNS failover vs anycast failover. Made a video of how I implemented anycast https://www.youtube.com/watch?v=tsXpQHi7Udo
Re:TCP Over IP Anycast – Pipe Dream or Reality? by _ao789 on Friday November 27, 2015 @07:09AM (#10636931)
Interesting video, thanks for sharing
Re:TCP Over IP Anycast – Pipe Dream or Reality? by dsl on Friday November 27, 2015 @08:42AM (#10637246)
Same here. The only people dismissing TCP anycast at this point are folks who haven't been running anycast deployments.

It also helps to respond to DNS with multiple VIPs within your anycast /24. Browser retry logic is pretty nifty and solves most of the real world problems.

Re:TCP Over IP Anycast – Pipe Dream or Reality? by federico3 on Friday November 27, 2015 @08:30AM (#10637203)
> individual sessions are typically much shorter than route change intervals

True for web traffic, not true for large file transfers, many multiplayer games, VPNs.

> live connections can get rejected and closed immediately by the new server allowing for rapid recovery

...if the client protocol allows recovery.

Re:TCP Over IP Anycast – Pipe Dream or Reality? by riteshm on Wednesday December 02, 2015 @01:23PM (#10664758)
Glad to hear that Paul! This is Ritesh, author of the LinkedIn article.

Lets chat offline. I am curious to see if you guys faced some issues that we are seeing.