|
Steve Gonczi wrote:
> Have you guys given any thought to the desirable frequency of such
> load updates? This is esp. interesting in a high load situation...
> Assume you have 50 real servers, and you are getting 5000 requests per
> second.
> Also, assume your real servers can handle max. 1000 HTTP requests per
> second.
this sounds like an excercise in probability. Assuming the request packets
arrive randomly, and the average load/real-server is 100/sec (as stated),
what fraction of the time will servers be asked to handle 1000/sec?
The other part of the problem is that the limit of 1000 HTTP requests/sec
is not hard. The latency at the client end will just go up. Connections
will not be refused. So nothing catastrophic happens at 1000/sec.
Another part of the problem is that packets don't arrive randomly. The
majority of requests to places like CNN come at lunchtime on the US east-coast.
Nothing much is happening at 3am. Presumably depending on your clients,
you'll have a similar di-urnal fluctuation.
> This means your load distribution can significantly change in as short of a
> time
> as 0.1 seconds.
For extra credit, at a load of 100/sec, what fraction of the time
will the load double in 0.1 sec?
Joe
--
Joseph Mack PhD, Senior Systems Engineer, Lockheed Martin
contractor to the National Environmental Supercomputer Center,
mailto:mack.joseph@xxxxxxx ph# 919-541-0007, RTP, NC, USA
|
