Messages in this thread | | | From | "Stephen C. Tweedie" <> | Date | Wed, 17 Nov 1999 00:03:37 +0000 (GMT) | Subject | RE: Max tcp connections |
| |
Hi,
On Fri, 12 Nov 1999 12:28:26 -0800, "David Schwartz" <davids@webmaster.com> said:
> But the thing is, if events fall back to poll under high load, > how can they be more scalable than poll? That strikes me as an absurd > argument.
Think about it --- the amount of work a server can do is bounded by the cost of doing each unit of work. That cost is O(1) for the event model and O(n) for poll.
The event model only ever uses poll if the event queue overflows, which is expected to be a rare occurrence (it *never* happens during heavy local-load benchmarking with phhttpd).
--Stephen
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |