Comments on: Benchmarking Mail Relays and Forwarders http://doc.coker.com.au/papers/benchmarking-mail-relays-and-forwarders/ Documents that are not date based Sat, 15 Jan 2011 18:04:42 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Icann http://doc.coker.com.au/papers/benchmarking-mail-relays-and-forwarders/comment-page-1/#comment-1685 Icann Thu, 10 Jun 2010 00:20:41 +0000 http://doc.coker.com.au/?p=27#comment-1685 his means outbound smart host gateway systems, list servers, and forwarding services. his means outbound smart host gateway systems, list servers, and forwarding services.

]]> By: etbe http://doc.coker.com.au/papers/benchmarking-mail-relays-and-forwarders/comment-page-1/#comment-359 etbe Thu, 31 Jul 2008 06:22:59 +0000 http://doc.coker.com.au/?p=27#comment-359 http://etbe.coker.com.au/ Rob: I will write about some other Postal tests at my main blog (see the above URL). The test was not contrived to abuse DNS, it was designed to simulate one aspect of mailing list performance. http://etbe.coker.com.au/
Rob: I will write about some other Postal tests at my main blog (see the above URL).

The test was not contrived to abuse DNS, it was designed to simulate one aspect of mailing list performance.

]]> By: Rob http://doc.coker.com.au/papers/benchmarking-mail-relays-and-forwarders/comment-page-1/#comment-358 Rob Tue, 29 Jul 2008 13:01:21 +0000 http://doc.coker.com.au/?p=27#comment-358 This test seems contrived specifically to abuse DNS and test alias expansion performance, while avoiding the huge performance bottleneck of the queue process. In order to follow RFC, every file written to the queue is opened O_SYNC and brings with it horrible write performance. Try a 1:1 or a 1:3 relay test and watch performance tank. No message store delivery numbers? :-) This test seems contrived specifically to abuse DNS and test alias expansion performance, while avoiding the huge performance bottleneck of the queue process. In order to follow RFC, every file written to the queue is opened O_SYNC and brings with it horrible write performance. Try a 1:1 or a 1:3 relay test and watch performance tank.

No message store delivery numbers? :-)

]]>