[Seaside] Seaside scalability?
dale.henrichs at gemstone.com
Fri Sep 18 16:34:09 UTC 2009
You are right that the 350 request/second benchmark was done with custom seaside and application mods.
This post (http://gemstonesoup.wordpress.com/2007/10/19/scaling-seaside-with-gemstones/) provides information on benchmark runs of up to 230 requests/second on a 4 core machine with no custom mods to seaside or the application.
Still benchmarks are benchmarks.
The benchmarks do demonstrate the feasibility of taking a Seaside application that was developed to run in a single Squeak/Pharo image and deploying it (unmodified) in a system with multiple GemStone vms running across multiple cores (leveraging shared session state across vms).
----- "Philippe Marschall" <philippe.marschall at gmail.com> wrote:
| 2009/9/17 John Chludzinski <john.chludzinski at gmail.com>:
| > Does anyone have a measure of the scalability of Seaside? What are
| some of
| > the larger projects (web sites) hosted by Seaside? The
| powers-that-be (here
| > at work) are questioning this. They are still uncomfortable using
| > "exotic", i.e. Smalltalk as opposed to Java. ---John
| Are you sure they meant scalability and not something else like
| perfomance? Pretty much anything that doesn't need session
| for failover recovery scales more or less linearly. I doubt that's
| information they are looking for.
| If you need session replication GemStone is the only Smalltalk
| that will do. And take that 350 req/s with a grain of salt, that
| requires hacks into Seaside and your code that are by no means
| seaside mailing list
| seaside at lists.squeakfoundation.org
More information about the seaside