<?xml version="1.0" encoding="UTF-8"?>
<!-- generator="FeedCreator 1.7.3" -->
<rss version="2.0">
	<channel>
		<title>Intro to Caching,Caching algorithms and caching frameworks part 5</title>
		<description>Comments for Intro to Caching,Caching algorithms and caching frameworks part 5 at http://www.jtraining.com , comment 1 to 2 out of 2 comments</description>
		<link>http://www.jtraining.com</link>
		<lastBuildDate>Thu, 11 Mar 2010 01:41:02 +0100</lastBuildDate>
        <generator>FeedCreator 1.7.3</generator>
		<item>
			<title>...</title>
			<link>http://www.jtraining.com/blogs/intro-to-caching-caching-algorithms-and-caching-frameworks-part-5.html#comment-67</link>
			<description>Hi Manik.

Thank for your comment, you are right i didnt test it with &quot;mvcc&quot; configuration, i will make sure that i run it using &quot;MVCC&quot; and i will take into my consideration also Infinispan and re-run my tests on the new version of CacheBenchnarkFWK

Thanks,
Ahmed - sneake75@hotmail.com</description>
			<pubDate>Fri, 12 Feb 2010 20:59:20 +0100</pubDate>
		</item>
		<item>
			<title>...</title>
			<link>http://www.jtraining.com/blogs/intro-to-caching-caching-algorithms-and-caching-frameworks-part-5.html#comment-66</link>
			<description>Hi there.  First, a short disclaimer: I'm the project lead of JBoss Cache, the founder and project lead of Infinispan, and the guy who wrote the first version of the CacheBenchmarkFramework which you used in your tests.  :)

A couple of comments on your extensive analysis:

1)  You mention JBoss Cache 3.x using MVCC as a default locking scheme (which is correct), but in your tests (the charts you have generated), you have JBoss Cache configured to use pessimistic locking which is [i][b]much[/b][/i] slower.  If you want to test defaults, you really should re-run your tests with MVCC rather than pessimistic locking.  This article I wrote contains some performance comparisons between an early beta of JBoss Cache 3.0 using MVCC versus pessimistic locking:  [url]http://jbosscache.blogspot.com/2008/09/naga-beta1-ready-to-rock-and-roll.html[/url]

2)  You really should be including Infinispan in your analysis.  See this article I wrote on Infinispan as a local-mode cache, which includes some benchmarks.  [url]http://infinispan.blogspot.com/2010/02/infinispan-as-local-cache.html[/url]

3)  The CacheBenchmarkFramework has evolved significantly as well.  It is no longer a part of the JBoss Cache source tree and is now its own standalone project on Sourceforge, and has been heavily revamped.  I encourage you to use this version instead.  [url]http://cachebenchfwk.sourceforge.net/[/url]

Cheers
Manik - a guest</description>
			<pubDate>Thu, 11 Feb 2010 14:39:12 +0100</pubDate>
		</item>
	</channel>
</rss>
