<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
	>

<channel>
	<title>NiftyName &#187; compare to others and try</title>
	<atom:link href="http://www.niftyname.org/category/compare-to-others-and-try/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.niftyname.org</link>
	<description>Open Source Hardware As A Service</description>
	<lastBuildDate>Tue, 10 Nov 2009 17:59:45 +0000</lastBuildDate>
	<language>en</language>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.0.4</generator>
		<item>
		<title>Storage I/O benchmark</title>
		<link>http://www.niftyname.org/2009/05/storage-io-benchmark/</link>
		<comments>http://www.niftyname.org/2009/05/storage-io-benchmark/#comments</comments>
		<pubDate>Wed, 27 May 2009 15:47:36 +0000</pubDate>
		<dc:creator>admin</dc:creator>
				<category><![CDATA[compare to others and try]]></category>
		<category><![CDATA[amazon]]></category>
		<category><![CDATA[aws]]></category>
		<category><![CDATA[benchmark]]></category>
		<category><![CDATA[ec2]]></category>
		<category><![CDATA[flexiscale]]></category>
		<category><![CDATA[gogrid]]></category>
		<category><![CDATA[matrix]]></category>
		<category><![CDATA[storage]]></category>
		<category><![CDATA[sysbench]]></category>

		<guid isPermaLink="false">http://www.niftyname.org/?p=344</guid>
		<description><![CDATA[4 cloud computing storage performance compared]]></description>
			<content:encoded><![CDATA[<p><script type="text/javascript"><!--
jQuery(document).ready(function(){
jQuery(".vertgraph").each(
     function(i,elt) {
         maxref=478;
         max=0;
         val=jQuery(elt).children().children().contents();
         jQuery(val).each(
            function(i,v)
            {
                if(parseFloat(String(v.data))>max)
                    max=parseFloat(v.data);
            }
         );
         ratio = (maxref-100)/max;
         jQuery(elt).children().children().each(
                function(i, elt){
                    h = parseFloat(jQuery(elt).contents()[0].data);
                    jQuery(elt).height((12+(h*ratio))+"px");
                }
          );
     }
);
});
// --></script></p>
<h3>Introduction</h3>
<p>We performed all the tests with the multi-threaded benchmark tool sysbench (version 0.4.8).</p>
<p>
<a  href="http://sysbench.sourceforge.net/" onclick="javascript:pageTracker._trackPageview('/external/sysbench.sourceforge.net/');" >http://sysbench.sourceforge.net/</a></p>
<p>This tool was primarily written for MySQL benchs and is very I/O intensive. Nowadays it&#8217;s one of the most appreciated tool for general I/O benchs.</p>
<h4>Read bench commands</h4>
<p>sysbench &#8211;test=fileio &#8211;max-time=60 &#8211;max-requests=100000000  &#8211;file-num=1 &#8211;file-extra-flags=direct &#8211;file-fsync-freq=0  &#8211;file-total-size=128M &#8211;file-test-mode=rndrd cleanup</p>
<p>sysbench &#8211;test=fileio &#8211;max-time=60 &#8211;max-requests=100000000  &#8211;file-num=1 &#8211;file-extra-flags=direct &#8211;file-fsync-freq=0  &#8211;file-total-size=128M &#8211;file-test-mode=rndrd prepare</p>
<p>sysbench &#8211;test=fileio &#8211;max-time=60 &#8211;max-requests=100000000  &#8211;file-num=1 &#8211;file-extra-flags=direct &#8211;file-fsync-freq=0  &#8211;file-total-size=128M &#8211;file-test-mode=rndrd run</p>
<h4>Write bench commands</h4>
<p>sysbench &#8211;test=fileio &#8211;max-time=60 &#8211;max-requests=100000000  &#8211;file-num=1 &#8211;file-extra-flags=direct &#8211;file-fsync-freq=0  &#8211;file-total-size=128M &#8211;file-test-mode=rndwr cleanup</p>
<p>sysbench &#8211;test=fileio &#8211;max-time=60 &#8211;max-requests=100000000  &#8211;file-num=1 &#8211;file-extra-flags=direct &#8211;file-fsync-freq=0  &#8211;file-total-size=128M &#8211;file-test-mode=rndwr prepare</p>
<p>sysbench &#8211;test=fileio &#8211;max-time=60 &#8211;max-requests=100000000  &#8211;file-num=1 &#8211;file-extra-flags=direct &#8211;file-fsync-freq=0  &#8211;file-total-size=128M &#8211;file-test-mode=rndwr run</p>
<h3>Forces in presence</h3>
<p>The goal is to be the most objective as possible, so we choose the closest configuration on each instance provider.</p>
<h4>Flexiscale</h4>
<p>ubuntu amd64 8.04 LTS</p>
<ul>
<li>1 GiB RAM</li>
<li>20 GiB of storage</li>
</ul>
<h4>Gogrid</h4>
<p>Centos amd64 5.1</p>
<ul>
<li>1 GiB RAM</li>
<li>database-server instance</li>
</ul>
<h4>EC2</h4>
<p>ubuntu amd64 9.04</p>
<ul>
<li>small instance</li>
<li>1.7 GiB RAM</li>
</ul>
<h4>NiftyName</h4>
<p>ubuntu amd64 9.04</p>
<ul>
<li>1 GiB RAM</li>
<li>50 GiB of storage</li>
</ul>
<h3>Read</h3>
<h4>Transfer MB/sec : we are twice fast as the fastest competitor !</h4>
<div class="vertgraph">
<ul>
<li class="flexiscale">0.9</li>
<li class="gogrid">88.9</li>
<li class="ec2">2.5</li>
<li class="niftyname">178.3</li>
</ul>
</div>
<blockquote><p>EC2, which is the most popular offer, suffers of very poor read performances.</p></blockquote>
<h4>I/O requests per second : it matchs read transfer values</h4>
<div class="vertgraph">
<ul>
<li class="flexiscale">54.4</li>
<li class="gogrid">5693.2</li>
<li class="ec2">160.0</li>
<li class="niftyname">11414.4</li>
</ul>
</div>
<h3>Write</h3>
<h4>Transfer MB/sec : EC2 is a (slighty) more serious competitor</h4>
<div class="vertgraph">
<ul>
<li class="flexiscale">0.9</li>
<li class="gogrid">0.7</li>
<li class="ec2">3.7</li>
<li class="niftyname">24.1</li>
</ul>
</div>
<h4>I/O requests/sec</h4>
<div class="vertgraph">
<ul>
<li class="flexiscale">56.4</li>
<li class="gogrid">42.9</li>
<li class="ec2">239.3</li>
<li class="niftyname">1541.6</li>
</ul>
</div>
<h3>Conclusion</h3>
<p>NiftyName supremacy is undeniable. NiftyName does not forget performances against features. What is done is well done, we are trying to reach the highest quality level, it&#8217;s significant for a common production environment.</p>
]]></content:encoded>
			<wfw:commentRss>http://www.niftyname.org/2009/05/storage-io-benchmark/feed/</wfw:commentRss>
		<slash:comments>0</slash:comments>
		</item>
	</channel>
</rss>

