<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	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/"
		>
<channel>
	<title>Comments on: Bandwidth isn’t cheap. Disk isn’t cheap. CPU isn’t cheap.</title>
	<atom:link href="http:///2009/05/22/bandwidth-isnt-cheap-disk-isnt-cheap-cpu-isnt-cheap/feed/" rel="self" type="application/rss+xml" />
	<link>http://blog.palantirtech.com/2009/05/22/bandwidth-isnt-cheap-disk-isnt-cheap-cpu-isnt-cheap/</link>
	<description>Articles from the Engineering Group at Palantir Technologies</description>
	<lastBuildDate>Tue, 24 Jan 2012 09:51:18 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.2.1</generator>
	<item>
		<title>By: Chris</title>
		<link>http://blog.palantirtech.com/2009/05/22/bandwidth-isnt-cheap-disk-isnt-cheap-cpu-isnt-cheap/comment-page-1/#comment-238</link>
		<dc:creator>Chris</dc:creator>
		<pubDate>Thu, 01 Oct 2009 03:37:26 +0000</pubDate>
		<guid isPermaLink="false">http://blog.palantirtech.com/?p=961#comment-238</guid>
		<description>Have you evaluated consumer grade flash SSDs for this application?  You can pick up 256GB SSDs at something less than $700 per unit; you can probably get Intel&#039;s X25-M 160 GB SSDs for less than $500 per in bulk.  That&#039;s something between $2.8k and $3.2k per TB of SSD vs your figure of $5k/TB using disk.

It sounds like your data access is dominated by random reads.  The latency for individual reads on an SSD array will be in the low hundreds of microseconds as opposed to a few milliseconds for the disk array.  When reading lots of small, sparse records in a database you&#039;ll get at least an order of magnitude more IOPS, power efficiency should improve, and storage density should increase significantly.

Better, cheaper, faster, smaller?  What do you think?</description>
		<content:encoded><![CDATA[<p>Have you evaluated consumer grade flash SSDs for this application?  You can pick up 256GB SSDs at something less than $700 per unit; you can probably get Intel&#8217;s X25-M 160 GB SSDs for less than $500 per in bulk.  That&#8217;s something between $2.8k and $3.2k per TB of SSD vs your figure of $5k/TB using disk.</p>
<p>It sounds like your data access is dominated by random reads.  The latency for individual reads on an SSD array will be in the low hundreds of microseconds as opposed to a few milliseconds for the disk array.  When reading lots of small, sparse records in a database you&#8217;ll get at least an order of magnitude more IOPS, power efficiency should improve, and storage density should increase significantly.</p>
<p>Better, cheaper, faster, smaller?  What do you think?</p>
]]></content:encoded>
	</item>
</channel>
</rss>

