<?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"
	>
<channel>
	<title>Comments on: IBM 31.4 GB Solid State Drive Supported in all Blade Models</title>
	<atom:link href="http://www.bladevault.info/2008/03/05/ibm-314-gb-solid-state-drive-supported-in-all-blade-models/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.bladevault.info/2008/03/05/ibm-314-gb-solid-state-drive-supported-in-all-blade-models/</link>
	<description>Tips, Tricks, and Gotchas for IBM and HP Blades &#38; Servers, Virtualization Products, and Storage</description>
	<pubDate>Fri, 29 Aug 2008 00:38:50 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.5.1</generator>
		<item>
		<title>By: Aaron Delp</title>
		<link>http://www.bladevault.info/2008/03/05/ibm-314-gb-solid-state-drive-supported-in-all-blade-models/#comment-344</link>
		<dc:creator>Aaron Delp</dc:creator>
		<pubDate>Sat, 19 Jul 2008 11:45:49 +0000</pubDate>
		<guid isPermaLink="false">http://www.bladevault.info/2008/03/05/ibm-314-gb-solid-state-drive-supported-in-all-blade-models/#comment-344</guid>
		<description>Good Morning Jason and thank you for your comments!  Now that you mention it I remember sitting on a technical call when it was introduced.  The reason for the slow write performance is because of the technology used to actually write.  I don't remember the details but basically it had to do an erase of the sector before it did the write and then I believe it did a checksum after the fact.  Anyway, I remember IBM saying there were multiple operations when a write happens and yes, write would be slow with the first generation of the drives.

We were told at the time that this will improve in future generations.

Thank you for your comment!!</description>
		<content:encoded><![CDATA[<p>Good Morning Jason and thank you for your comments!  Now that you mention it I remember sitting on a technical call when it was introduced.  The reason for the slow write performance is because of the technology used to actually write.  I don&#8217;t remember the details but basically it had to do an erase of the sector before it did the write and then I believe it did a checksum after the fact.  Anyway, I remember IBM saying there were multiple operations when a write happens and yes, write would be slow with the first generation of the drives.</p>
<p>We were told at the time that this will improve in future generations.</p>
<p>Thank you for your comment!!</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jason Woods</title>
		<link>http://www.bladevault.info/2008/03/05/ibm-314-gb-solid-state-drive-supported-in-all-blade-models/#comment-343</link>
		<dc:creator>Jason Woods</dc:creator>
		<pubDate>Fri, 18 Jul 2008 20:42:34 +0000</pubDate>
		<guid isPermaLink="false">http://www.bladevault.info/2008/03/05/ibm-314-gb-solid-state-drive-supported-in-all-blade-models/#comment-343</guid>
		<description>Err, the 73 GB SAS drive read and wrote at close to 40 GB/min (not 20 GB).  Oops.</description>
		<content:encoded><![CDATA[<p>Err, the 73 GB SAS drive read and wrote at close to 40 GB/min (not 20 GB).  Oops.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jason Woods</title>
		<link>http://www.bladevault.info/2008/03/05/ibm-314-gb-solid-state-drive-supported-in-all-blade-models/#comment-342</link>
		<dc:creator>Jason Woods</dc:creator>
		<pubDate>Fri, 18 Jul 2008 20:41:20 +0000</pubDate>
		<guid isPermaLink="false">http://www.bladevault.info/2008/03/05/ibm-314-gb-solid-state-drive-supported-in-all-blade-models/#comment-342</guid>
		<description>I did some experimental work with the dual 16GB SAS SSD drive last fall on a HS21XM.

When putting them in RAID1, I was able to get 80+ GB/min read, but only 20 GB/min write.  With a single IBM 73GB SAS drive, I was able to read and write at around 20 GB/min.  With no RAID, and writing directly to a single SSD "disk", write speeds stayed around 20 GB/min, but the read speed dropped to 40 GB/min (obviously because of the lack of RAID).  Above performance data was from tests performed with sequential read/writes.  Random reads/writes drew close to the same number (about 20% loss in performance) on the SSD, but HDD speeds dropped considerably.

Basically, the 16 GB SSD drive pack was great at many small random reads, but any writing is slow as molasses.</description>
		<content:encoded><![CDATA[<p>I did some experimental work with the dual 16GB SAS SSD drive last fall on a HS21XM.</p>
<p>When putting them in RAID1, I was able to get 80+ GB/min read, but only 20 GB/min write.  With a single IBM 73GB SAS drive, I was able to read and write at around 20 GB/min.  With no RAID, and writing directly to a single SSD &#8220;disk&#8221;, write speeds stayed around 20 GB/min, but the read speed dropped to 40 GB/min (obviously because of the lack of RAID).  Above performance data was from tests performed with sequential read/writes.  Random reads/writes drew close to the same number (about 20% loss in performance) on the SSD, but HDD speeds dropped considerably.</p>
<p>Basically, the 16 GB SSD drive pack was great at many small random reads, but any writing is slow as molasses.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
