<?xml version="1.0" encoding="UTF-8" ?>
<?xml-stylesheet type="text/xsl" href="http://blogs.technet.com/utility/FeedStylesheets/rss.xsl" media="screen"?><rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:slash="http://purl.org/rss/1.0/modules/slash/" xmlns:wfw="http://wellformedweb.org/CommentAPI/"><channel><title>Some more thoughts on disk IO and calculations….</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx</link><description>Thanks to everyone whose posted questions and comments to my previous blog post ! Since I’ve had so many questions on calculating IO, I thought I’d go into a bit more. First, there are many ways to calculate the disk IO per user. Measuring IO: What to</description><dc:language>en-US</dc:language><generator>CommunityServer 2.1 SP1 (Build: 61025.2)</generator><item><title>Exchange, XP</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#251820</link><pubDate>Wed, 03 Nov 2004 21:55:00 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:251820</guid><dc:creator>Windows Server Clustering </dc:creator><description /></item><item><title>re: Some more thoughts on disk IO and calculations….</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#255790</link><pubDate>Thu, 11 Nov 2004 18:16:00 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:255790</guid><dc:creator>David Wilhoit (kidego) Exchange MVP</dc:creator><description>How does a SAN and the write-back cache affect these numbers? Do I still need to allocate spindles according to I/O, or does the cache raise the number of I/Os per disk signifigantly?&lt;br&gt;&lt;br&gt;</description></item><item><title>re: Some more thoughts on disk IO and calculations….</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#257489</link><pubDate>Mon, 15 Nov 2004 11:57:00 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:257489</guid><dc:creator>Steve McGovern</dc:creator><description>Do you have metrics for write latency when using SAN synchronous replication for the log and database drives? </description></item><item><title>re: Some more thoughts on disk IO and calculations….</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#257910</link><pubDate>Tue, 16 Nov 2004 00:28:00 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:257910</guid><dc:creator>Nicole Allen</dc:creator><description>David -	The number of spindles you have backing your databases needs to be based on throughput (the IOs per second).  You should not depend on WriteBack Cache to reduce the numbers of IO.  Since the database writes are random, you can’t expect coalescing to significantly reduce the number of writes.  The advantage of the write back is that writes to the cache return the OS immediately – this reduces the write latency, but does not increase the overall throughput.  The data still needs to be written to disk.  That means if you overburden it, it will just fill the cache because the disks won't be able to keep up.  For normal IO on a healthy diskset, you should see writes around 10-15ms (factoring some queue delay, head seeking, and data transfer time).  With WriteBack caching, you'll see 1-5 ms write times.  In summary: write cache reduces the latencies on a well performing system, but does buy any reduction in the number of IOS.</description></item><item><title>re: Some more thoughts on disk IO and calculations….</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#265776</link><pubDate>Thu, 18 Nov 2004 18:00:00 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:265776</guid><dc:creator>Nicole Allen</dc:creator><description>David - I was unable to post a response to your previous question on the earlier blog because the blog is closed for comments...so I'm posting a reply here:&lt;br&gt;&lt;br&gt;Generally, read cache isn’t as helpful as writeback cache.  However, with the extent to which your disks are overused, that may not be sufficient.  Have you considered going to Raid10?   &lt;br&gt;-Nicole</description></item><item><title>re: Some more thoughts on disk IO and calculations….</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#265777</link><pubDate>Thu, 18 Nov 2004 18:01:00 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:265777</guid><dc:creator>Nicole Allen</dc:creator><description>Sorry, we don't have anything ready for publication at this time.&lt;br&gt;</description></item><item><title>Exchange, XP</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#267022</link><pubDate>Sat, 20 Nov 2004 01:00:00 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:267022</guid><dc:creator>Windows Server Clustering </dc:creator><description /></item><item><title>MSDTC and Exchange clusters </title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#270366</link><pubDate>Thu, 25 Nov 2004 19:43:00 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:270366</guid><dc:creator>Praveen Kumar's Blog</dc:creator><description /></item><item><title>MSDTC and Exchange clusters</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#354508</link><pubDate>Mon, 17 Jan 2005 21:25:00 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:354508</guid><dc:creator>Evan's WebLog</dc:creator><description /></item><item><title>&amp;amp;quot;Requesting data from the Exchange server...&amp;amp;quot; or &amp;amp;quot;Outlook is trying to retrieve data…&amp;amp;quot; client messages</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#405356</link><pubDate>Wed, 25 May 2005 20:59:44 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:405356</guid><dc:creator>You Had Me At EHLO...</dc:creator><description>Based on the questions that we got on another post, it seemed appropriate to address the &amp;amp;quot;Requesting...</description></item><item><title>'Requesting data from the Exchange server...' or 'Outlook is trying to retrieve data…' client messages</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#405362</link><pubDate>Wed, 25 May 2005 21:00:12 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:405362</guid><dc:creator>You Had Me At EHLO...</dc:creator><description>Based on the questions that we got on another post, it seemed appropriate to address the &amp;amp;quot;Requesting...</description></item><item><title>'Requesting data from the Exchange server...' or 'Outlook is trying to retrieve data…' client messages</title><link>http://blogs.technet.com/exchange/archive/2004/11/03/251743.aspx#405368</link><pubDate>Wed, 25 May 2005 21:01:38 GMT</pubDate><guid isPermaLink="false">d5e57398-b9ef-4490-9955-07cbb4e4a80d:405368</guid><dc:creator>You Had Me At EHLO...</dc:creator><description>Based on the questions that we got on another post, it seemed appropriate to address the &amp;amp;quot;Requesting...</description></item></channel></rss>