<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Re: SAS 6G vs. 3G for controllers/disks in Disk</title>
    <link>https://community.hpe.com/t5/disk/sas-6g-vs-3g-for-controllers-disks/m-p/4582581#M7461</link>
    <description>When looking at how long a disk I/O takes you have the time to seek but also the time to transfer the data from the disk to the controller.  3 Gb/s gives you a maximum theoretical transfer rate of ~300 MB/s.  a 6Gb/s link should double that to ~600 MB/s&lt;BR /&gt;&lt;BR /&gt;If you are doing I/Os with large transfer size say 1 MB (MegaByte) then the time it takes for one transfer is 3.3 ms at 3 Gb/s and 1.7 ms at 6 Gb/s. Assuming an average seek time of 3.5 ms then an I/O would take about 6.8 for 3 Gb and 5.2 ms for 6 Gb drives and controllers.  This timing is greatly simplified.  This give you a I/O rate of 147 I/Os per second for the 3 Gb and 192 I/Os per second for the 6 Gb drive and controller.  &lt;BR /&gt;&lt;BR /&gt;Of course if your data size is much smaller then the difference would be smaller.  Looking at an 8K transfer size we get an I/O rate of 283 and 284 I/O per second for 3 Gb and 6 Gb respectively.&lt;BR /&gt;&lt;BR /&gt;With RAID5 you also need to look at time to rebuild the raid set after a failure.  While the raid set is rebuilding you can experience significant performance reductions.  Again a faster transfer time will reduce this time.&lt;BR /&gt;&lt;BR /&gt;Hope this helps</description>
    <pubDate>Fri, 19 Feb 2010 07:11:57 GMT</pubDate>
    <dc:creator>Cass Witkowski</dc:creator>
    <dc:date>2010-02-19T07:11:57Z</dc:date>
    <item>
      <title>SAS 6G vs. 3G for controllers/disks</title>
      <link>https://community.hpe.com/t5/disk/sas-6g-vs-3g-for-controllers-disks/m-p/4582580#M7460</link>
      <description>&lt;!--!*#--&gt;Hello Disk Experts!&lt;BR /&gt;&lt;BR /&gt;A few questions for you regarding SAS 6G vs. 3G.  I'm using a RAID 5 configuration, so please keep this in mind for the below.&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;From what I understand, a spinning hard drive (vs. SSD) will never hit anywhere near the limit for 3G speed for sustained read/writes.&lt;BR /&gt;&lt;BR /&gt;1) Is there any reason to get SAS 6G spinning hard drives, and if so, in what situations and why?&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;From what I can gather, and these are rough numbers from Seagate's Cheetah drives, a SAS 10K drive can push out a sustained 88MB/sec/drive, and a SAS 15K drive can do about 120MB/sec/drive.&lt;BR /&gt;&lt;BR /&gt;If a 3G controller can maintain it's theoretical 3G throughput (doubtful), then I would need at least four SAS 10K drives or three 15K SAS drives before I could possibly see any real benefit to 6G.&lt;BR /&gt;&lt;BR /&gt;2) Is this logic flawed, and if so, how?&lt;BR /&gt;&lt;BR /&gt;3) What is the maximum sustained throughput on a 6G controller, e.g., the HP P410 (462862-B21)?&lt;BR /&gt;&lt;BR /&gt;4) In what situations does it make sense to use a 6G controller with 3G drives?&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;Many thanks!</description>
      <pubDate>Thu, 11 Feb 2010 18:47:18 GMT</pubDate>
      <guid>https://community.hpe.com/t5/disk/sas-6g-vs-3g-for-controllers-disks/m-p/4582580#M7460</guid>
      <dc:creator>Cory_K</dc:creator>
      <dc:date>2010-02-11T18:47:18Z</dc:date>
    </item>
    <item>
      <title>Re: SAS 6G vs. 3G for controllers/disks</title>
      <link>https://community.hpe.com/t5/disk/sas-6g-vs-3g-for-controllers-disks/m-p/4582581#M7461</link>
      <description>When looking at how long a disk I/O takes you have the time to seek but also the time to transfer the data from the disk to the controller.  3 Gb/s gives you a maximum theoretical transfer rate of ~300 MB/s.  a 6Gb/s link should double that to ~600 MB/s&lt;BR /&gt;&lt;BR /&gt;If you are doing I/Os with large transfer size say 1 MB (MegaByte) then the time it takes for one transfer is 3.3 ms at 3 Gb/s and 1.7 ms at 6 Gb/s. Assuming an average seek time of 3.5 ms then an I/O would take about 6.8 for 3 Gb and 5.2 ms for 6 Gb drives and controllers.  This timing is greatly simplified.  This give you a I/O rate of 147 I/Os per second for the 3 Gb and 192 I/Os per second for the 6 Gb drive and controller.  &lt;BR /&gt;&lt;BR /&gt;Of course if your data size is much smaller then the difference would be smaller.  Looking at an 8K transfer size we get an I/O rate of 283 and 284 I/O per second for 3 Gb and 6 Gb respectively.&lt;BR /&gt;&lt;BR /&gt;With RAID5 you also need to look at time to rebuild the raid set after a failure.  While the raid set is rebuilding you can experience significant performance reductions.  Again a faster transfer time will reduce this time.&lt;BR /&gt;&lt;BR /&gt;Hope this helps</description>
      <pubDate>Fri, 19 Feb 2010 07:11:57 GMT</pubDate>
      <guid>https://community.hpe.com/t5/disk/sas-6g-vs-3g-for-controllers-disks/m-p/4582581#M7461</guid>
      <dc:creator>Cass Witkowski</dc:creator>
      <dc:date>2010-02-19T07:11:57Z</dc:date>
    </item>
  </channel>
</rss>

