Open-E logo
0 Liked

    What you can expect from SSD?

    Hard disk performance is limited by mechanical parts velocity and there is no solution for this issue.

    Servers are using RAID for redundancy and for performance. Thanks to RAID technology it is possible to scale single hard disk performance.

    The team of disks works faster, but applications are needing increased improvements as over periods technical evolutions.

    There are storage offers using DDR RAM or SLC Flash that claim 2-20 GB/s throughput or 100K to millions of random IO/sec

    But the cost of such systems are extremely high.

    Recently RAID controller vendors like LSI and Adaptec offer SSD cache options. If the applications require higher random IO/sec such solutions are worth looking into.

    Adaptec claims 8 times faster IO than HDD-only arrays.

    LSI claims even 50 times faster IO. They have simulated web servers that re-read hot spot activity. With CacheCade enabled reached at 14,896 IO/sec and compared to 273 IO/sec with disabled CacheCade.

    Such SATA/SSD hybrid can significantly  improve applications with random IO patterns. They work with single SSD as a minimum , the solution is relatively inexpensive.

    Adaptec maxCache™ 64GB SSD Cache Performance Kit SRP = $1,795

    And  LSI CacheCade software pack has a suggested retail price of $270 plus X25-E Extreme 64GB cost $700 makes about $1,000.

    Very interesting will be to know what kind of random IO/ps can be reached if we use SSD only.  We have created RAID 5 with 4 * INTEL X25-M SATA SSD. The RAID 5 was fully initialized then the IOmeter test was started.

    Results below:

    RAID 5 with 4* INTEL X25-M SATA SSD 120GB

    Transfer Request SizePercent SequentialPercent RandomPercent WritePercent ReadMB / secondIO / second
    2KB01001000178707
    2KB010001004020266
    4KB01001000297259
    4KB010001005614266
    64KB01001000991580
    64KB010001002203524

    In order to be in a position to appreciate the above results we have run the same test pattern with RAID 5 with 12 * HDD SAS 15K (SEAGATE ST373455SS )

    Results below:

    RAID 5 with 12* SAS 15K

    Transfer Request SizePercent SequentialPercent RandomPercent WritePercent ReadMB / secondIO / second
    2KB010010004.42300
    2KB010001004.82470
    4KB010010008.42220
    4KB01000100102630
    64KB01001000721140
    64KB01000100981570

    So:

    Random Write IO/sec with 2k block is3.8times better
    Random Read IO/sec with 2k block is8.2times better
    Random Write IO/sec with 4k block is3.3times better
    Random Read IO/sec with 4k block is5.4times better

    and:

    Random Write MB/sec with 2k block is3.9times better
    Random Read MB/sec with 2k block is8.3times better
    Random Write MB/sec with 4k block is3.5times better
    Random Read MB/sec with 4k block is5.6times better

    And the comparison was not quit fair because the HDD array has 12 drives and SSD array only 4.

    We are using cheaper (MLC) non-enterprise version of the SSD. The main problem with MLC SSD is the write endurance which is about 11TB for 120GB drive and 15TB for 160GB.

    If drives are in a RAID 5 the write endurance can be multiplied by n-1. Also there is option to leave some spare space for the Wear Leveling mechanism. For example 10% reserved space expands write endurance by 2.8 times.

    So 4 * 160GB SSD in RAID 5 with 10% reserved space = 3 (n-1)  * 2.8 * 15 TB = 126 TB write endurance. If your application is not needing heavy duty write orientations then you can consider it for production.

    If the application is heavy duty write oriented you need to consider Enterprise SSD. Here is the write endurance comparison:

    ModelEndurancePrice
    INTEL X25-M SATA SSD 120GB15 TB$400
    INTEL X25-E Extreme 32GB1 PB (1024 TB)
    $370
    INTEL X25-E Extreme 64GB2 PB ( 2048 TB)
    $720
    WD SiliconDrive N1x 128 GB701 GB/day$1500

    How much is 15TB. Sustain write 24/7 with 20MB/sec, so 15TB will be written in about 9 days.

    How much is 2PB. Sustain write 24/7 with 20MB/sec, so 2PB will be written in about 3.4 years.

    The RAID 5 array with 4 * Intel X25-E 64GB drives will cost $2880 and is about $15 per GB.

    The RAID 5 array with 4 * Intel X25-E 120GB drives will cost $1600and is about  $4.4 per GB.

    In comparison to 15k SAS HDD, the RAID 5 array with 4 * Seagate 148GB (15k) makes 4 * $300 = $1200  and is about $1.3 per GB.

    This shows about 12 times more expensive but at least 3 to 8 times faster with random IO.

    We did not test the enterprise SSD but mainstream SSD only. The random IO with enterprise SDD would be even better.

    Rating: / 5.

    No votes yet

    7 Comments

    • Martin Rozanski

      January 19, 01 2011 03:22:07

      Very interesting , good read about endurance can be found here – http://www.storagesearch.com/ssdmyths-endurance.html
      BTW – what about 1M IOPS – http://www.fusionio.com/products/iodriveoctal , don’t want to know the price 🙂

      VA:F [1.9.22_1171]
      Rating: 0 (from 0 votes)
    • Paul Snellock

      March 17, 03 2011 03:15:12

      Intresting article but has a few confusing parts which make it difficult to decide how best to implement somthing like this. On the one hand is states
      Such SATA/SSD hybrid can significantly improve applications with random IO patterns.
      Then in your testing you get a similar result but state
      In order to be in a position to appreciate the above results we have run the same test pattern with RAID 5 with 12 * HDD SAS 15K (SEAGATE ST373455SS )

      So to get the sort of result you are getting can I use lower cost much larger drives which reduces footprint and cost which accoring to the article works or do I need more higher quality drives to get the result

      Until this is clear you could end up spending much more than you need to

      can sombody comment on this so I can work out how best to put this into practice

      Thanks

      Paul Snellock

      VA:F [1.9.22_1171]
      Rating: 0 (from 0 votes)
    • Sean Leyne

      October 12, 10 2011 09:17:05

      There is a typo in the article:

      The RAID 5 array with 4 * Intel **X25-E** 120GB drives will cost $1600and is about $4.4 per GB.

      should read

      The RAID 5 array with 4 * Intel **X25-M** 120GB drives will cost $1600and is about $4.4 per GB.

      VA:F [1.9.22_1171]
      Rating: 0 (from 0 votes)
    • Michael Smith

      February 16, 02 2012 01:58:16

      Have you done any performance testing with a SAN based VMware server environment, or are these SSD cached products geared more toward enterprise application storage for high intensive applications?

      We are looking to squeeze more performance out of our datacenter environment. We run multiple VM host machines connected to our SANs. I was wondering if we would be able to use this type of a solution, or if SSD caching is not suited toward hosting virtual server environments.

      VA:F [1.9.22_1171]
      Rating: 0 (from 0 votes)
      • Jason

        March 22, 03 2012 09:02:09

        It is a storage solution, so you can use it in any way that a traditional HDD can be used, therefore the VMWare storage would be able to use it. As this is an Open-E article, I assume these SSD’s are being used in an Open-E environment?
        The limiting factor as far as I can see is that the SSD has an endurance factor. As the article states, if you write 15Tb of data under the conditions they have shown then it would take 9 days for the SSD to wear out. If you want better endurance then you look at better spec’ed SSD’s.
        If you look at your actual data write profile for your business then you should be able to see how long the SSD should last in your environment. Of course it is RAID solution so if one SSD fails you replace it just like you do with any failed HDD in the RAID.
        Commonly I am seeing organisations use SSD for the Virtual Desktop delivery where very fast access results in short startup times for the end user. There is not as much write activity to the SSD as most of the write activity is happening on servers not the virtual desktop.

        VA:F [1.9.22_1171]
        Rating: 0 (from 0 votes)
      • Bunty

        July 25, 07 2012 09:36:22

        I have been using SSD drives in my laptpos and desktops for years. I must say that I will never go back to an IDE/SATA/SCSI drive again except for mass storage. Speed is excellent and I have used multiple brands for personal testing. Money is equal to time and I work to much already.

        VA:F [1.9.22_1171]
        Rating: 0 (from 0 votes)
    • Aadam

      February 12, 02 2018 10:15:26

      SSD is quite expensive compared to hard disk but in the coming generations, SSD will be quite common.
      Anyways, Good explanation though.

      VA:F [1.9.22_1171]
      Rating: 0 (from 0 votes)

    Leave a Reply to Michael Smith

    Cancel reply