Don't make me pull the prior art lever...

Put down the trademarked words and nobody gets hurt.

-ASB: http://XeeSM.com/AndrewBaker


On Sat, Oct 17, 2009 at 8:31 PM, Andrew Levicki <[email protected]>wrote:

> Indeed®™!
>
> 2009/10/18 Sean Martin <[email protected]>
>
> I dont think that's necessarily true. If you were to do a cost-benefit
>> analysis between traditional 15k drives and SSDs (or EFDs), you might be
>> surprised what you find.
>>
>> Take an Exchange server that experiences an incredible amount of read IO
>> at the information store. You might need 10 traditional spindles to meet
>> demand, but only 2 SSDs. If the 15k drives cost $1000 a piece, and the SSDs
>> cost $5000 a piece, it might make more sense to go the SSD route.  Obviously
>> storage capacity needs to be taken into consideration as well.
>>
>> - Sean
>>
>> On Oct 17, 2009, at 3:48 PM, [email protected] wrote:
>>
>> We need prices to fall significantly to reap the benefits of the sizes we
>> need.
>>
>> Sent from my Verizon Wireless BlackBerry
>> ------------------------------
>> *From: * Andrew Levicki <[email protected]>
>> *Date: *Sat, 17 Oct 2009 19:12:41 +0100
>> *To: *NT System Admin Issues<[email protected]>
>> *Subject: *Re: "Why RAID 5 stops working in 2009 | Storage Bits |
>> ZDNet.com"
>>
>> In my opinion, we're on the cusp of seeing solid state storage becoming
>> the norm and we will be able to put hard drives out to pasture or use them
>> more for backups than tapes.
>> Although we have much faster hard disks nowadays than ever, it's amazing
>> that we are still at the behest of such a mechanical device for our mission
>> / business critical data. Solid state FTW.
>>
>> Regards,
>>
>> Andrew
>>
>> 2009/10/17 Angus Scott-Fleming < <[email protected]>[email protected]
>> >
>>
>>> Scaremongering, or legitimate things to worry about?  Lots of the
>>> "Talkback"
>>> comments are that ZDNet is over the top these days, but it seems to me
>>> he's got
>>> some legitimate points.
>>>
>>> ------- Included Stuff Follows -------
>>> Why RAID 5 stops working in 2009 | Storage Bits | <http://ZDNet.com>
>>> ZDNet.com
>>>
>>>  Disks fail
>>> ÿÿ While disks are incredibly reliable devices, they do fail. Our best
>>> data -
>>> ÿÿ from CMU and Google - finds that over 3% of drives fail each year in
>>> the
>>> ÿ   first three years of drive life, and then failure rates start rising
>>> fast.
>>>
>>> ÿ   With 7 brand new disks, you have ~20% chance of seeing a disk failure
>>> each
>>> ÿ   year. Factor in the rising failure rate with age and over 4 years you
>>> are
>>> ÿÿ almost certain to see a disk failure during the life of those disks.
>>>
>>> ÿÿ But yoÿ´re protected by RAID 5, right? Not in 2009.
>>>
>>> ÿ Reads fail
>>> ÿ   SATA drives are commonly specified with an unrecoverable read error
>>> rate
>>> ÿÿ ÿ (URE) of 10^14. Which means that once every 100,000,000,000,000
>>> bits, the
>>> ÿ   disk will very politely tell you that, so sorry, but I really, truly
>>> caÿ´t
>>> ÿ   read that sector back to you.
>>>
>>> ÿ   One hundred trillion bits is about 12 terabytes. Sound like a lot?
>>> Not in
>>> ÿÿ 2009.
>>>
>>>  Disk capacities double
>>> ÿÿ Disk drive capacities double every 18-24 months. We have 1 TB drives
>>> now,
>>> ÿÿ ÿ and in 2009 wÿ´ll have 2 TB drives.
>>>
>>> ÿÿ ÿ With a 7 drive RAID 5 disk failure, yÿÿ´ll have 6 remaining 2 TB
>>> drives.
>>> ÿÿ As the RAID controller is busily reading through those 6 disks to
>>> ÿÿ reconstruct the data from the failed drive, it is almost certain it
>>> will
>>> ÿ   see an URE.
>>>
>>> ÿ   So the read fails. And when that happens, you are one unhappy camper.
>>> The
>>> ÿÿ message "we cÿÿ´t read this RAID volume" travels up the chain of
>>> command
>>> ÿÿ ÿ until an error message is presented on the screen. 12 TB of your
>>> carefully
>>> ÿÿ ÿ protected - you thought! - data is gone. Oh, you didÿ´t back it up
>>> to
>>> ÿÿ tape? Bummer!
>>>
>>> --------- Included Stuff Ends ---------
>>> More here with links: <http://blogs.zdnet.com/storage/?p=162>
>>> http://blogs.zdnet.com/storage/?p=162
>>>
>>>
>>> --
>>> Angus Scott-Fleming
>>> GeoApps, Tucson, Arizona
>>> 1-520-290-5038
>>> +-----------------------------------+
>>>
>>

~ Finally, powerful endpoint security that ISN'T a resource hog! ~
~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/>  ~

Reply via email to