Hard Drive Recovery Group | RAID 5 Is Not Doomed, Despite Reports

As recently as October of 2008, a story on the newswires predicting the failure (pardon the pun) of RAID 5 as an effective data storage system has caused some disturbance in the IT community, if only because of its somewhat opinionated nature. The story, which ran on ZDNet as well as some of the other main IT industry rags, noted that SATA hard drive capacity, now reaching above the 2 terabyte level, could be a drag on future RAID 5 growth, and, and here's where the controversy comes in, actually INCREASE the potential for read errors during disk construction.

In the article, the ZdNet columnist notes that SATA drives can suffer from unrecoverable read rates of 10^14, which extrapolates that these disk drives are unable to read a sector once every 100 trillion bits read. As many familiar with the data storage industry have noted, drives are already well above the two terabyte capacity level, which with this logic in mind, makes it all but inevitable when attempting disk reconstruction from a seven drive RAID 5 system failure.

Where things begin to get shifty is where the article claims that during recovery, the array volume is likely to declare itself unreadable, thus stopping the RAID recovery or self-repair. From this point, the article also notes that once this event occurs during the recovery process, it is more than likely all 12 terabytes stored on this hypothetical RAID system will be lost. Or in the least, require professional RAID data recovery services.

A Dead RAID On Every Corner?

While certainly what the ZDNet article boasts would seem to be a huge boon for the RAID recovery industry ("Imagine a world with dead RAIDs on every corner!"), we believe this to be far from the case. The writer then goes on to throw in several basic premises about hdd technologies, such as the fact that mechanical parts tend to fail, tight data density can also increase failure rates, and that RAID 6 could be an expensive, but effective solution to this perceived issue.

Click here to get a RAID recovery quote.

The real problem with the article is actually quite simple. What the article does note is the "soft error rate," which is the rate at which data could possibly be misread incorrectly a single time, but NOT multiple times. Therefore, in everyday RAID storage system use, this number is really useless. Because of ECC encoding (Error Correcting Code) in most hard disk drives, the likelihood of failing to record more than a few bits and being unable to correct from it is almost impossible. Using the soft error rate is therefore a poor rate with which to base a story on.

What the story should have been based upon is sector size and amount of ECC per hard disk sector. Additionally, the idea that a RAID will fail simply because of the size of its component hard drives is quite a fallacy as well. While technically the more drives connected increase the chances of a single drive failure, it is the number of drives connected to the system that ensure that this single drive CAN be recovered WITHOUT damaging the full array.

A prime example of this might be a RAID 10 system (Hard Drive Recovery Group has recovered hundreds of RAID systems, including RAID 50 or "RAID of RAIDs"). In order for a RAID 10 to fail and have its data unrecoverable, two of the drives in the array must have the EXACT data mirrored and failed during a very, very short time period during which the initial failed drive is not repaired before the second drive fails. As you can tell, there are a lot of ifs ands or buts in this scenario. Not only is it unlikely, but highly improbable. The result? RAID 5 will live on. The low price of hard disk drives ensures this is all but a certainty.



	RAID 5 Doomed? The Myth Debunked As recently as October of 2008, a story on the newswires predicting the failure (pardon the pun) of RAID 5 as an effective data storage system has caused some disturbance in the IT community, if only because of its somewhat opinionated nature. The story, which ran on ZDNet as well as some of the other main IT industry rags, noted that SATA hard drive capacity, now reaching above the 2 terabyte level, could be a drag on future RAID 5 growth, and, and here's where the controversy comes in, actually INCREASE the potential for read errors during disk construction. In the article, the ZdNet columnist notes that SATA drives can suffer from unrecoverable read rates of 10^14, which extrapolates that these disk drives are unable to read a sector once every 100 trillion bits read. As many familiar with the data storage industry have noted, drives are already well above the two terabyte capacity level, which with this logic in mind, makes it all but inevitable when attempting disk reconstruction from a seven drive RAID 5 system failure. Where things begin to get shifty is where the article claims that during recovery, the array volume is likely to declare itself unreadable, thus stopping the RAID recovery or self-repair. From this point, the article also notes that once this event occurs during the recovery process, it is more than likely all 12 terabytes stored on this hypothetical RAID system will be lost. Or in the least, require professional RAID data recovery services. A Dead RAID On Every Corner? While certainly what the ZDNet article boasts would seem to be a huge boon for the RAID recovery industry ("Imagine a world with dead RAIDs on every corner!"), we believe this to be far from the case. The writer then goes on to throw in several basic premises about hdd technologies, such as the fact that mechanical parts tend to fail, tight data density can also increase failure rates, and that RAID 6 could be an expensive, but effective solution to this perceived issue. Click here to get a RAID recovery quote. The real problem with the article is actually quite simple. What the article does note is the "soft error rate," which is the rate at which data could possibly be misread incorrectly a single time, but NOT multiple times. Therefore, in everyday RAID storage system use, this number is really useless. Because of ECC encoding (Error Correcting Code) in most hard disk drives, the likelihood of failing to record more than a few bits and being unable to correct from it is almost impossible. Using the soft error rate is therefore a poor rate with which to base a story on. What the story should have been based upon is sector size and amount of ECC per hard disk sector. Additionally, the idea that a RAID will fail simply because of the size of its component hard drives is quite a fallacy as well. While technically the more drives connected increase the chances of a single drive failure, it is the number of drives connected to the system that ensure that this single drive CAN be recovered WITHOUT damaging the full array. A prime example of this might be a RAID 10 system (Hard Drive Recovery Group has recovered hundreds of RAID systems, including RAID 50 or "RAID of RAIDs"). In order for a RAID 10 to fail and have its data unrecoverable, two of the drives in the array must have the EXACT data mirrored and failed during a very, very short time period during which the initial failed drive is not repaired before the second drive fails. As you can tell, there are a lot of ifs ands or buts in this scenario. Not only is it unlikely, but highly improbable. The result? RAID 5 will live on. The low price of hard disk drives ensures this is all but a certainty. Professional Help Is Here Remember that at Hard Drive Recovery Group, a consultation is always free. Don’t take chances when it comes to your mission critical RAID-stored data. Call us first. Did this article help? Submit it to: del.icio.us \| Digg it \| Simpy \| Yahoo MyWeb Click here to request a free RAID data recovery quote. All material on this website is copyright 1996-2010 Hard Drive Recovery Group. All Rights Reserved.	24 Hour RAID Repair RAID 5 Data Recovery RAID 10 Data Recovery SAN Tape Recovery RAID System Repair SNAP Server Data Recovery RAID Failure Recovery SAN Data Recovery RAID Configurations Defined RAID Drive Recovery SCSI Data Recovery RAID Tape Recovery LTO Recovery RAID 5: Doomed? RAID Manufacturers RAID File Restoration Exchange 5.5 Data Recovery RAID Drive Repair RAID Hard Drive Failure RAID Disk Recovery Dell PowerEdge Data Recovery Repair Your RAID RAID Recovery Repair Exchange Disk Read Error Mac RAID Recovery Linux RAID Recovery Repair RAID 5 RAID Drive Failure HD Error Index Controller Error Drive Not Ready Fixed Disk Parameter Table or BIOS Error
	RAID Data Recovery \| Mac Data Recovery \| FAQ \| Flat Rate Pricing \| Repair Hard Drive \| Instant Quote Hard Drive Repair \| Why HDRG? \| Emergency Tips \| Packaging Instructions \| Hard Drive Data Recovery Laptop Data Recovery \| Hard Disk Data Recovery \| Tape Data Recovery Hard Drive Crash \| Hard Drive Data Recovery \| Hard Disk Drive Repair \| SQL Database Recovery MS Exchange Data Recovery \| NTFS Recovery \| Oracle Disaster Recovery \| Data Retrieval Exchange Disaster Recovery \| Hard Drive Recovery \| FAT Recovery Data Recovery \| Disaster Recovery \| Disk Recovery Service \| Clean Room \| Drive Recovery \| Site Map Notebook Data Recovery \| HDD Recovery \| Hard Drive Failure \| Emergency Data Recovery MS Exchange Recovery \| Linux Recovery \| Exchange Data Recovery \| MS SQL Recovery Exchange Database Recovery \| Testimonials

RAID 5 Doomed? The Myth Debunked