Hitachi increases reliability with 3TB enterprise-class drive

[Jos]

Hitachi has introduced a new member to its Ultrastar series of enterprise-class hard drives, one that the company claims will deliver greater reliability and save on energy costs at the same time. The Ultrastar 7K3000 is available in 2TB or 3TB variants featuring a 6.0 Gbps SATA or SAS (Serial Attached SCSI) interface. It comes with a 7,200RPM spindle speed, 64MB of cache and a power consumption of 5.6W or 6.6W while idling for the SATA or SAS versions, respectively.

Like its predecessor -- the 3.5-inch A7K2000 -- the new drive has five platters, but it increases the areal density on those disks by 50% from 400GB per platter to 600GB. This means that datacenters could save energy by delivering the same level of storage using fewer drives. Hitachi also claims to have increased the mean time before failure rating of the drive to two million hours, resulting in a failure rate around 40% lower than the industry average (1.2 million hours MTBF).

Both Ultrastar 7K3000 models are available with regular and bulk data encryption (BDE) options. Although Hitachi made no mention of pricing, the SATA versions are already shipping, while the SAS models are expected around mid-year.

Permalink to story.

https://www.techspot.com/news/42128-hitachi-increases-reliability-with-3tb-enterprise-class-drive.html

 

[9Nails]

24 hours per day
x 365 days per year
= 8,760 hours in one year.

2 Million hours is about 228 years. Call me a pessimist, but I doubt the actual MTBF numbers will hold true. I'm considering it lucky to get 5 years from a drive nestled lovingly in a rack with air and power conditioning.

 

[Guest]

I agree with you completely.

 

[Rick]

9Nails said:
2 Million hours is about 228 years. Call me a pessimist, but I doubt the actual MTBF numbers will hold true.

You are correct if you're comparing small numbers of drives. MTBF numbers are almost entirely meaningless for consumers.

An MTBF of 1,200,000 hours actually represents the total number of hours by a very, very large number of units, not just one or two. You need a VERY large sample size to come close to the MTBF.. like an enormous data center filled with them.

A higher MTBF drive *should* indicate a more reliable drive, but it in NO way indicates how long ANY of those drives will last on their own.

For example, if you have 50,000 of those hard drives in your data center, then you can reasonably expect one to fail every day (on average) (ie. 1,200,000 / 50,000 = average hours it takes for one of those drives to fail). If you have a million, then you you could reasonably assume you're going to have 1 bad hard drive every hour.

If you have two in your PC though, they will not each last for 4400 hours. Manufacturers have no way of testing a drive until it dies (that process could take many years) so they just test thousands at a time to see how many fail. There is where the MTBF numbers come from.