Seagate has beaten Western Digital to the post in breaking the one terabit per square inch data barrier on a disk platter. (Now we know why they were so anxious to acquire Maxtor and their R&D facility!)
Currently, with a data density of 620 gigabits per square inch, the maximum capacity of a 3.5 inch is 3TB. With this announcement from Seagate, we are likely to see hard drive capacities shoot up to 6TBs for 3.5” drives and 2TB for 2.5” models. All of this is possible by using lasers to heat tiny areas of the platter (a.k.a Heat Assisted Magnetic Recording, as discussed on this blog two weeks ago).
This sounds all very well in theory but from a data recovery point of view a defective drive head could now cause even more damage. Currently, a defective drive head can cause bit corruption. At worst, a physically damaged drive head can have a scouring effect as it moves across the platters. But having a defective drive head which also has a microscopic laser attached to it could compound or magnify even small drive head defects. Seagate have not mentioned anything in their press releases about how the laser will impact power consumption and more importantly has not mentioned anything about how the laser will impact read /write performance. We look forward to purchasing one of these drives to put it through it’s paces in our lab.
For most 2.5” inch drives (or simply laptop hard drives), 1 or 2 platters (platters are the actual disks where the data is stored) have been the de facto standard for a number of years now.
But, around this time last year, Seagate brought out their Savvio drive with 3 platters in it. Now Western Digital have thrown their hat into the ring and brought out their own three platter 2.5” drive called the S25. This disk, like the Savvio, has a capacity of 900GB. On their website, Western digital claim the drive has a whopping 2 million hour MTBF (mean time before failure). Now, over the years I have heard some ludicrous claims proffered by drive manufacturers but claiming this MTBF figure (i.e. half the drives will fail within 228 years of operation) is farcical.
One of the largest hard drive studies ever untaken in the world “Failure Trends in Large Disk Drive Population” was conducted by the engineers at Google. It helps shed some light on some of the manufacturer’s claims. They noted “situations where a drive tester consistently ‘green lights’ a unit that invariably fails in the field” In other words, in manufacturer’s tests, they pass drives that would fail in real life.
Storage could be considered a critical area of the I.T. industry. Everyday we rely on hard drives from surfing a webpage, to storing a thesis and to companies storing whole databases on them. Computer users (consumer and enterprise) want storage options that have reasonable performance coupled with a reasonable level of giga/tera-bytes at a reasonable price. They do not want spurious numbers or claims pulled out of the air so some marketing department can get their egos stroked. With realistic expectations of hard drive failure rates, users can formulate better disaster recovery plans. Forewarned is forearmed.