Seagate Moving to Solid-State Enterprise Drives in 2009

By Aaron Heibert
published

Have you made the jump to solid-state yet? Probably not, due to pricing – but they are out there, and not cheap.

A few months back Seagate revealed that they would start selling enterprise based Solid-State Drives (SSD) some time in 2009.

Some industry experts believe that selling the Solid-State idea to industry IT managers will be difficult because there is currently no clear and cut way to describe endurance or life expectancy of SSD’s. Seagate is currently working on this problem with JEDEC standards body.

Rich Vignes, senior manager of market development believes, “As companies like Seagate start to demonstrate field-proven reliability and endurance in enterprise applications, we’ll overcome those (solid-state drive) endurance fears.”

With time, SSDs will catch on since they offer much better mean time between failures (MTBF) than standard mechanical based hard drives, they generate much less heat, require much less power, and can also be compacted into a smaller form factor as well. This presents many attractive key qualities for SSDs – the trick now is to deploy and convince the industry to make the move.

Quoting Gregory Wong, an industry analyst at Forward Insights, said, “IT managers tend to be conservative, so the qualification time will be quite long—nine months to a year, and early adopters will be Web 2.0 companies such as Google and Facebook.”

As it currently stands, consumers won’t be going mainstream with SSD any time soon at all. Prices may continue to trickle on the downward slope, but don’t expect leaps and bounds just yet. Let the enterprise market play out first.

Aaron Heibert
6 Comments Comment from the forums
  • blackened144
    We design and use our own supercomputer at my job for data manipulation. The drives are always the weak link in our systems. We have been testing 10 node systems using 15k SAS drives, but I cant wait until we start to get some SSD drives in here.
    Reply
  • I don't really agree with you saying: "and can also be compacted into a smaller form factor as well."

    Small,in 2,5 or 1,8" drives is not an issue in IT business.
    But can you honestly say you've ever seen a 500GB SSD,and how large it is?
    I mean,currently the largest SSD I've seen is a 160GB SSD, and they could't bring that one in 2,5" form factor!
    the 3,5" bay was all filled with flash chips!

    I wonder if companies will believe that, when they can get a 1 to 1,5 TB harddrive in 3,5" form factor.
    At this rate we'll need a drive the size of a CD/DVD rom (5,25")to host 320 to 512GB,which is still far off the 1,5TB 3,5" drives

    Also, the possibility is there that the larger the SSD's size, the more it'll consume on power. Although maybe many modern SSD's have better power management than previous drives.
    SSD drives not always consume less power than HD's.
    In some cases the reverse is true.
    Reply
  • estreetguy
    ProDigit80I don't really agree with you saying: "and can also be compacted into a smaller form factor as well."Small,in 2,5 or 1,8" drives is not an issue in IT business.But can you honestly say you've ever seen a 500GB SSD,and how large it is?I mean,currently the largest SSD I've seen is a 160GB SSD, and they could't bring that one in 2,5" form factor!the 3,5" bay was all filled with flash chips!I wonder if companies will believe that, when they can get a 1 to 1,5 TB harddrive in 3,5" form factor.At this rate we'll need a drive the size of a CD/DVD rom (5,25")to host 320 to 512GB,which is still far off the 1,5TB 3,5" drivesAlso, the possibility is there that the larger the SSD's size, the more it'll consume on power. Although maybe many modern SSD's have better power management than previous drives.SSD drives not always consume less power than HD's.In some cases the reverse is true.
    You obviously have no idea what you are talking about. Mechanical based arrays need alot of room and alot of cooling. This is what controls the design of the enclosures needed.

    250GB SSDs have been made by Intel already - not publicly available yet though. Its only a matter of time. The 250GB model that Intel has developed is 2.5" form factor as well.

    With the cost of rental per square foot - datacenters pay ALOT of money for space, not to mention the cost of opperating the large air conditioning units in big datacenters - wow, you have no idea do you?
    Reply
  • Estreetguy - Are you serious? Your blasting him for not knowing what the largest SSD drive is.

    Eventually SSD will get to 1.5TB+ in a 2.5" form factor, but how long will that be? By the time you have 1.5TB SSD, 2.5TB+ drives will be mainstream.

    Prodigit80 is right, also confirmed by Toms Hardware that current SSD drives actually drain your battery quicker (uses more power)than a traditional hard drive.

    Manufacturers will not change the way they build their arrays, they will simply claim that they are compatible with (2.5") SAS or SATA drives. SSD drives currently have SATA interfaces, although enterprise SSD drives will likely debut some day with SAS interfaces to work in Dual Controller systems better. You will still need to ventilate the drives, but even with traditional drives that does not require a tremendous amount of real estate in an enclosure. The controller runs the fans at the speed it needs to keep the internal temperature in the enclosure to a preset threshold. Ideally, if the SSD drives required less power then the power needed to cool the enclosure would be reduced, hence your power savings.

    As for your paragraph about data centers, you aren't changing drive form factors, so a 1U server would still be a 1U server. So floor space would not change. Data centers are built to cool X number of WATTS per square foot, so what you as the customer fills that floorspace with is not as relevant. Data centers power savings are determined by what technology they leverage to provide that cooling.

    SSD is hardly a game changer, you will still need to use Raid with a SSD based array in the enterprise. Optimistically, you will be paying the premium for SSD to get a lower drive failure rate and an elimination of latency.
    Reply
  • nekatreven
    none of you are going to be able to agree and none of you are without errors in your statements.

    The power test Toms did was simply not conclusive, especially in light of the PAGES of comments about how the test was flawed or should have been conducted differently, even after the correcttion was issued

    This is without mentioning that SSDs are still relatively new and HDDs are ancient. By the time ssds mature, it will be like comparing flopppies to cds, both in power usage and capacity. HDDs wouldn't even have the capacities they do today without perpendicular recording, and there are only so many breakthroughs left for hdds
    Reply
  • Pixels303
    they mention something like Wear leveling. The reason why hard drives exist still today is for that reason. In the beginning Tri-M engineering used a SD card for windows XP operating system in a compact 4" computer system rather than using conventional hard drives to reduce it's size. After a few hours of operation the device failed simply because Microsoft writes so often to the hard drive even when it is idle. Hard drives may have latencies and be bulky hot fragile things, but they perform very well over trillions of re-writes and reads. If I understand correctly, this Solid state stuff uses similar technology as your SD or CF memory cards which is great for read speeds and low power, but questionable for how many times you can rewrite them before some data goes missing. I'd be curious what that figure would be.
    Reply