Hard Drive Crunch Looks Worse…

Global HDD supply to be 35% short in 4Q11, say PC makers according to Digitimes.

The consequences of this are very interesting:

  • When ARMed smart thingies are increasingly attractive to consumers with lowering prices, Wintel with rising prices may become less attractive.
  • The Christmas season when a lot of retailers ship huge volumes and make their annual profits will be hit hard.
  • This situation could take 9 months or longer to correct by increasing production.
  • This will increase interest in server-centric computing where few hard drives can serve many.
  • Wintel, which depends on every-increasing volumes to pay for extravagant life-styles will take a serious hit. What will M$’s client division’s bottom line look like with 35% fewer hard drives to license?

So, if there weren’t already enough good reasons to switch to GNU/Linux and thin clients, necessity may be the final push. I expect M$ will have to dip into cash reserves to keep paying off its “partners”. I expect more partners will look for smarter ways of doing things. I expect more consumers will find alternatives to Wintel. Since “8” is nowhere in site, M$’s empire will be negatively affected for months. It’s all good unless you need more storage… but at least you have the option to put that storage on a server.

This entry was posted in technology. Bookmark the permalink.

9 Responses to Hard Drive Crunch Looks Worse…

  1. oe says:

    Good point.Perhaps I should slice some of the ram into a swap ramdisk. Nice that UNIX philosophy woven into GNU/Linux reduces all the hardware to devices.

  2. oiaohm says:

    oe Memory cards do have a weakness. Don’t use them for swap and other rapid rewriting processes they don’t particularly like it.

    Raid setup memory cards at 32 meg is very decent to run a lot of OSs.

    But the idea of TB and larger storage might end up on hold for some people due to short falls.

  3. pogson says:

    Ray wrote:“Problem: the servers also need hard drives too.”

    Solution: Use nearly full/well-utilized hard drives on a few servers instead of under-utilized hard drives on every client.

    Example: Last year where I worked each client had 40gB of storage in a hard drive but the lab used thin clients with only four SCSI drives in RAID 1 very comfortably, saving 20x40gB = 800gB of hard drives and their noise and power consumption and bulk with 4 drives giving better performance. With a new server, I would use 500gB SATA drives in RAID 1 and could easily run a whole school from one server. At Easterville’s Chemawawin school we used a single file server with four SATA drives to serve the whole school of 600 accounts. The thin clients had no storage except optional USB drives.

    Do the maths. Calculate capital costs, power consumption, storage, bandwidth… GNU/Linux terminal servers provide the best performance at the lowest cost in money, time, resources for everything except video which we can do very well by other means.

  4. oe says:

    “Stuff not effected memory cards up to 32 GB.”

    You can run a full-blown netbook linux (Easy Peasy, MeeHo and UNR, among others) in the early ASUS eee’s with only 2GB of pure SSD.

  5. Ray says:

    Problem: the servers also need hard drives too.

  6. oiaohm says:

    Phenom other than the fact that the biggest factories for SSD production are no more because they were in Japan in the area that was destroyed this year or are off limits due to the radiation mess.

    Yes we have had the perfect crisis this year Phenom.

    Really hard drive production and SSD production really need to be spread out better around the globe to prevent this issue.

    Better tech is not the issue. Better redundancy in factory placement around the world. So criss in a single country cannot cause major upsets.

    35 percent short fall is way worse than what was predicted. 10 percent was the expectation.

    Stuff not effected memory cards up to 32 GB.

  7. Phenom says:

    You are very right about speed, too. That is just one more reason to make your original forecast less viable.

  8. pogson says:

    There are multiple reasons to get larger drives. One is that larger drives are faster because there are more read-write heads/platters and larger caches. The higher areal density decreases the average seek time for any given information and increases the average transfer rate. The 40gB drives my school was using transfer about 40MB/s and the 500gB drives transfer about 100MB/s.

  9. Phenom says:

    Pogson, you are wrong. Again. HDDs go bigger because people happen to need the extra space for data. Data, Pogson, not some imaginary overrated OS requirements that exist in your mind.

    And, HDD supply shortage will cause… Increased efforts to make SSDs cheaper, and more efficient.

    Every such crisis only results in better and cheaper technologies in the long term.

Leave a Reply