Are enterprise drives worth the power?
Data center power consumption is getting a lot of press play lately. The issue: increased density means that a data center rack that used to need a 2kw may now need 6kw. And for every dollar spent to power equipment, another 40 cents is spent on cooling that equipment. It is a real problem.
Storage: power hog?
Most of the published press concern seems to focus on pizza boxes and blade servers, not big iron arrays, probably because Google raised the issue years ago. Yet EMC is responding with a couple of technical marketing papers (see Power Efficiency and Storage Arrays and EMC Symmetrix DMX-3 Electrical Power Estimation and Configuration Planning. HDS doesn’t seem to have any papers addressing the issue.
Feets, do your stuff
Nice tap dance by EMC in the latter paper. The simple fact is that all 3.5″ drives used in big iron arrays are power hogs. Fibre Channel alone adds about 2 watts per drive, according to Seagate’s Cheetah data sheet, which isn’t too surprising when you consider that FC drives have two interfaces. Even 7200 RPM SATA desktop drives use about 12 watts instead of FC 15k drives 16 watts. And it looks like SATA interfaces are slightly less power-efficient than PATA drives – on the order of 5% – while SAS drives are even worse – about 40% higher than SCSI. Oops!
So what is a “green” array vendor to do?
Seagate thinks it has the answer in their new Savvio 2.5″ enterprise drives. At 8 watts they use about 40% less power than equivalent 3.5″ drives. Of course, they are 70% smaller, so vendors will be able to cram even more drives into a smaller box, increasing power use per rack.
Maybe it’s time to get back to RAID
Striping was the original performance hack for disks. To my mind, the storage power problem isn’t going to get much better until vendors start using disks that were designed from the ground up for low power consumption: notebook drives. These average about 3 watts each and are designed to be shut down frequently. While each drive is pretty slow, stripe across a half dozen or so and the performance is better than you could buy 15 years ago. Didn’t people have databases and OLTP 15 years ago?
Or for a really radical solution. . .
Flash-based SSDs are even more frugal – at least in power – using a fraction of even the best notebook drive’s power. Perhaps the day will come when a data center will be cooled by a couple of room-size air conditioners and Samsung will be selling a couple of hundred million enterprise flash drives a year.
Comments welcome of course. And for all you Americans out there, don’t forget to vote tomorrow, Tuesday, November 7th.
Robin,
3.5 inch disks will stay around for a while.
Big-iron mechanical form-factors are not suitable for 2.5 inch disks. This will require / enable new designs and everything will shrink… and then it would be too small to sell … for so much.
Also… Datacenter designers should be worried about power wastage in 1U servers… with all of this unused ‘commodity’ motherboard infrastructure such as multiple PCI slots, boot disks, etc.
It seems that the latest wave of innovation is limited to packaging these power hungry, outdated ‘reference design’ motherboards into cargo containers and back-to-back rack mounted solutions.
Well, Intel’s power hungry Netburst architecture was replaced for an architecture originally designed for laptops. It certainly isn’t illogical to think that another designed-for-notebooks architecture – disks – could replace power-hungry disks.
As for power-wastage: saw an interview in Computerworld with Ken Brill, founder of the Uptime Institute, where he states that in the data centers he’s seen, anywhere from 10-30% of the servers are “dead” – not running an application but still powered up. Problem: no one is responsible for identifying and turning off unused servers. Kind of like government programs.
Depending on the strength of the power issue, I think the turnover to 2.5″ drives could happen a lot faster than most of us are thinking. IMHO, disk performance is a steadily shrinking issue as enterprise data continues to cool, and while they are currently more costly per GB, the savings in power, cooling and floorspace would wipe out a lot of those differences. EMC will be the last vendor to make the change, so the question is: who will be the first? On past performance, I’d say HP, but maybe IBM or Sun is finally willing to show some leadership. Stay tuned.
Robin
Robin, I had to chuckle when I read your November 6th blog (Will power kill the 3.5″ drive?). I work in the storage industry and am old enough to remember the horse and buggy days (pre 1990) when 3 terabytes of IBM 3380 “DASD” would completely fill a 50,000 ft2 data center. Energy costs were tens of thousands of dollars per month to keep this spinning and cooled. Today, the average modular storage array ships with 6 to 7 TB (or more) and costs a very small fraction of this in terms of energy to keep the storage running. You have to give the 3.5″ drive some credit for this (along with RAID technology). Yesterday a hero, today a goat…….
As the economic and social costs of energy become more expensive, storage vendors will respond. RoHS is one example where vendors stepped up and did the right thing albeit because they’d have lost business without adapting. Small form factor drives will one-day become as reliable and economical as they are energy efficient. Don’t doubt they’ll be found in enterprise storage as the market demands it.
Remember, the storage industry is always looking to be green (as in $$$) as well as environmentally friendly.
I enjoy your blog, thanks for the rant.
Big-Iron Bob,
I remember those days too. When a 500 MB disk was the size of a washing machine and cost $50k. I sold a lot of DEC’s RA81 disks: 1.8 GB in a 42 inch high rack for ~$60k. They went like hotcakes.
Yet this points to a larger issue: power efficiency, as well as cost and speed, have enabled computing to become ever more woven into the fabric of our lives, culture and civilization.
At the time of the RA81, the 5MHz VAX 11-780 supermini was a 6 foot high behemoth in 3 19″ racks and required 220v power. The backplane’s bandwidth was less than USB 2 today. There is probably a microcontroller in a clock radio that is slower than that VAX today, but I wouldn’t bet on it.
So now Intel’s 70 watt NetBurst architecture chips are too hot and power-hugnry for today’s Internet Data Centers, so the new Core Duo chips start at less than 14W, meaning you can pack several where you had one CPU before. And what will those CPU do for storage? Several 15W drives? Not for long.
We’re on track to have a 1TB 2.5″ laptop drive in 2010. Seagate’s Savvio drive is a step in the right direction, but its greater power density means more cooling problem, not less. I don’t think people will be willing to pay that price for very long, so I hope they are working on a 5 watt Savvio.
And yes, I am an MBA-carrying capitalist. I want to see a very healthy and profitable storage industry. What bothers me is that I see an industry that is getting ready to drive off a cliff. Maybe that will be today’s post.
Robin