Storage Effect

500GB, 7200 rpm - who needs 3.5″?

Seagate announced two new 500GB notebook drives.  So what?

• The 50-year history of the disk drive is all about cramming more and more bytes on less and less real estate.  The real estate shrinks when drive formats drop a size.
• 500GB 2.5″ drives mean we’re close to not needing the capacity advantage of 3.5″ drives. It’s the beginning of the end for the 3.5″ form factor.  Servers have mostly made the switch with 2.5 SAS drives like Seagate’s Savvio.
• 500GB 2.5″ 7200 rpm drives mean notebooks can get desktop performance without sacrificing capacity.  Expect rapid adoption of 7200 rpm vs. 5400 rpm in notebooks now that there is capacity parity and less of a premium in power consumption.

SAS drives get bigger and smaller to take share from SATA for business applications

IDC data from InfoStor shows this year and next are the golden age of SATA drives in the enterprise. 

It’s not that the trend for high capacity storage abates in the future; it’s that SAS drives are expanding their capabilities to replace SATA in many applications. 

• Bigger SAS: 1 TB SAS drives, matching the largest available SATA capacity.
• Smaller SAS:  2.5″ SAS drives, rapidly replacing 3.5″ SCSI, FC and SAS drives. 

Why settle for an interface originally designed for PCs if you can get the same thing in SAS for a little bit more?

SATA drives won’t go away of course - they still provide the most capacity for the dollar.  If it’s good enough for an application, people will continue to use it. 

Have you made the jump to SAS?  Why or why not?

Becomes the first major supplier to offer a 2.5″ SAS storage system

Dell uncharacteristically took the role of technology leader and launched the MD1120, a direct-attach storage system with 2.5″ SAS drives for their PowerEdge servers (thanks Blocks and Files). It’s likely that their major competitors (and others) will follow with their own announcements in the near future.

Why 2.5″ SAS?

Make no mistake - they may be small, but they are the cream of the crop.  Fastest (for 10K rpm), most reliable, highest data integrity.  Oh - and they use less space and a lot less power than 3.5″ drives.

Don’t confuse 2.5″ SAS drives with notebook drives.  They’re similar in size, but that’s about the only thing they have in common.

The beginning of the end for 3.5″ enterprise drives

The only fatal flaw for 2.5″ and storage has been capacity. These drives are already the standard for servers, but storage system makers couldn’t make the numbers work with only 147GB per drive. 

It looks like 300GB may be the tipping point.  Seagate recently launched the first 300GB 2.5″ SAS drive, the Savvio 10K.3. 

What’s your 2.5″ storage plan?  Is it time?

Servers are already there; 300GB could be the tipping point for storage systems

Seagate announced a 300GB version of its Savvio 10K rpm drive, the highest 2.5″ SAS capacity yet available.  Is this the dawning of the Age of Small Form Factor for enterprise storage?  Information Week thinks so.

The server market has already made the turn.  HP and other high-volume server makers have committed to 2.5″ SAS drives for performance servers.  Power, size and reliability benefits make it a no-brainer.  And while 2.5″ hasn’t yet taken over servers from system builders, the switchover has begun.

For storage systems, the capacity just hasn’t been there. At 300GB, it may be.  A 2.5″ drive is 70% smaller than 3.5″ SAS drives, which max out at 450GB today (although 600 GB will be along before long). 

What say you?  How are you using 2.5″ 10K or 15K drives in storage applications today?

Even best-in-class rebuild times expose data to hours of risk

Blocks and Files points to an Demartek study (sponsored by Pillar) showing that the Pillar Axiom 500 rebuild times are much shorter on high capacity arrays that similar EMC or NetApp systems.

The glaring data beyond Pillar’s performance, though, is the teeth-clenchingly long times that data is one drive failure away from catastrophic loss in every case. 

The tests were conducted with about 50 500GB drives per system using RAID 5 (RAID 4 for NetApp), meaning the arrays can be rebuilt if one drive fails, but not two.  So during the rebuilds of from 3 to 23 hours, if another drive fails, all data is lost. 

Insert 1 TB drives into the equation, and your rebuild time (and vulnerability) doubles.

RAID 6 and other dual-failure protective schemes make this problem go away, but cost a little in capacity. 

How are you dealing with this?  I’ve heard that RAID 6 is gaining traction for 7200 rpm high capacity enterprise drives like Seagate’s Barracuda ES that are less reliable than 15K SAS enterprise drives (see Seagate’s 3.5″ Cheetah and 2.5″ Savvio for reference). 

Does your RAID vary by drive class?  What other magic do you apply to make this work?

Common-sense storage innovations slash storage energy use

Xyratex has taken a second step into the lower-power storage arena with the OneStor SP1224s that uses 2.5″ SAS drives.  In February they added software to their RAID platform that allows OEMs to selectively spin down drives not in use.

Lisa Hart at Xyratex says “This is like having a light switch on your wall.” Exactly!  Everyone understands that leaving lights on in an empty room is a waste.  Storage is not so different: 

1. Use smaller drives.  The OneStor SP1224s provides double the performance & half the watts with high performance 2.5″ SAS drives instead of 3.5″ drives.  This is the equivalent of having a smaller room that gets plenty bright with a smaller bulb. 
2. Turn off drives when not in use.  MAID technology is going mainstream after pioneering efforts by Copan and others.  It’s not for every application, but hits the spot for many of today’s high-growth capacity apps.
3. Turn off parts of drives when not in use.  Seagate’s PowerTrim technology on Barracuda ES and Cheetah 15K 3.5″ drives takes the MAID approach to the drive level by selectively turning off the lights in some “rooms” of the drive’s electronics in a smart way that doesn’t impact performance. 

Expect to see more vendors applying smart power management at all leveles to make a big dent in data center Watts per terabyte (or petabyte!).

Comments please: who’s using MAID, PowerTrim or 2.5″ SAS today?  Who’s planning to?

2.5″ enterprise drives reduce storage system power

Conventional wisdom is that 2.5″ enterprise drives are great for servers, but need more capacity to be viable in storage systems.  Problem is, conventional wisdom didn’t account for the vast quantities of these systems packed into an internet data center.

Mario Apicella, InfoWorld’s Storage Advisor, takes a look at Infortrend’s B12 storage system and says the future is now for 2.5″ SAS drives. 

Why?  Less real estate, obviously.  But also much lower power for the same performance: 2.6 Watts less per drive than same-capacity 3.5″ drives.  And dramatically less weight to ship, move around and pile on floors.  

Re-engineering current systems for smaller-format drives is the easiest, most effective way to make storage more energy friendly.

Here’s more info on the Savvio 2.5″ 15K SAS drive Mario referred to in his post. 

Who’s using 2.5″ SAS today?  Servers and/or storage.  What’s your experience?

Actual enterprise drive reliability meets expectations

One failure in a million hours?  It’s claims like these that seem extreme to some people when they look at enterprise disk drives.  Yet a study of 39,000 NetApp systems by a researcher have found that these drives fail at a 1% annual failure rate (AFR).  Robin Harris summarizes the study in his blog.

The translation from AFR to MTBF (Mean Time Between Failures) is not exact, but this AFR number puts MTBF in the million-hour ballpark,  showing that disk drive specifications do indeed portray actual reliability performance.

It’s hard to test for a failure every 1.6 million hours

This is not an exact science, because to prove that any one drive will only fail on average every 1.6 million hours (the spec for the Seagate Savvio drive), you’d have to run a whole bunch of drives a whole bunch of years.  This study is a nice real-world validation!

The multinationals have moved to 2.5″ server drives.  Are you ready?

System builders for the most part have yet to adopt 2.5″ drives in mainstream servers.  Yet Seagate sees these drives accounting for 34% of SAS/SCSI/FC units shipped in the 12 months ending in June 2008.

What’s the disconnect?

The multinationals have made the shift to 2.5″ for performance servers.  They made the jump because they can provide servers with lower power, smaller footprints, and better reliability by using these drives.  Seagate’s Savvio drive is a big part of that transition.

Your customers are seeing these solutions now. What will you say when they ask how your solutions compare?

The time is now to develop your alternative to these now-mainstream solutions from a lot of acyonym-named companies. The good news is that chassis manufacturers are ready now. 

Share your 2.5″ server challenges and successes.