The Register® — Biting the hand that feeds IT

Feeds

TMS flash array blows Big Blue away

Creams SPC-1 result

Cloud storage: Lower cost and increase uptime

Texas Memory Systems has absolutely creamed the SPC-1 storage benchmark with a system that comfortably exceeds the current record-holding IBM system at a cost per transaction of 95 per cent less.

TMS submitted a RamSan-630, a 3U box holding 10TB of single level cell flash that delivers – according to TMS – one million IOPS (10GB/sec) random sustained throughput, and 500 watts power consumption. It scored 400,503.2 IOPS compared to the 380,489.3 achieved by a six-node IBM SAN Volume Controller system with two back-end DS8700 arrays, at a cost of $18.83/IOPS.

Here's where the TMC box just blew Big Blue away: the RamSan-630's price/performance was $1.05/IOPS, 5 per cent of the IBM figure – or 95 per cent less, looking at it another way.

The IBM system needed 294.7TB of configured disk – we counted 2,048 146GB 15,000rpm drives – to deliver Application Storage Unit (ASU) capacity of 97.582TB. Its total system cost was an eye-watering $7.17m. The RamSan-630 cost $419,000, less than 6 per cent of the IBM system's cost. It needed 13.744TB of raw flash with an ASU of 8.117TB – far less storage capacity than the SVC system.

A TMS RamSan-620 achieved 254,984.21 IOPS at a cost of $1.13 in October 2009. It needed 4.9TB of SLC flash and the total system price was $287,858. The RamSan-630 also comfortably beat the IBM SVC system in terms of latency.

There is surely no point in IBM – or any other vendor – trying to beat TMS on the SPC-1 benchmark with anything other than its own flash-based system.

Also, given these numbers, why would anyone with a workload resembling SPC-1 and needing, say, more than 100,000 IOPS, choose anything other than a flash-based system?

SPC-2

TMS also submitted the RamSan-630 to the SPC-2 benchmark. This consists of large, sequential I/O generally required by file processing, database query, and video-on-demand applications. A single InfiniBand-attached RamSan-630 produced 8,323.13 SPC-2 MBPS with an SPC-2 price-performance of $49.37 per SPC-2 MBPS.

The top MBPS score is held by an IBM DS8800 with 9,706.74 and a price performance of $270.38. An HP XP24000 array scored 8,724.67 MBPS with a price performance of $187.45 while an HDS USP-V achieved the same MBPS number with its price-performance being marginally different at $187.49

The RamSan-630 was in the same MBPS ball park, but substantially cheaper, showing that it does well in the high-bandwidth area, as well. All in all, these are a good couple of benchmarks for TMS – and ones it hopes will help slay the high-performance storage array business. ®

Steps to Take Before Choosing a Business Continuity Partner

Latest Comments

not really

One reason the TMS system is cheaper is that there has always been a niche to get a good $/IOP score with a small quantity of flash. Few vendors occupy this niche, because there isn't a huge market for it.

The SVC config tested is over 97TB and has a "minimum" of 25% headroom to scale the performance as it only has 6 out of a possible 8 nodes, it's not an unrealistic config for Enterprise. Will the RamSan scale ten times and see a linear increase in performance or will it top out and simply add cost and capacity, but not performance, hence ruining it's $/IOP..

Horses for courses.

0
0

Re: "memory is faster than disk"

The SPC-1 benchmark runs for long enough that it captures something approximating steady-state performance of the flash.

It will not capture what might happen to the flash days, weeks, or months, down the road.

0
0

It is Cheaper...

...than IBM. But the IBM system is a classic example of a benchmarking product built for a big headline number but not something that anyone would purchase in the real world. Their costs for capacity and performance are both stupidly high.

I can see a place for the TMS unit where ultra-high performance over a relatively small amount of storage is required, but given it is still coming in at up to 5x the cost of capacity of other SPC-1 results it isn't exactly what you would call general-purpose storage. Saying that it is cheaper than IBM for both capacity and performance is no more of an accolade than "not in a position to stiff customers who have no idea what they are buying but know it needs to be blue".

0
0

More from The Register

 breaking news
What's HP got under wraps? Looks awfully flash and tape shaped
What happens in Vegas won't stay there - we've got the details
Microsoft borks botnet takedown in Citadel snafu
Stupid Redmond kicked over our honeypots, wail white hats
SCO vs. IBM battle resumes over ownership of Unix
Zombie lawsuit back and wants to suck the brains out of Linux
IBM's $1bn layoffs latest: Now axe swings in US, Canada - reports
Union claims 121 storage bods canned after dismal sales
NetApp musters muscular cluster bluster for ONTAP busters
Storage array OS overhauled to juggle more nodes, go down on you, er, less
HP adds 'Haswell' Xeon E3s to entry ProLiant servers
Gussies up MicroServer for SMBs, adds baby switches
Buffalo herds DDR3 RAMs into DriveStation's spinning rust corrals
Claims cache-packed gear keeps up with flash drives
'THINNEST EVER' spinning terabyte beauty slips out of WD fabs
Size-zero drive packs a whopping 143GB per millimetre