HP is blowing its 3PAR horn loudly, saying it has doubled 3PAR's customer base since buying it last September, and that 3PAR is now top dog on the SPC benchmark. The 3PAR business has had triple digit year-on-year growth (100 per cent at least) for the quarter ending July 31st. That means revenue of at least $92m. HP says 3PAR' …
Flash appliances and enterprise disk systems are in separate markets and fulfil different requirements. Customers need large amounts of capacity that delivers high-levels of performance to a LOT of attached hosts. You can't do this today with a flash appliance.
These benchmarks show efficiency and performance - the cost per IOP was excellent for a high-end device and didn't require additional external virtualisation appliances to stripe across multiple backend arrays.
A single virtualised array that can scale this far - not easy to do or EMC would have benchmarked the VMAX ages ago.
Pull your head out of Netapp's backside long enough and you may understand what storage is really about.
Bit strong there? We have used the Texas RamSans for a few years now and really like them, but we have them as front-ends for EVA arrays for a few select applications, with the RamSans providing the flash layer and the EVAs the scale. We do not use the RamSan/EVA combo for general storage. We have considered 3Par arrays as a means to replace both the RamSans and the EVAs in one frame, and they make sense for a lot of our general, virtualised "private cloud" needs. After all, if massive bandwidth and scale is what you want, you'd probably go for the hp P9500 or the EMC DMX, where you have lots of front-end ports, but then the price will make the 3Par look cheap! Or you could use cheaper scale-out products like the P4000 series. Horses for courses and all that, but you often end up with comparisons between two dissimilar products, which implies there are a lot of "idiots" out there working with the varied requirements of corporate storage.
PS: Has anyone asked Micheal Dell what he thinks of hp's growing the 3Par base by "100%"? >:D
Agree to much of what you said, but the STAC M3 is very much a multi-user benchmark, so I believe the 11Million IOPs are valid.
Couldn't agree more!
Spot on - compare the 3PAR array with other similar high-end disk arrays from other vendors and it kicks ass! Course small flash appliances can deliver higher IOPS per GB or lower cost per IOP, but not many customers have a 13TB workload that needs 400,000 IOPS! More customers have a blended workload, 100-200TB, and require a couple of hundred '000 IOPS which the 3PAR array nails! When flash devices have the functionality and resilience of a full featured disk array then make the comparison but at the moment the flash appliances are Indy Cars (stripped down, bare spec, all about the speed) vs the 3PAR Ferrari!
Front end ports
How many host ports do you need? The P10000 will support 192. The P9500 will only support 160.
11 Million IOPs
Not sue fo the similarities of the benchmark, but I recently saw a financial services benchmarking suite from STAC Research, which had a storage array from www.kove.com at 11.7Million IOPs.
Seems 3Par/HP might have along way to go yet?
Awful article - 11 million IOPS ?
No disrespect to Violin, but Chris's article is pretty poor, enterprise is not just about performance or even capacity, take a look at many of the flash only arrays and it's pretty easy to pull their architectures apart. Does just buying the one really give you the availability and serviceability features of a more traditional enterprise class array like the V-Class, what about value add features, Sync / Async replication, snapshots / restore etc, etc.
Kove is DRAM, so not apples for apples, that's like vendors short stroking the array benchmark from cache, it's fast but has almost zero use case for the vast majority or workloads, that's partly why SPC exists, to avoid the smoke and mirrors used in Vendor benchmarking. 3PAR V-Class are squarely aimed at the vast majority of Enterprise workloads, not the corner cases handled by flash only such as Kove, Violin etc.
Chris, the 3PAR SPC numbers are very impressive when you compare to the real competitions SPC results. They could have added SSD to this test but then it would have skewed the results to the point they're no longer directly comparable against other vendor arrays. So why not beat the real competition hands down, with a real world solution (not a lab queen) running good old spinning rust which can be directly compared to previous submissions.
Kove, (appears to be), a persistent, (so not cache), block device that supports snapshots!
Persistent so what ?
It's DRAM and running a completey different and unrelated benchmark, what bit of this are you not getting ? Your argument is like comparing a 747 with the Space Shuttle. So what if it's persistent, anyone can make DRAM persistent it's done all the time with cache, but it doesn't make the numbers any more valid. BTW I hope it's got a better system than battery backed DRAM, otherwise I'd want two as a minimum or some form of destage to disk. I haven't seen one these new age arrays yet that once off the slide deck can touch a real enterprise array for capacity, availability or features.
R&D pays off
All of HP different storage systems, like EVA, the Hitachi, MSA and so on are now renamed. All begin with a P followed by a number.
Unified storage: Mission accomplished!
Please learn how to interpret $/IOP for SPC-1
D from NetApp here.
3Par's $6.5/IOP are after the 50% discount. Other vendors state list prices. So please look at the full disclosure in the SPC page in order to understand the pricing, and calculate $/IOP based on list to keep it apples to apples.
Even though list pricing means less than nothing these days...
Good result overall even if almost 2,000 disks were needed, would be interesting to see it with RAID6 to provide a similar level of protection and efficiency to NetApp :)
Hi Chris, thank you for the recognition of the killer HP 3PAR benchmark! I think it has been stated already in some of the comment postings, but wanted to emphasize the point that $/usable terabyte has traditionally been a major customer buying factor. The Texas Memory Systems $/ASU (cost per usable terabyte) is 400% higher than the 3PAR array, and the economic efficiency for those aggregating lots of workloads in private and public clouds is a key factor.
Craig Nunes, VP of HP Storage Marketing
Must be an imposter! That was a technicaly coherent and reasonably short statement from a supposed hp veep!
“HP is blowing its 3PAR horn loudly, saying it has doubled 3PAR's customer base since buying it last September”
So let me get this right, maybe I have a warped view of things but......
IDC shows that just before HP bought 3PAR, 3PAR had 0.58% of the storage market share. This is of course market share and not customer base, but gives an indication of the size we are talking about here.
Anyway, I would have expected the increase in customer base to be significantly greater given the size of HP, the IT refresh cycle and the fact HP customers have an alternative to the EVA. Let’s say a company that made washing machines sold 20 washing machines in 2010. They were then acquired by Walmart and in 2011 they had sold 40 washing machines! Hmmm wait a sec!
Sure, It would be very hard to double your customer base in a year if you had a 15% market share in a competitive environment, but when you’re sitting at just over half a percent, you can churn out some great benchmarks like this one and you get acquired by one of the biggest IT companies out there I would have expected the customer base to increase by more than 100% in a year.
I’m starting to think 3PAR could have doubled their customer base without the HP acquisition and HP is actually blowing a horn they should be a bit ashamed about.
Credit where it is due!
I would be very surprised if 3PAR had doubled its' customer base in 9 MONTHS organically! In a storage market that is growing single digits per annum, to grow 100% in 9 months is impressive! Plus the wording suggests doubled the number of customers, so doesn't include upgrades / additions to existing customers. There have been some very high profile, similar acquisitions - Dell Equalogic and Compellent, EMC Isilon, IBM XIV, etc. All were big storage companies with comparable market share buying smaller companies with disruptive technology. But I don't believe any of those acquisitions resulted in such an immediately increased uptake.
@Man Mountain. Impressive? Really?
Growing 100% organically in 9 months in a single-digit growth market is indeed impressive.
However, converting existing HP EVA customers running out of service into HP 3Par customers while loosing overall marketshare isn't.
@AC Did you mean loosing or was that losing ?
@AC Show us the market share numbers then, HP's aren't in yet, so you appear to be jumping the gun.....or then again you could just be trolling for HP's competition.