Slashdot Mirror
Start-up Claims SSD Achieves 180,000 IOPS
Lucas123 writes "Three-year-old start-up Pliant Technology today announced the general availability of a new class of enterprise SAS solid state disk drives that it claims without using any cache can achieve up to 180,000 IOPS for sustained read/write rates of 500MB/sec and 320MB/sec, respectively. The company also claims an unlimited number of daily writes to its new flash drives, guaranteeing 5 years of service with no slowdown. 'Pliant's SSD controller architecture is not vastly different from those of other high-end SSD manufacturers. It has twelve independent I/O channels to interleaved single level cell (SLC) NAND flash chips from Samsung Corp. The drives are configured as RAID 0 for increased performance.'"
there's enough for everyone in here
Neat.
"I'm just here to regulate funkiness."
They're fishing for a price point? Quick, everybody make a comment to the effect that such a drive is only worth about $10...
Orwell: "In a Time of Universal Deceit, telling the Truth is a Revolutionary Act"
Start-up Claims SSD Achieves 180,000 IOPS
Claims? As in no one else but the company has stated this "fact"? I wish this article waited for a review before being posted :S
Disclaimer: I am not god.
We may not be created equal
But we can be treated equal.
am i rite?
I used pre-production versions of these. I tested them with Terabytes of test data in random write tests. They are amazing, and can saturate a 1Gb FC connection with random writes. They are very resilient. We put these in my company's demo boxes to show that our architecture can compete with EMC. Kind of cheating, but we told them that it was a special drive that enables us to show the limits of our storage management architecture in a small, 1U box, instead of just showing you the limits of physical hard drives.
We beat their 8Us of EMC hard drives by 34% with just one of these 2.5" drives, and we had bottlenecks all over the place in our small demo box. And they did the testing, not us.
The thing about these drives is that they are more expensive ($/GB) even than registered ECC DDR2/3 RAM, which obviously is going to be even faster.
This looks like a pretty good device. Tho i haven't heard much about them until recently I'm still pretty skeptical about their claimed lifespan - something that would be able to handle 24/7 consistent read/write for a number of years. The other thing that leaves me scratching my head is the missing DRAM cache -- I thought the need to store information then write it in buffer was kind of important especially with writing as fast as SAS is supposed to be able to transfer it. If these were hitting the shelves today i'd probably wait it out.
"i lost my dignity on a slippery wiener"
And Intel's enterprise-class SSDs already offer sustained speeds of up to 250MB/s read and 170MB/s write, wouldn't read speeds of approximately 500MB/s and write speeds of over 300MB/s be expected?
... They are amazing, and can saturate a 1Gb FC connection with random writes. ...
A one gig fiber connection is no great shakes.
Did you mean something else?
So they will realize the Data is useless and go with what the CEO has asked for.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
in Raid 0 you are in deep deep do-do.
Most peole know that striping 2 or more disks can give a performance increase but the idea of putting business critical data in a Raid 0 config is IMHO just plain crazy.
1GbFC? How well can this stand up to modern 8GbFC or 10GbE iSCSI?
Or better 10Gb FCoE (lower overhead than iSCSI). In theory with a fast enough controller they should be able to do it for reads in a RAID1 configuration.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
The summary seems to end abruptly and the article.
One of their drive can saturate the bandwidth of 3Gb SAS doing real I/O (not single sector I/O off drive cache)
The same principle should be extendable to RAID5.
Several separate, smaller devices combined into a RAID5 array, all inside one 3.5" case. That would take care of failures in one of the sub-devices. In case the "mainboard" that connects them all and holds the SAS interface fails, make the "mainboard" exchangeable. Swapping it will revive the drive.
C - the footgun of programming languages
Only worth about 10$? You're crazy, I'd pay up to 20$ for such a drive!
I can move 2TB over 100M in just over a minute at normal walking pace, and that's if I have a single drive.
If I have a backpack or push-cart loaded with 50+ drives, we are talking tenths of a petabyte per minute per 100M.
If I load up an airplane full of high-density media and send them cross country or halfway around the world, we are now talking VERYBIGNUM bytes per day.
Still, this is way cool, and sneakernet isn't what you need if you want to copy rather than move your data.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
With all the fast SSD's I've tested I've found the controllers to be a bigger bottleneck than the SSD itself. I've seen 50% performance gains on the Intel x-25e's simply by hooking them to a second machine with a different controller. Even with the best performer (Intel ICH9) I still had the feeling that the controller might have been holding the drive back a bit. Haven't tried it with an ICH10 based board yet though so perhaps there's significant improvements there. (on further reading they claim to be using SAS, I'm not aware of any really high performance SAS chipsets, they all seem to be targeted at RAID's of traditional HDD's and so can't keep up with SSD, I'd really be interested in some details of their test).
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
What do you mean by "EMC hard drives"? Are you referring to SSD, SATA, or SAS? Also, what sort of device is controlling these drives? My Symmetrix with loads of global cache is going to be much more performance oriented than say an AX-100.
I guess what I'm getting at is that "8Us of EMC hard drives" doesn't really mean much to me unless they are all configured in some sort of super-dooper 45+1 RAID group. ;)
The thing about these drives is that they are more expensive ($/GB) even than registered ECC DDR2/3 RAM, which obviously is going to be even faster.
So, how much do they cost exactly?
Find a job you like and you will never work a day in your life.
Well, I don't know the whole setup, just that it was about 10 drives (15k) SCSI (not SAS) in a RAID 5. I don't know how much cache. It was a Clarion unit. But, the customer thinks, "Wow, your little box that I've never heard of has just beaten EMC." They don't get into the technical details when they make that sort of decision.
TFA said serial-attached SCSI (SAS) was currently 6Gb/sec going on to 12 by 2012. SATA III is also 6Gbit/sec.
0.5GB/sec is 4Gbit/sec, which is under the SAS limit.
Even if it were SATA @ 3Gbit/sec that would still be quite fast.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
Well maybe $20 for around 128gb but that is definitely stretching it.
I mean, I'm not often one to complain about moderation, but this is really a sad and depressing sign of how the only thing that matters in moderation is getting into the thread as fast as you can without bothering to come up with an intelligent thought first.
I mean, a comment that has one two words in it (counting the title!) that are nothing but opinion gets moderated "Informative?" All because it's the second post? Yeesh.
Posting AC so that I'm not being a signal-free karma whore too by way of this rant.
They're using SLC NAND flash which has a lower wear than MLC NAND but that doesn't mean there is no wear at all. It looks like a nice drive anyway.
have to wonder about the accurary of the following claim:
I have no problems with their claimed speed since frankly, if you run multiple smaller internal unit in parallel, you can pretty much get any speed you desire. But it's my understanding that the wearing out of the storage cells is a physical problem and in order for their claim to hold true, they've had to have done one of two things.
1. Made a fundemental break thru on SSD storage that doesn't have the wear problem.
2. Have enough storage and wear leveling to last for 5 years at maximum rated speed.
Re - your sig "How are sites slashdotted when nobody reads TFAs?"
My guess: Robot overlords, including those which are allegedly evil, allegedly non-evil, and others.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
Thing with these kind of prices is that you start off with the off the shelf price - if any - and then negotiate the real price. And this final price is - of course - confidential otherwise clients start comparing prices on the internet. If he would post the price they would directly point to the company that was paying the price and signed the confidentiality contract.
It's more the "with random writes" part that is impressive (but what size writes?).
Heck, I'd even pay $25 Canadian!
Be relentless!
Seems like the same massive advantage of an Application-Specific Integrated Circuit (ASIC) over general processors and even FPGAs that I see in video compression, a field I keep tabs on.
At one time I had wondered why a $100 camcorder could encode video in real-time, when my seemingly much more powerful desktop took hours. Answer: ASIC.
Some of you may be thinking, "Well, duh," but I am not an electrical engineer and thought it was intriguing when I first found out about ASICs.
Two problems:
1) They're bottlenecked by SAS, which, if they're using 3gbit controllers, probably won't go that much higher than ~500MB/s
2) Their cost is probably insane, if they're setting the upper bounds at $6000
By comparison, Fusion-IO claims 100,000 IOPS (not as high, but not far off) on their drives, and are about to introduce a new model for $895. They use a PCI-e 4x slot, which assuming v1.x, should give them about 10gbit/s (before overhead) to play with.
Also, Woz is their chief scientist, so bonus.
The newer version of SAS would bump up the interface to 6gbit, but then, PCI-e 2.0 would bump a 4x slot up to 20gbit/s.
In short, it seems to me that the future of super high performance drives is in PCI-e rather than SAS.
OR, the LSI 2108.
The 2008 is available onboard on several supermicro server boards. I've got one onboard on my X8DTH-6F
Any X8 supermicro board with a 6 in the model number, has the LSI 2008 (and any with a 3 in the model number has the LSI 1068e)
I'm currently using an add-on LSI 1068e since there isn't support for the 6Gb/s SAS chip as of yet, in my desktop unix "big cat" operating system of choice. The LSI 1068e works out of the box though. Bootable and everything.
Thing with these kind of prices is that you start off with the off the shelf price - if any - and then negotiate the real price. And this final price is - of course - confidential otherwise clients start comparing prices on the internet. If he would post the price they would directly point to the company that was paying the price and signed the confidentiality contract.
So, how much do they cost exactly?
The thing about these drives is that they are more expensive ($/GB) even than registered ECC DDR2/3 RAM, which obviously is going to be even faster.
That only means it would probably be better to use RAM for read-only applications. An application that needs to commit a write to a database, ensuring that the bits are actually written to a physical medium, will be be able to utilise flash rather than a hard disk. A lot of database servers would gain increased performance from such an arrangement.
Only worth about 10$? You're crazy, I'd pay up to 20$ for such a drive!
Heck, I'd even pay $25 Canadian!
Mod parent redundant. :)
XML is a known as a key material required to create SMD: Software of Mass Destruction
The secret: they're actually trying to price the 2gb model and scaling from there.
My Guess: /.'ers like pretty pictures.
Imagination drew in bold strokes, instantly serving hopes and fears, while knowledge advanced by slow increments...
The list price for Sun's 16GB flash drive (XTA7210-LOGZ18GB) is >$8,000 US. This is write-optimized flash for use in a ZFS hybrid storage pool.
E.g. The intended use is you use it as the "log device" (Sun equivalent to ext3 journal) to accelerate the speed of synchronous writes to your storage.
And it has a small fraction of the IOPs and speeds Pliant is claiming.
When they bring this to market, we can expect it to cost 15x to 20x that price per GB, easily, at least initially.
Not really since today's rate was 20 U.S. dollars = 21.6840012 Canadian dollars. ;)
Anyway I'm also Canadian, so he was willing to pay 5$CAD more than me, which is completely insane!
I recall once reading a sci fi story where a character in our future comes across old magazines and advertising. They remarked in astonishment and disbelief that manufacturers were actually allowed to make claims about their own products, since this was obviously open to utter abuse.
Sounds like a great idea to me, personally. Sadly, I don't remember the author, but Heinlein and Asimov come to mind as likely possibilities.
Of course, nowadays the manufacturers simply astroturf or hire PR firms (reviewers?) with some unrelated name to do the same, so I guess there's no going back to the truth.
It's not like the paper won't take the ink, or the internet the bits, if it's untrue...
A bit disingenuous, showing off the massive random IOPS of the SSD drive in your box vs. a non-SSD configuration of a storage array which can use SSD disks. "Kind of cheating" is an understatement. Your 8U of EMC disks became 10 disks somewhere here too. Anyone's 15krpm SCSI disks on a really good day may do something like 200 cache miss IOPS. 10 disks * 200 IOPS = 2000 IOPS. Well, yeah, finding an SSD drive that can outdo 2000 IOPS is trivial and NetApp, EMC, IBM, Compellent, HDS, etc etc will gladly provide you one if you want to compete on equal footing. It sounds like you're duping these people. But hey, it's not like that type of thing is uncommon in the storage industry!
No, nobody would buy Pliant SSDs at 15x the price of STEC SSDs since Pliant only has 5x the IOPS.
The $/GB of DRAM is misleading. Sure, 300 GB of RAM is cheap, but how much does the server cost that can hold it?
I think the point of the GP is not the life of the product, but the life of the company.
who cares if they have a money back guarantee or whatever if they aren't even going to be around?
Maybe nobody would buy them at 15x, but just b/c the improvement isn't worth it to them.
A new device with an N fold increase in IOPs against the previous model does not necessarily have a market value bounded by an N fold increase in value. If they are unique in the marketplace, and there is strong enough demand for the improvement, they can sell at a price greater than an N fold increase.
It's not as if you can buy 5 of the previous model SSD and acquire the same reduction in latency, when your writes are synchronous, you can't re-order (or pipeline/parallelize) your write ops.
Your IOPs get bound by the speed of your best SSD.
An application that needs to commit a write to a database, ensuring that the bits are actually written to a physical medium, will be be able to utilise flash rather than a hard disk.
Log files are sequential, so there is no speed benefit in using flash media instead of hard drives.
Generally drives in this class in the 70+GB range run anywhere from ~$10,000 - ~$15,000 each. I wouldn't look for this in your local PC shop anytime soon. These are still really HPC only items and are designed for extreme work loads. Not just booting linux or windows faster.
>you can't re-order (or pipeline/parallelize) your write ops
so this technology which attain super fast speed bundling controller chip channels can neither speed up that
You are using comparing HDD's in a Clariion to your SSD's. Not a very fair comparison. Why not compare your box to a Symmetrix V-Max with SSD's, or even a DMX with SSD's. What you've done is like sticking a Ferrari engine into a Lada and compared it to say a Ford Focus and saying that your's wins. Just because your engine is faster doesn't mean your product is better.
In an application where you need to ensure data is written to somewhere physical on commit the database will call fsync(). This can only be done 250 times per minute on a 15k rpm disk drive. That limits the database to 250 commits a minute. Battery backed or flash cache increases performance here - "The really enormous performance increases that have been found for update-heavy database loads come from another hardware enhancement, namely the RAID controller with battery-backed cache.... On a test involving a workload heavy on updates, with transactions involving relatively few updates, the introduction of this cache increased performance by a factor of 20."
Of course I meant 250 commits per second, not per minute. That's still assuming the best case of one transaction being written every rotation of the disk. Actual world results will probably be lower.
But I then would wand FOUR of them! ^^
Man, some people just deserve to be ripped off!
Any sufficiently advanced intelligence is indistinguishable from stupidity.
I stopped caring about speed. As long as it's fast enough to boot before I'm done in the kitchen and bathroom, and as long as it manages to stream HD movies and the like, it's OK.
I rather prefer a good ZFS pool and a set of reliable drives, that survive the first 3 years, but also the next 20!
Any sufficiently advanced intelligence is indistinguishable from stupidity.
Write speeds for a given rotational speed have been getting better over the years, and any real system is going to be using a battery backed cache anyway. With a properly tuned file system I don't see any benefit from an SSD for purely sequential writes. Add in cache on a controller card, and you don't even need that. Modern drives can cache the fsync() operations and perform them as a long single write.
All this talk of RAID is nonsense and doesn't apply to these drives. RAID stands for "Redundant Array of *Inexpensive* Disks". These SSD are probably bloody expensive.
The number of fsync's per minute depends a lot on the writes you are doing. If you are using a log-structured filesystem then fsync() can be handled asynchronously; you just don't return until the drive reports that it has committed those write (all writes are linear in LFS). The down side of this is that, until the garbage collector runs, your data is slow to read back. This isn't always a problem for a database. If you have a 4GB db then you can keep it all in RAM and stream the commits out to disk. Reconstructing it in RAM after a crash or power failure will be slow, but while it's running it will be fast.
I am TheRaven on Soylent News
That's why we need more development in RAM-based disk emulators. Much like the now-archaic Gigabyte i-Ram, I would kill for a PCI-E card that takes 8 or more registered RAM modules and spits out a bunch of SAS or SATA connectors to be Raid-0'ed, with battery backup. It would be cheaper than a high-end SSD and much much faster.
-Billco, Fnarg.com
So where do we meet up?
I'm late reading this article, but there *are* products out there that do exactly what you state. I can't recall any company names off-hand (and it'd sound too much like an advertisement anyway), but I think one of them was mentioned earlier by someone. The ones I've seen will take 6 sticks of ECC-R DDR2, and have a small external connector for power to maintain the contents of RAM while the computer is off. You're still limited by how many PCI-E slots you have in your servers (most 1U servers have 1-2 for instance), but it's a cheaper alternative to SSD.