Sitting on the exhibition floor at Powering the Cloud, you sometimes wonder how many of these companies will be still here next year: the sheer volume of flash startups is quite frightening – I can see five of them just from where I am sitting. There's Pure, Tintri, Nimble, Fusion-IO, Violin – but an ever-expanding number are …
By the time any of these startups make any significant money, the market will have moved. We won't be using flash with all it's issues, we'll be using memristor or PCM type technologies and one's that we'll use in a way similar to conventional memory. This is stop-gap technology and they've only got a few years to monitise it.
I wonder if you just don't understand how much more life flash has to it, or how far from real world use memristors or PCM are... or if you know and just don't care because work in the marketing department of HP?
I mean HP has given up sliding the memristor availability date forward to always be "some time next year" and have now moved it several years out. Which basically says they have no clue at all when or even if it will be commercially viable.
As far as using any of these things as if they were "conventional memory", that's gona happen for a count on your hand number of applications, there is a difference between using memory to build storage, and using storage to build memory. I understand the later idea is getting lots of buz that doesn't make it a good idea.
Phase Change Memory (PCM) has been shipping in Nokia Mobiles for a while now, and there are good reasons to believe that design choice will be mirrored in other consumer devices over the next few years. If that does happen, then the economies of scale, along with significant interest in PCM from the likes of Intel could drive the costs of PCM below that of MLC flash in a relatively short timeframe (shorter than the time many of these start-ups have been around). If so, building a strictly flash optimised technology base might seem kind of short sighted, then again, I don't think most of these guys are thinking much beyond a 3 year time-frame.
Right and there are no flash economies of scale?.... oh wait what is that other chip in the cell phone... its a NAND flash chip. A 64 Gb per die NAND flash chip... that Micron PCM chip next to it?... 128Mb, 50X smaller, it might get better, it should catch DRAM but its not going to catch NAND, different cell sizes. So kindly take your unsupportable FUD and sling it somewhere else.
And it would be horrible to build something really awesome today when random people are claiming that in a time frame greater than the replacement schedule for the product that maybe something better will come along, and it will obviously be so much better, but apparently if you optimized for one kind of memory technology none of that effort would be of any use in any other memory technology, assuming of course that out of nowhere this other technology instantly replaced NAND, which of course it isn't going to.
The bigger problem with FUD slinging like this is that if you actually looked at even the public semi roadmaps you would know your vision of the future isn't, but you just want to scare people into not buying something better now. Not impressed.
Choose your Flash vendor with care
Yes, many startups hoping to be bought. But that's the wrong motivation if wanting reliable Flash storage that does as promised by over enthusiastic marketing. My advice is look seriously at what EMC are doing with new product XtremIO. This is not JBOF (just a bunch of flash), its an intelligent array hardened and supported by a major vendor. Great inline deduplication feature that reduces the storage footprint and cost per GB. Great for VDI and server virtualisation.
This flash array will be around for years to come while other have passed away.
Re: Choose your Flash vendor with care
You forgot to quote the price of the array and your contact informations for potential buyers….
Reality is, XtremeIO is NOT yet on the market, while some of these startups have more than two years of selling and many happy customers. Sure they are startups with all the possible risks (basically one, to disappear living a customer with a dead system), but the fact some of them are selling means there are customers willing to buy those solutions.
I do not agree at all with the article, too much generic and vague. As always, it all belongs to choices: there are customers trusting the big vendors and never switching to a startup, but also others that see a real and big value in those "slight differences" and go out buying them right because of those features.
Also, not all solutions are "dual head + linux + zfs", this is (sorry) a really dumb statement. There are plenty of really different solutions out there, scale-out systems, object storage in the backend, and others. And this means mainly two things:
- freedom of choice for customers, that have more opportunities to find the best fit for their needs instead of choosing a "almost right solution with some drawbacks"
- a push to "the big ones" to accelerate their development/selling cycles
Surely some of them are out to be acquired. But when I see quotes about the big ones being EMC VNX (ex clariion...), HP 3Par, Dell Compellent, someone is forgotting they were aquired exactly as the new ones would possibly end up.
Or is it a startup is cool only when it gains a famous brand logo on their chassis?
Wait a doggone minute there...
One correction ; Netapp haven't "given up" on tuning with flash. NetApp was one of the first to employ Flash based caches in 2009 with FlashCache; that's 4 years ago. We still ship and use both caches and SSDs for performance in the latest Ontap 8, along with the all-flash EF540 and the soon to be delivered FlashRay.
It's interesting that you think that the startups you mention all want to be like NetApp when they grow up -- or perhaps that's what they told you? -- for it's certainly what they want you to believe.
The truth is they don't want to be like NetApp. No way. Not in a month of Sundays. That would mean putting in some hard work for the long term, and none of them are in it for the long term.
NetApp has got a whole set of experiences and innovation in this space built over decades of investment and learning that these new guys on the block just don't have. And won't have and don't want either, since their business model (as you note) is a desire to get bought and make the principals some money.
Lots of excitement...
...there is a lot of cool things coming out of the startups, but when I think long and hard about it I really can't see myself wanting to move away from the big players in all-flash or hybrid-flash systems. As we're currently discussing a storage refresh and working with multiple VARs and manufacturers, we always seem to come back to EMC (VNX with FAST), HP (3PAR with AO/DO features), Dell (Compellent with Data Progression across SLC/MLC), IBM (Storwise with Easy Tier), and NetApp (FAS with Flash Cache and Flash Pools).
We always get ramped up by the startup storage vendors only to be let down when we really look under the covers, mostly because the products are designed to do one thing well. It sure makes for a boring experience but at least if you have the big guys in your corner and a local SE you can choke, they'll try not to f#ck it up too badly (and they don't mind picking up a lunch tab every other month or so). If a startup gets it wrong, they might not even be around long enough to take your anger out on them.
Maybe the next time we look at a storage refresh, one or two of these vendors will still be around and worth taking a look at but until then I'm going to run safely into the arms of the big guys even if they run train on my budget for maintenance and support services.
who's gonna buy them?
IBM just bought TMS, EMC bought extreme, HP bought 3PAR, Dell bought compellent, NetApp bought LSI and has their flash ray thingie coming.. Cisco bought whiptail.
I think Cisco would be at least somewhat worried about pissing off NetApp/EMC even more (and thus losing some of their steam in their efforts to leverage N/E to get into accounts etc) by competing more directly with them. Perhaps if/when UCS growth tails off they will get to a point where they don't care and just go full steam into storage too.
The rest of the folks seem to be pretty happy with the various companies they acquired and probably are not likely to make further acquisitions anytime soon, IMO of course.
HDS, and perhaps Fujitsu(?) seem to be some of the only major folks left that haven't bought anyone recently ?(I don't count BlueArc since that is just a gateway). Anyone left out there?
Oracle I guess, Pillar is a door stop so I don't count them. They also have ZFS which I've seen almost nothing but complaints from customers on how poorly those systems are built/bad support etc. Though Oracle is all about the engineered systems, and putting the storage intelligence in the application layer where it's more efficient (Exadata), so I suspect not a high likelihood of Oracle going after these startups, unless it's *really* cheap.
I suppose there is a chance one of the cloud service providers may see value in some of this tech and acquire it (with the intention to use it entirely as a service provider role not re-selling appliances to customers). Though to-date it seems the cloud players are all about building stuff in house, no matter how crappy it may end up being.
Might we see something new here? Might some of these flash startups merge with each other, or acquire each other? That would be sort of interesting.
Something different from an emerging player
I think you missed one major point when mentioning Tintri vs the others. Tintri is different as we focus on VM environments only and change the point of management away from the traditional LUNS/VOLUMES/RAID GROUPS/ ETC to the VM. (For the record, yes I do work for Tintri).
This is something completely different then anyone else in the market space.
Providing VM level visualization and management (See end to end how your VM is performing, VM level snapshots, VM level QOS, VM level replication). Tell me anyone else in the industry who is doing this today. Designed from the ground up for VM's. It's not a respin of ZFS with a pretty GUI, or a we can do that to product.
So I make the statement that Tintri is different.
I look forward to more spirited conversation
Author is mistaken
As pointed out by a Tintri employee above, not all these systems use ZFS. Tintri does not, in fact most do not. There is one (Tegile) that does, but has added significant enhancements above and beyond the original open source ZFS release. So casting aspersions on the entire startup industry as ZFS based is inaccurate.
This market has a lot more play in it. Yes IBM has a reliable and fast product (the TMS), but it has absolutely no data protection features (snapshots or replication). That isn't so great. NetApp is shipping an LSI based design that was never designed for flash (the EF 540/550), which does have snapshots and cloning, but again not great performance since it isn't flash optimized. HP's 3PAR 7000 systems now have an all flash model, with great features, but performance is "meh" because again, not designed for flash. EMC is _finally_ bringing XtremeIO to market, but it to is lacking in features and somewhat in performance.
The list goes on. So again, there is a lot more play here and just because the big guys are shipping something that is all flash today, doesn't mean it can stand up in the near future. This market will continue to change DRAMATICALLY over the next 3 years.