In this blog I look beyond the obvious and try to find out why people and companies do what they do - and what it means for the rest of us. 
Sometimes I’m slow to pick things up, but it just dawned on me that when Quantum announced dedupe OEM deals with EMC and Dell a few months ago, they just put themselves into a very interesting position.
EMC is going to use the Quantum stuff in several of their offerings. Dell appears to be coming out with one product, at least to start. What’s interesting is that because all three folks use the same code in those products, they will interoperate. An EMC system will replicate to a Dell system which will replicate to a Quantum system. That opens up all sorts of interesting possibilities – which were lost on me previously.
First, it lets all 3 vendors make an “open” claim, and do so with some legitimacy. There are 3 places right now that you can buy a product that guarantees compatibility with the offerings from the other two. Assuming the code isn’t hacked at all (and I’m told it is not), that means if you bought a system from EMC but decided that you now hate the logo, you can replace it with a system from Quantum or Dell and it will just work. Or, you might buy someone with Quantum and you already have EMC, and you can just assume they will work compatibly. That’s interesting. They all have different packaging options and pricing structures, so in theory a buyer can shop between them and pick the right set of configurations and functionality for their own needs, knowing if things change, they have options.
In the mid-market, Data Domain owns the perceptive (and real) market value. I contend that in any break-out market, beating the incumbent victor with a “but mine is better” strategy is a fool’s errand. People are not interested in re-solving a problem that has been solved already when they must slay brand new dragons tomorrow. What does make this market potentially interesting however, is that while DD owns the hearts and minds of the masses, the actual footprint remains relatively small. DD has more than anyone (over 10,000 of these gizmo’s last I looked), but out of the millions of potential spots it could go, it remains a small percentage. They need to continue to grow fast to lock out the marauding armies from hacking away. Big giant OEM’s have the luxury of being big and giant, and as such, with a modicum of effort they can plop a lot of boxes in the world. While it remains to be seen if an initial buying decision will be based on the argument that de facto “openness” should be the top consideration – it clearly should be “a” consideration. History has proven that provided a choice in a market that is still in flux, with all things being equal the market will chose a “standard” that provides choice. It happened with the Cisco router. It happened with the NetApp filer. Cisco was the de facto standard because they connected to the most things (while the big guys tried to insist no one needed to connect to anything but their stuff). NetApp used standard hardware and NFS. Once the landscape is truly dominated by one player, then they have become the de facto standard and it doesn’t matter much if your solution is better, or more open, or makes toast for you. This market looks like there is still a little time to disrupt the Data Domain train – but not much.
I also have zero idea how aggressive EMC or Dell will be with this stuff, although it appears to clearly benefit their cause to be so. EMC has approximately 523 different dedupe offerings at last count, so I’m not sure the poor sales folk will be able to figure this out without some corporate direction. Dell is different, as they appear more focused and live in this market slice. Their issue is going to be Data Domain’s sales force – which is big and good and getting bigger and gooder. Yes, I said gooder.
At the very least, it will help Quantum get over Data Domain FUD. If I’m Quantum and DD tells my customer “whoa, they are going out of business and they are convicted felons and they hate babies”, (ok, they don’t say most of that), I say, “look, we aren’t going out of business – that’s absurd – BUT let’s just say we did all perish in a Hawaiian volcano incident, you can continue to buy and get support from EMC or Dell. You don’t think they are going out of business do you?” followed up (or proceeded) by, “Dell and EMC are fairly large companies with fairly comprehensive financial and technical acumen – and they chose to OEM from Quantum. I’m fairly sure we didn’t “bamboozle” both of them. They know what they are doing.” I’m not that smart, but I’m pretty sure I could leverage the Dell or EMC relationship to put a halt to that one.
Related posts:
blogs
As you can imagine I have an opinion on all of this. To make an open standard you have to publish your approach such that everyone can work with it and interoperate. Quantum shows no signs of doing this and infact enforces their patents with anyone that may violate it. The alternate for Quantum it o OEM their code to all vendors yet the rumors are they have an exclusive with EMC and EMC allowed DELL entry to that exclusive as EMC uses DELL to take storage to the mass market while EMC focuses on the enterprise. So what you have in the end is two vendors shipping their own hardware running Quantum code. That does not make a standard. It does make for two nice distribution partners if both actually pay enough attention to the products and space. EMC has Avamar, DL3D/4000 (Quantum( and now there seems to be an appliance that competes with DL3D/4000 by putting the Avamar dedupe in an appliance which would suggest that EMC may over time replace Quantum and that Quantum is a stop gap measure. Name the instances where an OEM software model has worked. Almost always it is a stop gap measure to buy time while the OEM builds or acquires their own solution Where is Riverbed’s, Data Domain’s or before they were acquired EqualLogic’s OEM? They built their own brand and channel. We believe Quantum will be engineered out in the end and they have made a huge stratetic mistake. This market is less than 1% penetrated and the playing will shift many times before it is over. It will be fun to see how it really turns out and to read all of the banter along the way to see who was truly right or wrong.
Bill Andrews
ExaGrid
Wow this creates a whole new definition for Data Storage Fuzzy Logic. Because Quantum, Dell, and EMC all sell the same box this becomes the basis for universal dedupe? A few years back when HP, Dell, and Overland all sold the same tape libraries did that create a universal standard too? Let me think..Um, no.
DrDedupe
—–Well, as suspected, this blog has generated a whole bunch of chaos – which generally is a good thing. I can’t tell you the last time I had a blog that generated so many behind the scenes “I don’t want to be identified” emails, as a matter of fact.
There are a bunch of issues, so let me start in decending order of relevance.
1. Many people, such as Larry above, have focused far to heavily on the left brained logic of “why”? or “how”? or “No way does EMC ever really do x, y, or z!” To you all, I shall re-emphasize the real point again for the right half of your brains – it doesn’t matter. The point was that IF the stars aligned, it is POSSIBLE that the collective market power of two mongo dominant players like EMC and Dell could have the strength to create a DE FACTO standard (defined as: the person with the most stuff installed, who owns the hearts and minds of the buying public) because there would now be lots of systems installed all using the SAME Quantum code base. Period.
A few years ago Dell, HP, IBM, etc. all starting shipping machines with Intel processors in them, although there were PLENTY of other choices – and low and behold a DE FACTO standard emerged. The world has not looked back.
2. I got several rants back about how just because they used the same code – it did not mean it would be compatible. I will only say this – I am not naiive enough to believe in fairy tales, nor that either entity – if they so chose – for whatever reason, good or evil, could not easily make their systems completely INCOMPATIBLE with relative ease. Why they would do that I have no idea, but could they do that? Absolutely. Conspiracy theorists abound.
3. There were plenty of heated, impassioned plea’s to get me to realize all sorts of technical death traps with either the Rocksoft/Quantum code base, or the security of all de-dupe systems, and even the CERN/Google supported “silent data corruption” issues that have little to do with de-dupe and much to do with data in general. Most of the points were excellent – but misguided in this context. With absolutely no disrespect intended, I say “who cares?” Is there really anyone out there that truly believes no matter what they have or what they buy, it is actually going to work perfectly? Of course not. Everything breaks, evertything sucks, and nothing is as advertised – but get over it. That has ALWAYS been the case in this business – but somehow we spend a hundred billion dollars a year or so anyway. My point is while maybe all the issues are totally accurate – it does not have significant meaning in the market as a whole. We will still spend money and we will still market until we drop. Perception is reality my friends.
Business is not a place for the righteous – it is a place for the right. You are not going to stop running IT because your disk is going to eventually become corrupted. You aren’t going to stop doing backups because EMC might screw with the Quantum code. Individually you may make a decision between one choice or another based on your opinons, but a choice will be made.
Bernie Madoff screwed so many people out of so much money he will burn in hell, I am sure – but does that mean people will stop investing?
All of your points are good my friends. You may be absolutely correct – but if you look at the bigger possibilities at play, you might see the forrest through the trees.
Final Semi-Related Point: The ONLY financial reference made at EMC’s analyst day was when Joe Tucci commented on how EMC has overtaken Data Domain in the dedupe space as EMC did $90M last quarter (mostly Avamar, which really isn’t the same thing whatsoever) and DD did $85 or so. A $15B company and the CEO is talking about $90M? A mistake from a normally mistake free company at such events.
First, why bring up a number that low, comparatively? You want people to focus on that small slice of the market? Second, and maybe the biggest issue in doing so, is that instead of adding value to EMC’s overall position (the rest of the messaging around virtualization, etc. was very strong) – it instead RAISED Data Domain’s value and interest. Where this might have been a little, niche sort of market, EMC has turned on a large spotlight on it now. If I’m Data Domain, I’d be doing cartwheels. You can’t buy that kind of love.
—–Steve
P.S. – Larry, you spelled your own name wrong I think, which is awesome.
Just so you have the facts straight: Of the EMC Q4 numbers analyst estimates, which I believe are substantially correct based on my own independent research, were that QTM based dedupe was $55 mill of the reported $90 mill. Last week EMC agreed to lend QTM 100 mill. I wouldn’t underestimate the EMC commitment to QTM or their success selling the DL series.
—–Who’s underestimating? I think lending Q $100M pretty much proves the point, no? EMC isn’t exactly known for humanitarian aid – they did this because they want and need Q to keep on keeping on.
I can’t win on this one. Earlier I was accused of being on the take by the monkey for even suggesting this (he is wrong on all counts), and you seem to feel I’m somehow dissing the mighty Q! Yet, if you all would read a little slower, you’ll see you all actually agree with me!!! – Steve
Steve, I didn’t mean that you were underestimating…the “I” was rhetorical. So here is some new validation of your point: In a con call with senior debt holders on Monday Q informed the group that in ADDITION to the $100 mill…EMC will prepay $40 mill in royalties to Q who will pay that to the senior lenders on the condition that they amend the bank loan to allow Q to buy back any converts that remain after the tender offer. So $100 mill loan plus $40 mill pre-paid royalty…Q’s debt goes down, it’s partnership with EMC is further strengthened and Q’s stock looks to be de-risked. And by the way I hear EMC is growing their tape biz with Q as well.
I think I’m closer to the Storage Monkey than you on this one. His problem was that you used the term “open standard” in both the title and body of the post, and there is no such thing as an open dedupe product, and I doubt there ever will be. Towards the end of your post, and in your comments, you seem to be making the DE FACTO argument, which is very different than being an open standard.
The DE FACTO standard in _VTLs_ (NOT DEDUPE) is FalconStor. Thanks to EMC, IBM, Sun, COPAN and others, there are far more FalconStor-based VTLs out there than probably all other vendors combined.
The CURRENT DE FACTO standard in DEDUPE is Data Domain. Many customers I talk to use Data Domain and the words “dedupe” almost synonymously. They’re not quite (nor do I think they ever will be) a Frisbee/Tivo/Kleenex, but they’re pretty stinking close. Like FalconStor in the VTL market, there are probably more DD boxes/customers than all other vendors combined.
The question is will EMC’s sales engine do for Quantum what they did for FalconStor? Will they supplant Data Domain as the DE FACTO standard in the dedupe market? They’ve got a long way to go before that happens. While they are close to their quarterly numbers, DD’s been selling that box for several years and they’ve got thousands of customers that EMC/QTM don’t have. They’ll need to significantly surpass DD’s sales numbers for a long time before they catch up to DD. But anything’s possible when you get the Hopkinton sales engine on your side.
Now, as to them being an “open standard?” No way, and I don’t think anyone ever will be. To become an open standard, you have to actually publish your standard. A dedupe vendor would have to say “you can either replicate to my box or accept deduped replications from my box as long as you adhere to this API.” And usually when that happens, the STANDARD becomes the standard, not the company. Perfect example: Sun’s NFS. They invented the standard, published it, and it’s the standard. But no one would accuse Sun of being the standard NFS server. If anyone gets that honor, it’s NetApp. The same goes true of IBM & Microsoft with SMB/CIFS. These are “open standards.” That has nothing to do with being the DE FACTO standard.
And since there is NEVER going to be a standard like THAT in the dedupe space, the answer to the question that your blog post’s title asks is “absolutely not.”
(Of course, as you French people say, “ne dis jamais jamais.”)