RSS
Email
Print
ShareNetApp Buys Data Domain
Posted by George Crump on May 21, 2009
NetApp is making a serious move into the disk backup and disk archive market, laying $1.5 billion on the table to acquire Data Domain. If you are a regular reader of my blog on Byte and Switch you know that in general I am skeptical of most technology acquisitions, but this one may be an exception. It has the potential to give NetApp an incredible advantage in the environment -- they just have to manage the integration.
NetApp's Responsibility -- Do no harm, but do something
NetApp of course positions this as non-disruptive to their current offerings. Really it is too soon to tell what will happen and to some extent not fair for us to expect NetApp have all the answers. What we have to hope is that NetApp lets Data Domain be Data Domain and where Data Domain is a better fit, have the NetApp products step out of the way, at least until the Data Domain technology can be integrated in.
The best advice is to let Data Domain run as a stand alone entity for a very long period of time and then slowly in the background pull some of the Data Domain technology into the traditional NetApp products.
Deduplication is not a feature -- its a market.
Comment by Funny on May 21, 2009 12:57 PM
I wonder what this says about NetApp's dedupe "feature". Uh huh, just turn it on!
Cackle!
Reply to this comment
Comment by AnOutsider on May 21, 2009 3:27 PM
As somebody who's been through an acquisition before, at the company that got bought, here are my thoughts.
Data Domain's product may have won, but their people need to be careful they don't come out on the losing side of this acquisition. The people at the company that's bought will always be viewed as outsiders, with limited opportunities in the parent company. If a choice position comes up, who's going to get preferential treatment, an established NetApp person, or an "outsider" Data Domain person? The NetApp person every time. So a lot of good Data Domain people will end up leaving because they're not going to be able to advance.
I don't see their senior management team and executives sticking around, either. A team that's built that good a product suddenly playing second fiddle and taking their marching orders from somebody else? Nope, I don't see that happening. They've already made their money, they're going to cash out, leave, and be top dog at some other venture. One year from now, I bet all their executives, senior management, and founders will be gone.
This isn't a merger, it's an acquisition, and Data Domain was the one that got bought. NetApp can try to put as much positive spin on things, and try to keep hands-off as much as they want, but ultimately they'll make the rules now, not Data Domain, and Data Domain will have no choice but to march to those rules. And Data Domain has two things going for them that I think are going to make it difficult, it not impossible, to do that: a very strong independent streak, and success in the marketplace.
The last thing on the financial side is, NetApp isn't growing by innovation, they're growing by acquisition. To me, that means they've entered a period of stagnation. Data Domain's stock had a lot of upside to it if they had remained independent. Not so with NetApp. Sure, Data Domain might enjoy a $8 premium per share right now, but I bet that's the only growth they see, as I don't see NetApp being worth much more than they are now, except for small incremental growth (that's just my gut feel, I'm no financial analyst).
I'll reserve total judgement for about a year from now when we see how things shake out. But I can't see how people in Data Domain are viewing this as anything but a sell-out. I bet one year from now, Data Domain will be a shell of its former self. With the economy hopefully picking up, I bet a lot of their good people jump ship.
I guess I'm just not a fan of acquisitions based on past experience, so my view is certainly biased. Unless a company was formed with the express purpose of being acquired, I just don't think they work out well for the people there.
Reply to this comment
Comment by Industry Insider on May 21, 2009 4:33 PM
Good news for Data Domain. This is a innovative company that came out of nowhere and grew in a very short period of time. By offering solid products and aligning themselves with other industry leaders (along with spending some cash on marketing); they were able to make a name for themselves and sell for a great price in this very uncertain business market.
Reply to this comment
Comment by Guerre on May 21, 2009 9:57 PM
George, don't lose sight of the fact that the vast majority of data stored on RAID systems (on a PB basis) does not benefit from de-duplication because it has no redundant information. Video, medical, CCTV, scientific, and any kind of scanned imagery does not benefit from de-dupe unless you have a bloody-minded IT staff that insists on using the tired paradigm of tape emulated backups (yes this is common in a data center but exceedingly rare in the PB-scale disk farms).
Reply to this comment
Comment by Chris on May 21, 2009 10:40 PM
I guess netapp finally realized how hard dedupe technology is actually to implement... about $1.5B as hard.
They kept insisting they had dedupe on all tiers, and this shows that marketing with out technology can only go so far.
I'm glad they got smart up and bought datadomain.
Now the trick is to implement good dedupe for online storagebecause their ASIS stuff just doesent work
Reply to this comment
Comment by George Crump on May 22, 2009 10:15 AM
Guerre: Very good point and I couldn't agree more. In fact in a recent article. I write how much more important compression is on primary storage than dedupe. See: http://tinyurl.com/p3acdp
Reply to this comment
Comment by HappyDDUPshareholder on May 22, 2009 1:53 PM
OK. I agree they bought a better mousetrap for backup, but archiving? Come on. The DD solution still has major scalability challenges for large enterprises. If the model is to sell boat loads of little DD boxes, then they have the right solution. But, I'm not sure that's really what an enterprise wants.
DD did a great job of bailing out before they ever really had to fix that problem!
Don't you think there are better techologies out there?
Reply to this comment
Comment by Mike on May 22, 2009 2:05 PM
LOL! "Deduplication" is very much, just a feature. Because it is a relatively new innovation, it has had a transient lifetime as an independent market, but it is fast-tracking to an embedded feature. (There was once a "market" for software video playback and object-oriented databases, too)
NetApp primarily bought DD for the brand-awareness and customer-base; there's no IP in DD suite that NetApp didn't have already.NetApp's chronic problem is not technical innovation but instead market development -- something which DD did very well and now gets rewarded for.
EMC continues its strategic shift to a software & services company, and getting away from bent metal -- there's no compelling reason for them to do a me-too in this space, for them it is all commodity.
Reply to this comment
Comment by George Crump on May 22, 2009 2:06 PM
Happy, When it comes to archiving, my opinion is the focus changes from speed of injest to other things like scalability (as you imply) but also retention, availability, audit-ability.
Reply to this comment
Comment by Carl on May 22, 2009 2:08 PM
I always sucks to get acquired, but your conspiracy-theory mindset is laughable. Your attitude will be the biggest factor in you getting laid off, not any broad conspiracy.
Reply to this comment
Comment by joe on May 22, 2009 3:42 PM
How much you wanna bet this came around when Commvault added dedup to their backup software (not just dedup to any disk but to tape as well) in April 09 (without having to have a seperate appliance like DataDomain). With that addition they became competitors to DataDomain instead of working together. This may be a move on both sides (Datadomain and NetApp) to deal with Commvault's recent addition of a very useful dedup solution that they didn't have before which is only about a 20K addon compared to 100K or so for a seperate appliance.
Reply to this comment
Comment by HappyDDUPshareholder on May 22, 2009 4:33 PM
Joe
Interesting point on CVLT. But that's a 20K premium for something that many storage vendors already have built-in for free. Look at Permabit for example. Dedupe is just part of how the product works, it doesn't cost extra. And for
So, back to NTAP. $1.5B for a feature that others already have (including themselves)? I guess their own stuff was really that bad??
Reply to this comment
Comment by CVLT user on May 22, 2009 7:36 PM
joe you must work for commvault CVLT. but i have been using it for 3 years now at my current company, and i used it in 2001 at another company. it is still the same product that it was in 2001. the dedupe stuff works fine, i paid for it and i like it, but commvault backup is still the same old antiquated way of doing things as before, even if they changed the name from galaxy to simpana, and the archive and replication is impossible to do without a huge amount of work and cost upfront. i am seeking alternatives, but i wont plug other vendors here. i doubt anyone merged because they feared commvault or their commodity dedupe. commvault only has a decent backup solution. dedupe is commodotized already, soon the current paradigm for tape and disk backup will be as well.
Reply to this comment
Comment by backupmonkey on May 22, 2009 11:03 PM
The CVLT product is in no way going to be as successful as a ddup, exagrid or emc dl3d type appliance based solution. The deduplication ratios are half of what the appliance vendors provide, based on CVLTs own admission (10:1 versus 20:1 or better for the others, due to much larger block sizes) which means more disk at the same retention, and more bandwidth for replication (crucial for tape elimination). The cost isn't just '$20k' and you're done, because the sheer amount of processing required by deduping data is going to require new (and maybe more) media agents/backup servers. Just look at how much hardware is stuffed into the ddup boxes, and their performance is barely better than a tape library at best. Not to mention you still need to buy disk! The appliance model will stick around for deduping backups because it makes sense; offload the heavy lifting to a dedicated device and let your media agents manage backups/restores.
Reply to this comment
Comment by backupmonkey on May 22, 2009 11:23 PM
With regards to your assertion that de-dupe is a market and not a feature, George, I would challenge your position.
De-duplication has become ubiquitous; certainly it's prevalent in storing backups (DDUP, QTM/EMC, Sepaton, ExaGrid, IBM Diligent, etc), but it's migrated into the primary storage arena (NTAP), backup software (CVLT, CA, Symantec, EMC Avamar), WAN acceleration (Riverbed)... and these are just a few examples. It IS a feature now in, I would argue, backup appliances (VTL or NAS based, take your pick) and backup software, for just two markets that are relevant to this site/discussion. Look at Overland for a case in point; how much success did they have with the now-dead REO-D after an arguably successful run selling lots of non-dedupe REO VTLs for a few years. Not enough to sell many or keep them solvent without outside money... Any 'real' player in backup appliances has de-dupe as a feature, but it's not the only one in the product. Replication, reporting, data integrity checking, etc etc are all features of these types of products, as is de-dupe. De-duplication is great, but it's just an enabling technology. It allows you to replace tape with disk at a much lower cost than would otherwise be possible, and eliminate or reduce offsite tapes by WAN-friendly replication; that's pretty much it! It's no more a market than a car engine - the engine is required for the car to become mobile, but the market is for the automobile itself.
I am not sure yet if NetApp's move is going to be a good one long-term or not. Certainly they've added a couple thousand net-new customers to their install base, and picked up some additional channel partners as well. They have given themselves an easy entry into the backup storage market, as neither of their offerings (a-sis or the VTL) are particularly good ones. But, they are a player that sits higher up in the enterprise food chain than DDUP has been, and the DDUP product line is not particularly enterprise-ready, regardless of their marketing messaging. I'll be interested to see if they try to continue the trend DDUP has been making lately to take the DDUP product line higher up in the enterprise, or use the DDUP install base and existing product to get more of a foothold in the mid-market.
Reply to this comment
Comment by George Crump on May 23, 2009 1:10 AM
backupmonkey, first very good comments. I think in the piece that I said that its NetApp that does not see dedupe as a feature because I don't think you would pay $1.5B for a feature. That said, I think the answer is somewhere in the middle, maybe it is a feature that is so important (or hot) that you can make a market out of it? While dedupe has become almost ubiquitous, clearly they are not all created equal. Some vendors are still struggling with their dedupe implementations. Data Domain's seems pretty flexible and solid.
Also your point that is has to be more than just dedupe if very well made. All the backup appliance has to have replication, reporting, data integrity, performance and scale.
Reply to this comment
Comment by George Crump on May 23, 2009 1:16 AM
Joe,
I have a lot of respect for Commvault's product and their strategy. That said one of my complaints with their current presentation of the capability is when the do a price comparison to dedupe appliances they do not include the total cost of the software. They assume you already own Commvault and certainly if you are a Commvault customer this is accurate and fair. But if your not, you have a fair investment of dollars and time ahead of you prior to seeing the ROI from their Dedupe.
See my prior entry on backup conversions:
http://www.byteandswitch.com/storage/backup-recovery/backup-conversions.php
Reply to this comment
Comment by George Crump on May 23, 2009 1:18 AM
CVLT User, If you are comfortable please contact me privately, I'd really appreciate the opportunity to discuss your experiences first hand. It can be on or off the record.
Reply to this comment
Comment by Joe's boyfriend on June 14, 2009 2:30 PM
Mike,
How does crow taste. If Dedupe was a feature, then it is the most expensive feature ever. Sure it is a feature for all the storage companies attempting to jump on the DDUP bandwagon, but in reality it is very much a product category as witnessed by EMC's own bid.
Reply to this comment