Chapa shoots at Mr. Backup and misses

My long time friend David Chapa just wrote a blog entry called “Mr. Backup makes predictions for 2010. He shoots and misses!“  I can’t let a title like that go without a response, so, here goes.

He took some pot shots at an article that I wrote of predictions for 2010.  First I want to say that I hate predictions articles. Who the heck knows what’s going to happen this year?  Who would have guessed what happened last year with the whole Data Domain/NetApp/EMC deal?  It’s less than two weeks into this year and Google’s changed the game again with the introduction of their cloud storage service.  But TechTarget asked me to write one so I did.

His first complaint is that I predict target dedupe to grow and that’s a no-brainer.  It may be so, but there are some vendors predicting otherwise, so I thought I’d at least say that.  He then hints at why shouldn’t a customer just put their backups on a NetApp that has dedupe built in it?  How do I count the reasons?

1. It will inevitably end up costing them more. Their dedupe is less robust (read lower dedupe ratio) than other dedupe systems, it’s only within a single flex-vol (it’s less global than data domain), etc.  That translates into buying a lot more disk — and nothing’s free.

2. The performance is less than other dedupe systems.  If you’re going to be a target for backups, you need massive performance.  NetApp’s got some nice systems, but thousands of MB/s they don’t have.

So you would get more performance and more storage for less money by buying a real target dedupe system. (NetApp knew that, which is why they tried to buy Data Domain.)  I like NetApp for a lot of things, but as a target for regular backups (e.g. NetBackup/NetWorker/TSM/CommVault, etc.) not so much. I didn’t mention NetApp in the target dedupe section because nobody else sees them as that either.

He then asks a what if. Sorry I don’t do what ifs.  What if NetApp had won Data Domain?  Would I have to respond to this blog post?  What if monkeys fly out of my butt?  I don’t know.

He complains that I didn’t mention NetApp in my source dedupe section. Um, cause they don’t HAVE a source dedupe product? I listed all the vendors I knew of.  NetApp doesn’t have a source dedupe product.  He says that they have other ways to back up VMware.  That wasn’t what it was about.  It was about source dedupe — and they don’t have a product in that space.

The next complaint is that I left NetApp out of the cloud backup section.  Again, cause they don’t have a cloud backup product.  Hello?  I know you guys are pushing your storage as the platform for the cloud. I even know that one cloud storage provider is using it as their storage.  But that paragraph wasn’t about cloud storage; it was about cloud backup — and you don’t have a cloud backup service.  And as to cloud backup — who got mindshare last year?  That would be Mozy/EMC.  Sorry.

And then he thought it was really weird that I mentioned the vStorage API and that it would really change things.  Let’s see… VMware is HUGE, right?  Like 95% of datacenters huge, right? And VMware backup is a pain, right?  And vStorage is finally poised to fix that, right?  (Well, except for the whole VSS thing that I didn’t know about when I wrote the article.)  That to me seems like something poised to really move in 2010.  What’s so strange about that?

And again I didn’t mention NetApp when talking about things that would change VMware backup in 2010.  That’s because (while I actually think they are one of the two best ways to back up VMware) I don’t think they’re really going to move much more of it in 2010 than they did in 2009.  I’m not saying they won’t sell it; I’m just saying it’ll be a lot like 2009.

Now I get it. He didn’t like that EMC, Mozy, Data Domain, Avamar, and VMware (all EMC products or companies) all came up as game changers in 2010, and not one NetApp product made the list.  Sorry.  I like their products, but I don’t see any games being changed in 2010 because of them. Sorry I have to say it like that, but he asked why.

Written by W. Curtis Preston (@wcpreston), four-time O'Reilly author, and host of The Backup Wrap-up podcast. I am now the Technology Evangelist at Sullivan Strickler, which helps companies manage their legacy data

2 comments
  • hey Curtis! Wow, that didn’t take long ๐Ÿ™‚

    Now I don’t know that I would call it complaining, but rather curious.

    And I wouldn’t call them pot shots either, I just felt that you missed mentioning a company, NetApp, who does have technologies in those categories that are compelling.

    As I mentioned in my blog (http://blogs.netapp.com/barandgrill/2010/01/mr-backup-makes-predictions-for-2010he-shoots-and-misses.html), I have a great deal of respect for you and your work and it was in that light that I wrote the blog and it is in that same light I respond.

    Now as for #1 – the target dedupe. What I said in my blog was “what’s to stop a customer from carving off disk..” – not that they had to go out and purchase it, but they already had it in house. My point is that if a customer has it in house and they can use dedupe for free, take advantage of it during this economic downturn.

    As for the “what if”, taken in context you made a statement in your article,

    “Other reasons for target dedupe’s increased adoption have more to do with some of the companies that are pushing it. EMC Corp. acquired Data Domain last year, spent $2.1 billion dollars, and their sales teams have been given marching orders to show the stockholders that this wasn’t a bad move. “

    My question was where is the “than” – other reasons for target dedupe’s increase…has more to do with companies that are pushing it, “than”?

    This wasn’t a pot shot, this was trying to dig further, is there a than? Such as more to do with companies pushing it than the technology itself or more to do with companies pushing it than customers building this into their 2010 plans? It left me hanging. So then I asked what if EMC hadn’t bought DD and continued with QTM – would that have still been a true statement – would it still have more to do with companies pushing it or would it have been something else. I understand you don’t like “what ifs”, that part of your article really intrigued me and wanted to hear more.

    So in the source dedupe portion, I wouldn’t have touched that but you put virtualization and backing up virtual servers in there and wrote about how source dedupe helps the virtual environments. Okay this is where I started turning my head to the side because this is where I thought you were going to give kudos to NetApp for how we are changing the game with virtual environments and how VMs sitting on NetApp storage with dedupe is probably one of the most efficient solutions and coupled with NetApp’s DataMotion how customers can move not just one VM but hundreds or thousands at a time with integration into vStorage.

    Now on to cloud backup…I don’t think that I’ve ever told you about companies like BT, Rackspace and Joyent (http://www.netapp.com/us/company/leadership/cloud-computing/differentiate.html) and how they use NetApp as their cloud infrastructure.

    I know you were talking about more home based backup products like Mozy, Crashplan and Carbonite and you’re right NetApp doesn’t have a household or home product like those but we have been selected by very large companies to be their cloud infrastructure which includes data protection for their customers.

    So for the vStorage API – the only thing I thought was weird there was it was the only thing that wasn’t a general category. ๐Ÿ™‚ Okay, that was a pot shot – I was picking on you there my friend.

    Yes, I did find it interesting that you listed mostly EMC products in your article. Not just because I work for NetApp but because what I see from the NetApp side (things that are generally available today), there are alot of technologies that will be game changers for 2010.

    What I should have said at the outset of my blog was “Kudos to Curtis for writing this prediction article for 2010 – he nailed the top end categories, Deduplication, Cloud, Virtualization, Tape Backup and CDP (near-CDP).”

    Those clearly will be key areas in which 2010 will find impact, so maybe my title was a bit much.

    Oh and you don’t need to apologize – I asked the question and I appreciate your honest answer.

    Chapa

  • As to the “what’s to stop” question… The answer’s nothing. But I don’t find a lot of customers with a lot of disk just sitting around waiting for backups, so…

    As to the “companies pushing it,” maybe it was just bad sentence structure on my part. The first reason that I said target dedupe would increase was that it’s right for people. Then what I meant to say was “and then it will also grow because of some of the companies that are pushing it and their reasons for and methods of pushing it.”

    The VMware piece of the source dedupe section was only meant as a “oh and BTW.” It didn’t occur to me to then morph that section yet again to say “put your VMware on NetApp.” Sorry.

    As to cloud backup. I don’t know how many times I have to say it. I don’t care what BT/Rackspace/Joyent have as their infrastructure. You don’t have a cloud backup service. And no, I’m not just talking home stuff here. I’m talking about petabytes of commercial data under protection with commercial cloud backup apps. Even if any of them were using your storage, you still don’t have a cloud backup app.

    As to mentioning a lot of EMC products, what can I say? You know there’s no love lost there with me and them, but they kicked butt in 2009. (They specifically kicked YOUR butt when they stole DD away from you.) That’s not my fault.

    When I see a mass migration to things NetApp (like I did to Mozy, VMware, Data Domain, and hopefully soon to the vStorage API) I’ll be glad to talk about it.