Everything I know about NetApp tells me they have solid products with good and sometimes great functionality. But if NAS is becoming a commodity that only buys them time, not immunity.

It is like the mini-computer companies in the ’80s. As PC performance kept rising, the weak ones with the least product differentiation went under first. DEC, the largest and strongest, was the last one standing and - if they had adopted to network and processor commoditization - could still be alive and kicking today, just as HP is.

So I’m not bashing NetApp. I like them as a company. But they have been seduced by enterprise IT and they need a climb-down plan NOW, not in 5 years.

Their large and loyal customer base gives them maneuvering room. I hope they make the most of it.

Robin

I have been a NetApp, EMC, HDS and HP customer in large environments (0.5-1PB disk). As best I can tell, NetApp pulls it all together as well if not better than the competition. They cetainly provide a better unified solution than anyone else. Not perfect, nothing is, just better.

NetApp IP-SAN iSCSI support, manageability, integration, experience and scalability is second to none, period. iSCSI is THE disruptive technology to FC-SAN. Just a matter of time as iSCSI sails on up on the perpetual ethernet wind. Netapp will continue to lead here for a long time.

NetApp NAS is second to none in major enterprise. There are some very interesting new players, but they have a long way to go and certainly do not have the technical core competence that Robin refers to. The other majors have have been playing catchup for years but have never really matched the NetApp special sauce. To suggest that commodity NAS will eat away at the NetApp core value or business viability suggests a slightly shallow understanding of the market size, growth potential and forces at play. Until NAS is offered with high performance out of the “cloud”, there is no real fear. Up till then, all these commodity NAS players actually end up driving more customers towards the big players. Then there is VmWare, Oracle, SAS etc over NFS…….take a look. Then checkout direct NFS (Oracle) and parrallel NFS.

FC - NetApp is growing in this segment faster than anyone else as far as I can tell. Ultra high-end is still something yet to be achieved, but again, it is only a matter of time. Key thing to remember, most revenue is not made in the high-end, only the most margin. Netapp is gaining quickest where it matters, in the middle. The low-end is ripest for iSCSI.

The scalability issue. Yes this can be an issue at the ultra high-end where storage is not planned and executed with a considered strategy. Note that this is most environments, due primarily to management/political interference rather than lack of technical aptitude. I have seen it with every vendor I have touched. Ask someone about legacy FC environments with a micture of EMC hardware running and you will understand a true scalability and manageability nightmare.
How is Netapp helping? The spinnaker/ONTAP GX integration is a core component from what I can tell. We have all heard it before, just treat the storage as one big, albeit intelligent, pool. (I think ZFS is built almost entirely on this principle.) The manageability or “intelligence” is where it really matters, the hardware and OS are cool, but nothing if they don’t address the core customer problems. HDS tried it one way and it is kind of cool, but doesn’t really solve enough of the manageability problems to set the world on fire. HP are HDS in that space. IBM SVC does some very cool stuff, but has major scalability issues in practice. EMC are still working on InVista. Not entirely their fault, getting agreement in the FC space is like hearding cats. But by the time it is ready it will be addressing a storage market segment in decline. Revenue to be had for sure, but not really a clear future.

VmWare/Xen etc can present something interesting and new in the storage space, but they have their hands so full in other areas of rapid development that I think pervading a crowded storage space is far from a market reality. Enter some group of geeks with a smart virtual storage appliance. They need to be very clever mind you given the almost complete lack of standards in the server virtualization space.

Software coupled to an expensive solution? Err how is this different from the entire enterprise storage marketplace? A huge portion of the storage dollar is to be made in the software space. If you aren’t spending it with Netapp, you are spending it with someone else because you still need a solution to your business problem.

Now, a good discussion on how storage in the cloud will influence the fortunes of various storage players, that is something I am keen to rant about.

–END RANT—

FC, iSCSI, NAS
Application integration with VMWARE, Oracle, MS Exchange..etc…

It is great to have one box that does it all, if you are a small shop with 10,20,30TB. If you are a large enterprise customer with hundreds of TB, this model does not work so well. Netapp systems do not scale well, and using their software couples you an expensive solution.

I mean, if someone wants to have a NAS to service NFS for example, a SnapAppliance or Iomega NAS will indeed work. But bear in mind that the performance will be really slow. Depending on the business needs, and if this is fine for the sysadmin then so be it.

Also, sure someone may build a NAS using Linux or FreeBSD box with all of his/her custom configurations/optimizations. Been there, done that. But a forethought sysadmin should realize the time and efforts needed for this, and the fact that s/he won’t be working on the same place/same position forever. This means, if s/he ever leave the replacement could very well rip the whole solution and replace it with something that is supported.