It’s interesting to consider which storage startups have been successful in the last 6 to 7 years. The most successful companies I’m seeing are the ones that actually sell disk storage in an integrated system across standard interfaces.
Companies that attempts to integrate functionality directly into the network like Storspeed (NFS/CIFS caching), the NAS switching startups I’m aware of (e.g., Acopia Networks), and the network file encryption players like Decru really had to play by the incumbent’s rules. For example, Netapp and EMC were threatened by the full architectures that Acopia had, and feared getting commoditized in the backend. I think Decru was allowed to play because it had such a narrow niche focus and it addressed a real customer concern (performance while encrypting) but wasn’t too threatening. But if you look over the last 3-4 years, the acquisitions/IPOs/growth have predominantly been in the storage hardware space by players that actually sold the storage in stand-alone boxes: LeftHand, EqualLogic, 3par, Compellent, Data Direct, Isilon, Data Domain, BlueArc (very solid growth and likely to get bought soon), Spinnaker, and so forth.
Ocarina Networks is the only recent storage software startup that got bought for more than $100 million, and I think that was a fluke. They were not getting much traction in the field, and Dell for some reason overpaid. They were not getting enough traction to keep their VC’s in the game, and then the Dell Bluebird offer happened. Good for them, but you can’t just depend on bluebirds.
So why is it that the stand-alone storage hardware players are winning? I think there
are a couple of reasons:
(1) To deliver adequate “performance” for the customer, you have to control the disk storage hardware and software to provide a complete solution that addresses the
customer’s primary “performance” metric, e.g.,
- Compellent — performance metric is simplicity of deployment for SMB’s of small- to medium-sized arrays
- 3par — autonomic self-management for large arrays
- Data Direct — streaming large file performance and very large capacity and high density (HPC/media focus)
- Equallogic — not sure, but I think it was their strong iSCSI focus and support
- Isilon — scalable NAS storage for large capacity customers accessing mostly large files
- Data Domain — deduped storage for backup
(2) By specializing along a particular metric of performance, you can create a “jobs-to-get-done” purpose-oriented brand that helps customers and partners identify you as the go-to company to get that job done (e.g., Isilon is the go-to company for large-file scale-out NAS, BlueArc is the very high performance NAS player, Compellent is the SMB-friendly storage array that provides sophisticated features like thin-provisioning); and by doing both hardware and software, you can fully control that performance metric in a way you can’t if you just do software.
(3) Related to controlling hardware and software, you need to productize along industry standard interfaces (iSCSI, FC, NFS/Ethernet, CIFS/Ethernet) and not try to get overly funking in what you try to do across those interfaces — I think this is what hurt Acopia. I loved Acopia’s technology but they had to get outside where the standards were in a pretty aggressive way. Sometimes that’s a good approach, but it didn’t play out well for them.
(4) Partly it’s just a numbers game: VC’s like the larger revenue you can garner at lower volumes versus a low volume software provider; the latter has to have a high ASP to get decent revenue in the beginning, but high ASP’s generally make it harder to sell higher volume, it’s kind of a vicious cycle. So I think that VCs in the early stages have a top-line growth focus, and they are willing to put up with a small (or even negative for a while) bottom line while the volume grows. Perhaps that’s why so many of these storage hardware/software startups survived while a lot of software startups died.