It's war: Six groups of hopeful startup fleas are attacking the six incumbent storage elephants. They hope to dislodge them and grow into self-sustaining or acquired businesses at the big guys’ expense.
The six incumbents are Dell, EMC, HDS, HP, IBM and NetApp. They face attacks from storage startups to their NAS filer products, to their iSCSI arrays, to their Fibre Channel arrays, to their object storage products, to generally everything they do from those who want to replace their arrays with cloud storage.
The first storage startups’ attack comes from filer accelerators like Alacritech, Avere and CacheIQ. They stick a multitiered accelerator box, often using solid-state storage, at the server or array end of a NAS link and by caching filer data offer dramatically faster data access for reads and writes to older filers. Don't bother to upgrade to that expensive new filer, they say; instead, turbo-charge it with our scalable boxes, which can logically group multiple filers into a single resource.
GridIron does a similar thing to SAN arrays, but its approach – that of accelerating SANs -- is different from the other attackers of block access arrays, which instead aim to replace SAN arrays.
Meanwhile, the iSCSI block access and unified file/block arrays offered by the six incumbents are being attacked by a host of flash-based iSCSI storage startups’ products; NexGen Storage, Nimble Storage, SolidFire, Violin Memory, XtremIO and others. Coraid is attacking here too with its ATA-over-Ethernet arrays.
The incumbent SAN heartland of Fibre Channel SAN arrays is being attacked head-on by all-flash storage startup Pure Storage. Various other vendors allow Fibre Channel access to their flash arrays too, such as Texas Memory Systems, but may lack the software functionality needed to dislodge the incumbents.
Another attack on the incumbents with flash boxes comes from Astute and Tintri with flash arrays built specifically for virtualised server data storage.
On the object storage front, storage startups such as Amplidata, Caringo, Cleversafe and Scality are circling the incumbents in this still-nascent storage protocol field.
Meanwhile, cloud gateway startups and players like Nasuni, Ctera and Riverbed are looking at just about everything and saying, “We can cache locally and use enterprise-class cloud storage for everything but the hottest data. Get rid of your in-house storage mess, and do it better for less with the cloud, using Amazon, Joyent, Nirvanix or some other flavour-of-the-month cloud storage service.”
These hungry startup nippers want to grow fast. They say their technologies, designed from the get-go for today’s storage requirements, are much better than the decades-old incumbent vendor storage offerings. And, often being flash-based, they outperform the last crop of startups bought by the incumbents, meaning EqualLogic, LeftHand Networks, 3PAR, Compellent, etc.
It's a hugely hopeful world out there in storage startup land. They love their shiny new technology and are full of evangelising and revolutionary marketing zeal. The incumbents mimic their technology; have you seen HP's all-flash P4900 (LeftHand) array, for example? This is all to the good. It is technology evolution in action -- digital Darwinism. Without the prod from these hungry startups, the incumbents would get rich, lazy and fat. But instead they have to develop their own me-too technology or buy a startup to gain it.
If they do it well, the storage startups are repelled or absorbed rapidly. If they do it badly or under-estimate the threats, the startups can progress and we see the likes of, for example, a Data Domain emerge. It was so successful because backup to disk just keeps on getting bigger and data deduplication helps with that problem better than anything else. The result? NetApp and EMC both tried to buy Data Domain, with EMC eventually succeeding.
The same story lies behind the Diligent purchase by IBM, and Ocarina by Dell. This pattern of startup success and incumbent acquisitive response lies behind the iSCSI storage startup acquisitions by first Dell and then HP, and purchases of Compellent, 3PAR and Isilon.
These were the last decade's startups. Now we have a flock of new ones knocking on the incumbents’ customer doors, saying, "Let me in. We have much better products."
Whether they do or not will only be decided by the market. The startups’ products may solve problems you will pay money for. Or they may solve problems that only loom large in their engineering teams’ eyes but mean nothing to you or any but a handful of other users.
Storage tech startups are a huge gamble. If they pay off, you get better technology and become more efficient. It’s terrific that this happens, because without it we would be stuck, with no data deduplication, no iSCSI arrays, no thin provisioning, data tiering or flash arrays.
The wars against the incumbents are a welcome feature of our industry, and long may they continue.
Chris Mellor is storage editor at The Register.
This was first published in April 2012