Is storage driven by disruptive technology innovations? Largely, no. With the exception of solid-state drives (SSDs) and disk as a replacement for tape, storage is driven mostly by sustaining innovations.
A series of classic innovations in storage technology can be seen to start with Auspex Systems' invention of network-attached storage (NAS) in the late 1980s. NetApp then created the NAS appliance, which sold to smaller shops and contrasted sharply with Auspex's large and costly systems. NetApp developed its product's capability to the point where it was too late for the slow-reacting incumbent to respond. The result? Auspex Systems is now long gone. NetApp's filer appliance innovation certainly disrupted Auspex Systems, but it didn't disrupt storage-area network (SAN) storage sales because customers still needed block access to data and SAN storage. In the wider context of the storage market, what NetApp did was considered a nondisruptive move because there was little or no cannibalisation of SAN sales.
A similar point can be made about content-addressed storage (CAS), which has essentially been pioneered by EMC's Centera. The product has built out a new market segment for reference information storage alongside existing SAN and NAS storage, but hasn't disrupted anything else in the larger scheme of things. Sure, newer players like Caringo are trying to disrupt Centera, but the CAS market is stable. The same goes for Fibre Channel (FC) SANs. They haven't been disrupted by iSCSI because block storage access over Fibre Channel takes place in large SAN fabrics, and iSCSI is perceived to be best suited to smaller SAN applications for smaller customers.
iSCSI could do more and could replace some Fibre Channel SANs, but so far the technology has built out a new segment of SAN storage that FC, because of its cost, can't be applied to.
Dell happily sells iSCSI SAN storage alongside block storage products -- EqualLogic alongside Clariion -- and Hewlett-Packard sells its LeftHand iSCSI products along with Fibre Channel EVA arrays.
So where is disruptive storage innovation happening?
Solid-state drive technology is beginning to replace Fibre Channel drive technology, which seems to be a clear example of storage disruptive innovation. SSD technology has evolved from consumer electronics applications to the point where you can stick a controller in front of lots of NAND chips and make a cost-effective replacement for expensive, short-stroked Fibre Channel drives. STEC has done this and Intel is following suit. The leading drive suppliers, Seagate and Western Digital, have their own SSD initiatives, but neither one has made a real presence in the Fibre Channel hard drive replacement market.
PCIe-connected solid-state drives, like Fusion-io's ioDrive, could add another tier of storage to servers and may erode Fibre Channel drive sales. Solid-state drives represent a pretty clear-cut case of disruptive innovation and hard disk drive suppliers are finding that their Fibre Channel products could be replaced by SSDs. The 1.8-inch hard drive market could also be disrupted into non-existence by SSDs, but this will probably take several years to play out.
Another ongoing disruptive storage innovation concerns disk backup and tape. With hard disk drive arrays gaining SATA bulk storage capacity and offering backup to disk, virtual tape library (VTL) functionality and data deduplication, we've arrived at a tipping point where cost and data access advantages are starting to displace tape storage.
We can be pretty confident that SSDs are displacing FC hard drives and are poised to displace Netbook hard drives if SSD technology gets its act (and pricing) together. We can be more confident about tape being disrupted. Are there any other disruptive forces facing us in storage?
Is cloud storage the next culprit?
There's an arguable case that cloud storage -- in the form of public clouds -- could be disruptive. If customers migrate to accessing their data in public clouds then they won't need to buy iSCSI or FC SANs, NAS filers or CAS reference stores, LTO tape libraries, Brocade and Cisco SAN fabrics, or clustered filers.
I don't think private cloud storage offers this threat, as it arguably wraps cloud storage access around evolved versions of existing storage products.
Disruptive innovation is a judgment of hindsight. We'll have to wait and see if cloud storage is disruptive or, in private cloud form, just another in a long line of sustaining storage innovations.
Chris Mellor is storage editor with The Register