In the storage industry, 2015 was both the best of times and the worst of times! This year saw some serious storage M&A activity. Dell offered to buy EMC. Western Digital finally owns HGST and now SanDisk too. IBM bought object storage supplier Cleversafe.
Underlying these events is an inescapable fact: Storage, like the rest of the IT industry, is changing rapidly. For example, a senior exec for Hewlett-Packard Enterprise in EMEA boasted that Chinese “white box” servers account for little of the overall server market, but he missed the point. The low revenue reflects the murderously low prices Quanta and others charge. In fact, this represents the future for HPE; competing for hardware sales with ODMs will be very tough.
Where servers go, so goes storage. We can expect boxes from SuperMicro, Quanta and Lenovo, for example, to make serious inroads in the storage space. While storage has been protected by software value-add, software-defined storage and the likes of Ceph are knocking down the walls and the real cost of hardware will translate to much lower prices per terabyte over the next few years.
This accounts for the Dell-EMC deal. EMC appeared to have seen the future, with software-focused products like ViPR aiming to embrace commodity hardware under the EMC software umbrella. Even that may not have been enough of a future, especially given the notorious reluctance of hardware-centric behemoths to become software-driven. The alternative is to become a bigger company, hence the deal with Dell.
HP’s split has more to do with the malaise in the PC segment. Desktop sales are in free fall. Tablets are being refreshed much more slowly, with the result that the whole segment is shrinking. But this also reflects a major change in storage. The volume of “consumer” drives is falling fast and no gyrations about capacities of 5 TB or 10 TB, or solid-state drive hybrids, can fix that problem. WD and Seagate are seeing their cash cow dying in front of them!
This brings us to that WD-SanDisk deal. WD has a strategy that includes SSDs. HGST, which is now an actual WD subsidiary, has some pretty good high-end drives. Adding SanDisk is a great move for WD, since it brings a broader range of products to its portfolio. From embedded drives to all-flash arrays, SanDisk covered a broad swath and having WD selling its products should boost sales a lot.
Seagate shows some symptoms of being left behind at the gate. Focusing on a soon-to-be highly competitive box business with likely a shrinking market with OEM customers, Seagate bought OEM box-maker Xyratex a couple of years back. It should have bought SanDisk instead to get a foot in the door of the SSD game, but now it’s too late!
Let’s look at why the timing on SSDs is so critical. SSD prices continue to drop. It’s much cheaper to buy an SSD than a so-called enterprise disk drive, and the latter business is beginning to dry up. 3D NAND technology could drive prices down to bulk storage drive levels in a year or two, whereupon only the most cautious CIO would stay with hard drives.
Yet flash technology itself has been threatened this year. Intel/Micron are talking up a new memory technology that is much faster than flash and also denser. We are a year or two away from real impact with 3D X Point, but life will get very interesting when it’s available. In the face of all of this, having a larger technology and market base in flash is a good thing, while access to flash replacement technology is even better. This is the real value of the SanDisk deal to WD.
Meanwhile, IBM extended its software portfolio by buying Cleversafe. This was an excellent move: IBM got a patent portfolio and some solid object storage technology, all of which can easily be moved into a software-defined storage environment. IBM seems to be positioning itself to have a strong software story while getting away from money-losing hardware. It remains to be seen if IBM is agile enough to win in this game, but it's making smart moves.
The storage industry also saw some shakeout in the all-flash array space. Pure Storage is doing well, but Violin is struggling. Overall, the segment is doing well, reflecting the low-cost boost to the SAN it delivers.
The success of AFAs, coupled with compression/deduplication and thin provisioning have caused the traditional disk array market to slump. All of these reduce capacity demand, with the AFA providing a vehicle to buffer data prior to size reduction. The impact in the enterprise is huge, with capacity needs reduced by as much as 5x for hard drives.
When this reduction is coupled with drives that have grown from 3 TB to 10 TB over the last year or so, the need for new hard drive array boxes has generally evaporated. Sales of so-called hybrid arrays, with a few token SSDs thrown in, have saved some of the market, but the AFA does a better job and typically now comes with compression/dedupe built in.
The SAN that the AFA plugs into is under siege. As predicted, Ethernet is starting its long march to kill off Fibre Channel, and the segment is weakening. FC will probably enjoy cash-cow status for a few more years, but any real growth is history. Emulex, one of the two FC Host Bus Adapter vendors left, sold itself to Avago this year.
Finally, the march of cloud storage and open source software such as Ceph and OpenStack Swift continues unabated. These technologies reflect the future of storage much more than EMC arrays.Jim O'Reilly was Vice President of Engineering at Germane Systems, where he created ruggedized servers and storage for the US submarine fleet. He has also held senior management positions at SGI/Rackable and Verari; was CEO at startups Scalant and CDS; headed operations at PC ... View Full Bio