Which storage technologies and vendors will fly in 2011?
Someone's gonna crash and burn
Questions At this turning point of the year, as Dell wraps up Compellent and InSite One, and Atrato slides under the waves, it's timely to look ahead at the storage world in 2011 and ask que sera?
Data growth and storage demand is seemingly unstoppable, and new storage tech has advanced quite a bit in 2010. Things are looking good for the industry in general.
It may be that multi-level cell (MLC) flash will be cheap enough, reliable enough and last long enough to kill off the 3.5-inch Fibre Channel hard disk drive (HDD) business in tandem with the rise of fast 2.5-inch HDDs. That same flash could bound out from its tablet computing base and establish defensible beach heads in the notebook market and, even, the desktop PC market. After all, what other way can Windows bloatware-bound notebooks and PCs fly than by feeding the Windows software flab to multi-core processors as fast as flash makes possible?
We might see 2.5-inch disk drives take over from 3.5-inch disk drives everywhere except for bulk data storage in arrays and in desktops and external drive. For sure we'll see 4TB 3.5-inch drives next year and 2TB 2.5-inch ones will be hinted at, won't they?
Here at El Reg we are waiting to see if HDD suppliers other than Seagate develop hybrid SSD+HDD products.
A big, big question is whether converged IT stacks will prove popular and start lifting the fortunes of system (server + network + storage) vendors at the expense of stand-alone storage vendors. We see four main issues here.
Will Michael Capellas' VCE deliver on the promise its execs and company bankrollers are hinting at? We are also looking at the prospect of IBM delivering an integrated platform combining its blade servers, Storwize V7000, and Blade Network Technologies' networking along with orchestration software and perhaps development tools? There is the possibility that Dell will buy into networking technology, the presumed main missing link in its converged IT stack ambitions. Lastly, Cisco could add storage in some way to its UCS and Nexus products to produce an all-Cisco converged IT stack.
Will cloud storage move into the mass market and take cloud storage gateways with it, brightening the prospects for start-ups such as Aspera, BridgeSTOR, Cirtas, Nasuni and StorSimple? We might see EMC's Atmos surge into prominence as cloud storage providers buy the box in droves. The cloud could also propel object storage into prominence, boosting the fortunes of Caringo and Dell with its DX6000 box.
We are also looking at EMC's Centera expectantly and asking whether Atmos will take over its object role as far as the cloud goes, and whether a specific Centera hardware platform will survive separately from the CLARiiON and Celerra arrays.
Will NetApp just keep on growing market share or will EMC, Dell, HP and IBM manage together to rein it in?
That depends upon HP managing to revive its mid-range fortunes with a refreshed EVA line and boosted down-market 3PAR box. It will require EMC to deliver on its mid-range unification theme by doing something in the Celerra and CLARiiON space, something combinatorial as we might say, and IBM putting more of its mid-range wood behind one arrow, looking like the Storwize V7000 and not the XIV array. That will possibly become a filler-in of IBM's mid-range storage catalogue, along with the DS5000.
All of the possibilities pronounced in the article don't last long.
Sure, magtape can do ten years at a stretch in optimal storage conditions, maybe longer if you have Das Ueber Data Vault, but what about longer storage?
What's the go in that case? How to circumvent data rotation, or at least extend rotation cycles and reduce data rot? Magtape?
Non-optical media? May as well rub it with adipocere at the start of the storage cycle, for all its recoverable worth for at least that may prevent oxidation.
What the fix, O Mighty Reg Pundits and Punters?
Netapp continue to grow, despite the issues developing when you stress the array with too much random I/O when the capacity reaches above approx 70%. Barry Burke to getting increasingly angry at this.
Brocade to be bought by HP or maybe an outside shot at EMC - buying back some of the IP they sold when getting rid of McData. Storage, networking, virtualization, middle-ware.
HDS to continue selling, with no fuss, and with terrible software.
EMC to release ControlCentre without host agents, and using SMC for all configuration changes. Still Java, still too bloated.
I won't make an attempt at XIV. Does it sell, is it selling, is it going to be axed, does it do what it says it does, would IBM sell a dud?
Re IBM XIV - StorageBuddhist
With respect, XIV is hardly the peer of VSP and V-Max, despite IBM wanting it to be so.
It doesn't come close on RAS features, software functionality, drive type/flexibility.