Feeds

Will HP 3PAR high-end storage arrays?

Crossroads for ex-EMC man Dave Donatelli

  • alert
  • submit to reddit

Secure remote control for conventional and virtual desktops

Comment What is the future for HP's high-end XP storage arrays? Which way will incoming HP server-networking-storage boss David Donatelli jump?

Donatelli left EMC to run a combined server, storage, and networking operation at HP. The company is busy creating an integrated IT stack consisting of servers, storage, networking, and system software. He is legally prevented from being involved with storage until 2010, but his influence in the server and networking sphere appears to being felt with the 3Com purchase, which bulks up HP's ProCurve networking business nicely.

The company has been acting as if owning core technology is its preferred way of operating. Instead of partnering with LeftHand Networks for isCSI SAN (storage area network) storage, it bought it. It is deeply involved with core tape technology in the DAT and LTO space. It has bought a storage management company, AppIQ, and two storage file companies, Ibrix and PolyServe, which makes its reliance on Hitachi Data Systems for its flagship storage offering, the XP array, surprising.

HDS has OEM relationships with both HP and Sun for its USP-V array, a very powerful array with a controller of great expandability and the ability to virtualise third-party storage arrays connected to it. There are just four high-end enterprise storage array suppliers, if we exclude NetApp from the top tier which many observers do.

There is EMC with its recently revised Symmetrix V-Max, IBM with its DS8000, HDS with its USP-V, and 3PAR with its InServe. There is also Pillar Data with its Axiom arrays, but these are generally thought to be not up there in the enterprise array stratosphere, occupying rungs on the storage ladder below that.

Were David Donatelli and his team at HP to cast their eyes over the XP product and decide that it was not suited to their high-end storage array requirements going forward to an era of prolific server and data centre virtualisation and cloud computing, then what are their alternatives?

First off, they could take whatever HDS has in its development pipeline. They might not like the idea of taking a box that Oracle/Sun will likely take, especially with Oracle preferring Sun now for storage hardware in its database machines, but business logic may put that feeling aside.

Secondly, HP could look to partner elsewhere, which means EMC, IBM, or 3PAR. The first two are extremely unlikely as HP competes strongly against both. The latter is a theoretical possibility, were 3PAR willing.

Thirdly HP could look to build its own high-end array technology. Fourthly, it could look to buy it in.

Building a high-end storage array from scratch is a massive effort. Revising an in-house array as EMC has done in a major way and as IBM is doing steadily with its DS8000 is expensive enough but starting from ground zero means major bucks, gazillions of them, and several years effort.

HP shows a preference for buying in technology it can use. Were it to decide to buy in high-end array technology there are only, I reckon, two candidates: 3PAR and Pillar. Of the two 3PAR has the better fit as Axioms compete more with EVA arrays than XP ones.

What we think we know about USP-V futures comes from talking to its CTO, Hu Yoshida. He says the future for HDS is a combination of scale-up within the array and scale-out, meaning adding more arrays. We're probably looking at the adoption of Intel's Xeon 5500 processors, with maybe 256 or more controllers, interconnected with a high-speed switched fabric, and supporting multiple tiers of storage ranging from SSD through 15K rpm drives to 7,200rpm SATA bulk storage disks. The new array will probably feature very good integration with server hypervisors, thin provisioning, automated data movement between tiers, multi-tenancy, and different qualities of storage service for tenants and applications using it.

The future for 3PAR seems to involve federating its clustered InServe arrays, with marketing head Craig Nunes saying: "It's a strategic concept for us...and a number of components are in active development today... We're going to be making an upcoming announcement about this." Clustering gives 3PAR scale-up and federation would give it scale-out.

It looks to us in El Reg towers that HP's high-end array future is either continued partnership with HDS and the coming new USP-V or the taking of its high-end storage destiny into its own hands with a purchase of 3PAR. ®

Secure remote control for conventional and virtual desktops

More from The Register

next story
Ellison: Sparc M7 is Oracle's most important silicon EVER
'Acceleration engines' key to performance, security, Larry says
Oracle SHELLSHOCKER - data titan lists unpatchables
Database kingpin lists 32 products that can't be patched (yet) as GNU fixes second vuln
Ello? ello? ello?: Facebook challenger in DDoS KNOCKOUT
Gets back up again after half an hour though
Hey, what's a STORAGE company doing working on Internet-of-Cars?
Boo - it's not a terabyte car, it's just predictive maintenance and that
Troll hunter Rackspace turns Rotatable's bizarro patent to stone
News of the Weird: Screen-rotating technology declared unpatentable
prev story

Whitepapers

A strategic approach to identity relationship management
ForgeRock commissioned Forrester to evaluate companies’ IAM practices and requirements when it comes to customer-facing scenarios versus employee-facing ones.
Storage capacity and performance optimization at Mizuno USA
Mizuno USA turn to Tegile storage technology to solve both their SAN and backup issues.
High Performance for All
While HPC is not new, it has traditionally been seen as a specialist area – is it now geared up to meet more mainstream requirements?
Beginner's guide to SSL certificates
De-mystify the technology involved and give you the information you need to make the best decision when considering your online security options.
Security for virtualized datacentres
Legacy security solutions are inefficient due to the architectural differences between physical and virtual environments.