Original URL: http://www.theregister.co.uk/2009/04/21/oracle_sun_storage/

Oracle stares into Sun for storage future

Just what is Oracle's business now?

By Chris Mellor

Posted in Storage, 21st April 2009 10:03 GMT

Comment Oracle becoming a storage supplier through buying Sun - who would have thought it and what does it mean for Sun's storage products?

There's long been a contradiction in Sun's storage array strategy. How can it push the open storage idea and simultaneously sell storage arrays with what amounts to proprietary controller hardware and software? If Sun carried through on its open storage vision, then in the long term, the arrays it OEM's from Hitachi Data Systems and others would be toast.

With Oracle buying Sun the Ellison-run enterprise's stated storage idea is to focus on open storage and systems. Oracle has long had a very successful strategy of commoditising everything in the hardware-to-operating system stack underneath its own software products, so that it could charge high margins for its software, but also put forward a low-cost total Oracle DBMS and app system, compared to other DBMS vendors with proprietary and/or higher-priced hardware. Think Big Blue and HP.

Now Oracle has its hands on a whole swathe of storage HW and SW, much of which is in a mess due to the contradiction between Sun's give-it-away-free open source SW strategy and its proprietary HW running proprietary software.

General server, storage and software supplier or not?

So far, Oracle has sold things that integrate and support its database and application SW crown jewels. With Sun it is now in the general server, storage and software businesses. Does it want to be? Will it continue to be? We don't know. Apart from general intentions of investing in Sun and growing its revenues, all we know is what the announcement singles out, such as a focus on Open Storage.

Let's assume Oracle is not in the general server/storage/SW supply business. Then we might assume that, as a working principle, Oracle will look at the Sun storage collection and focus on anything that can further its commoditise-everything-below-the-Oracle-SW-layer strategy. If other Sun products are in secure niche markets, don't cannibalise existing Oracle products, and represent an upsell opportunity for Oracle, then they should survive. Everything else, unless it can be adopted for use as a competition stopper that doesn't harm Oracle's own product revenues, will be toast.

So ... Sun's Open Storage 7000 product with a freely available Solaris + DTrace + ZFS + lots more SW stack should survive, and be embraced avidly by Oracle. The run of the mill Sun storage arrays, from the OEM'd HDS USP-V downwards, face strategic oblivion.

The mainframe StorageTek tape libraries represent a relatively secure mainframe niche with recurring revenues, a good customer base, and upsell opportunities. Expect this to continue. Ditto the mid-term future of StorageTek's proprietary tape format, but its longer term future depends on the combined Oracle/Sun organisation's view of the trend to more and more disk-based protection.

The hybrid storage/server products, like the X4500 Thumper, are being used by Sun to sell against Oracle data warehousing and business intelligence products. They could be bundled with Oracle's own products and survive, but the partner support for currently competing software, such as GreenPlumb, could be scaled down.

The Sun virtual tape library strategy may well continue, with a combined mainframe and open server VTL product being developed, as Sun is trying to do at the moment.

The MySQL database represents Oracle DMBS competition, but if Oracle could position it as a low-end stopper supporting its own database and not cannibalising its sales, then it could survive. Equally it might just be left to languish.

The Sun cloud storage effort, such as it is, may be developed as part of an over-arching Oracle cloud storage strategy.

Oracle, with Sun storage, is now a competitor to EMC. That might cause fireworks, especially if Oracle heads cloud-wards with vigour.

If Oracle will be a general supplier

If Oracle decides it is in the general server HW and storage HW supply business, as well as the soup to nuts, disk to app, integrated HW/SW business, then its job of optimising the two businesses, database plus app SW on the one hand and Sun's product tool box on the other, will be much, much more difficult. If it decides that this is indeed where it's at, then on the storage side it needs to look at replication, thin provisioning, working with VMware and Hyper-V, deduplication and all the other storage technologies in deployment and build out.

Actually, Oracle is already in the general storage supply business, but at arm's length, via Larry Ellison's funding of Pillar Data. There is now the tantalising prospect of that company being bought closer to Oracle's own storage hardware business.

Maybe Pillar CEO Mike Workman could even be asked to run the whole Sun storage show. That would be a shock for John Fowler, the current boss of Sun's systems business. On the other hand, Mike W. would get on well with Jon Benson, the long-lived StorageTek executive who reports to Fowler and runs Sun's StorageTek (Tape and non-hybrid server/storage drive arrays) business.

Potential toast rack

If Oracle confines itself to the integrated disk to Oracle app business, then whole swathes of Sun's product range are toast. However, whichever strategy Oracle decides upon, general server/storage supplier or integrated Oracle stack supplier, the logical outcome of the open storage strategy is that the proprietary arrays are history.

Much more detail will come out in the following weeks, as Oracle understands in more depth what it is buying and what it can do with it. It's going to be very interesting to see what it discards from its newly-acquired line-up, what it keeps, and how it plays its hand.

One aspect of that hand concerns NetApp. Oracle uses a lot of NetApp kit. The Sun NetApp lawsuit about ZFS IP infringement could be settled out of court with goodwill on both sides. It's a distraction, especially so considering that Oracle is facing a massive integration exercise in the shape of Sun.

Oracle has not had to integrate a HW vendor before. SW integrations like PeopleSoft are huge, but digesting a near-tier one Systems HW+SW vendor like Sun is up on another level altogether. It's not inconceivable that Oracle could choke on it.

The other storage server/storage system suppliers - Dell, HP and IBM - will be watching very closely, as will the storage suppliers with current Sun supply contracts. Everything is up for grabs.

Who will win and who will lose? El Reg is going to follow this with interest. ®