Oracle Places Its Storage Strategy Bets

Exadata Database Machine: This is a database solution that focuses on online transaction processing, data warehousing and database consolidationworkloads. One might ask why this is included in the storage portfolio. The answer is that this product incorporates integrated storage as Oracle hasre-engineered databases and storage to improve database I/O; this is, in effect, a not-so-secret storage weapon. If Oracle can get a customer to buy an OracleExadata Database Machine, then a significant amount of storage goes along as a bundled ride.

ZFS Storage Appliance: This is an enterprise network-attached storage (NAS) entry that has been engineered for Oracle environments. There are direct connections between it and the company’s database solutions, which helps the company pursue what might be called an "everything goes better with Oracle"-type of strategy.

StorageTek Tape: Oracle has not abandoned tape for data protection and archiving. In fact, StorageTek (acquired by Sun in 2005) offers the market’sbiggest tape library solution. Now, tape libraries are a horizontal market sale rather than a very targeted sale, such as in a database environment, butthere is no need for Oracle to have everything tightly coupled with its traditional products. However, as much as vendors might like the notion of selling largeenterprises everything they need, few of those customers fulfill all their IT needs with one vendor.

Pillar Data Systems’ Axiom: This latest acquisition gets Oracle back into the block storage area network (SAN) I/O storage business. Note that Pillar was majority-owned by Oracle CEO Larry Ellison, who has been an investor in Pillar since its birth about 10 years ago. That close relationship means that Pillar is likely to be accepted neatly into Oracle’s overall storage portfolio. Although Axiom is a horizontal rather than a vertical market product, Oracle may still be able to find some ways to exploit its architecture within its traditional environments.

Frankly, and even the company is likely to agree privately, the Oracle storage portfolio has nowhere the breadth and depth of any of its major competitors' (systems from EMC, HP and IBM), and any best-of-breed product claims could be successfully contested by those three competitors, as well as by Dell, NetApp and others. However, dismissing Oracle’s storage portfolio out of hand would be a mistake.

To understand why, we need to understand how the company is playing its cards. Even though we have to be careful not to extend the analogy too far, consider that in some games, players can get one or more cards (sometimes for purchase) to improve their chances. In this parlance, Pillar Data Systems qualifies as another Oracle card.

The question is, how strong a card is it? For whatever reason, Pillar has not had the success of some of its competitors, including Dell (with Compellent), EMC, HP (with 3PAR) and IBM. However, cut Oracle a little slack, as it just acquired the company. Oracle may have some ideas about how the Axiom architecture can play well with its traditional products, but it needs time to explore, develop, and announce what those would be.In games such as bridge and poker, where a number of hands are played over time, no one has an absolute single strong hand (in terms of high cards); the overall winner is the player who has a relatively high number of "good" hands (that is, a better hand than the other players).

Oracle’s strength as a vendor lies in customers with a strong commitment to the company’s core products (most of which are not storage). But the company’s argument, and it is a strong card, is that applications, data and storage, which have traditionally been separated, now require tighter alignment for optimal performance.

Leveraging expertise across the elements of the hardware and software components of the information infrastructure can improve both performance and efficiency. Oracle feels that it has superior application and data knowledge that it can integrate with storage, and it claims that conventional storage system vendors do not have the expertise necessary to tie everything together efficiently.

Although Oracle recognizes that its claims do not apply in all business or IT cases, it makes the bold assertions that applications can run up to 10 times faster with only a tenth of the amount of storage used in conventional systems--and, oh, by the way, that queries may run up to 70 times faster. This strategy is intriguing, but it is a lot like a "savings-of-up-to-75% sale," where the cost of at least one item is reduced by three quarters while most other product prices are reduced by only a little.

The question that customers have to ask and answer clearly is what these performance and capacity benefits mean in their particular environments--a lot or a little. The company is pushing the concept that its software runs better with Oracle storage, such as guaranteed storage performance on demand. So any CIO in a heavily Oracle-based IT shop has to at least give the company the opportunity to present its case, such as using an Exadata Database Machine incorporating integrated Oracle storage.

Oracle is essentially making a back-to-the-future trend. In the past, servers and storage were thought to need tight coupling in a direct-attached storage (DAS) mode with a single vendor’s hardware. That myth was dispelled when innovative storage-focused vendors like EMC started displacing server vendors’ storage even before shared storage in the form of storage area networks (SANs) arrived.Now the argument has returned, but with a twist. Application and software vendors like Oracle typically have a deeper knowledge of their product than storage vendors and so can design them to use storage better--and they do not have to share that knowledge with competitors.

Oracle knows that its storage strengths are not as a general-purpose storage supplier, although it may enjoy some success in the general market with its tape libraries. Therefore, the company has wisely decided to play its hole card, where it has a strong home court advantage in its installed base. Oracle knows where those customers live and what they need and is likely to be able to at least gain an audience.

If Oracle can demonstrate product integration that results in meaningful benefits, including both value-added performance (faster queries may drive added revenue)and economic efficiency (as represented by TCO or ROI calculations), then it has the chance to develop a growing revenue stream that incorporates a significant amount of storage.

But integration with its own products is not the only story that Oracle will need to tell about storage. It also needs to have a strong story on how it will integrate with the other aspects of the IT transformation process, such as virtualization and cloud computing.

The bottom line to remember is that Oracle is a fierce competitor that never quits or rests until it has won every possible chip. And that will continue to make the ITinformation infrastructure wars--including the storage war--very interesting to watch.

At the time of publication, Oracle is not a client of David Hill and the Mesabi Group.