Analysis: Behind Oracle's Acquisition of Virtual Iron

Oracle's move shakes up the virtualization status quo

You have to give Oracle Corp. credit. When seeing something that strikes its fancy, it snaps it up, such as soon-to-be-former Sun Microsystems Inc., for example.

Oracle struck again recently, picking up virtualization specialist Virtual Iron Inc. The Virtual Iron buy raised more than a few eyebrows, however, coming as it did less than three weeks after Oracle first vaulted into the hardware and virtualization fray with the acquisition of Sun.

What's so intriguing about the Virtual Iron deal? There's a feeling, after all, that it gives Oracle more virtualization technology than it can realistically use. On top of Virtual Iron's virtualization assets, and in addition to its own virtualization portfolio (centered on Oracle VM), Oracle will inherit both Sun's standalone Xen-based hypervisor (xVM) and the Solaris operating system's more sophisticated support for Logical Domains. In addition, over the last 24 months Sun did much to articulate a creditable virtualization management and provisioning strategy (xVM Ops Center), as well as a thin-client story (Sun Virtual Desktop Infrastructure, or VDI, software).

What does Virtual Iron bring to the table? It's a fair question, industry watchers concede. "That Oracle is looking to beef up its in-house virtualization assets is not especially surprising. What is less expected is that Oracle would make this acquisition on the heels of its purchase of Sun -- which has considerable in-house virtualization assets of its own," writes Gordon Haff, a principal IT advisor with consultancy Illuminata, in a posting on his blog.

Oracle, for its part, contends that Virtual Iron will help it shore up its dynamic data center management capabilities. Haff doesn't contest that claim. "Oracle says that Virtual Iron adds dynamic resource management and automation including capacity management, power management, and the ability to integrate with other software through an open, comprehensive, and scriptable API," he says. Moreover, Virtual Iron's hypervisor, like the hypervisor that Oracle inherited from Sun -- and like Oracle's own VM hypervisor -- is based on the open source Xen project, so Oracle obviously didn't acquire the company for its unique or best-of-breed hypervisor technology.

By the same token, that makes Virtual Iron much more than just a hypervisor: it fields a full-fledged virtualization stack. Notwithstanding the manageability and automation tools that Virtual Iron brings to the table, and putting aside, for the moment, the existence of three distinct Xen implementations, cases of redundancy will surely exist.

For any other vendor, the question might be "How much?", but redundancy is something from which acquisition-hungry Oracle has never shied. Its history is full of examples where it has purchased technology that it already possessed in-house, sometimes in a highly developed state. Consider its acquisition of the former Sunopsis nearly three years ago; that vendor gave Oracle best-of-breed data integration (DI) technology, something that Oracle -- which had recently released a beefier version of its Oracle Warehouse Builder DI tool -- ostensibly already had. (To be fair, Oracle articulated a credible use case for Sunopsis, citing the applicability of its specialty extraction, loading, and transform, or ELT, technology for its ongoing Fusion middleware efforts.)

There have also been cases in which Oracle determined to introduce overlap where none had previously existed. In 1995, for example, Oracle acquired the venerable Express online analytical processing (OLAP) engine from Information Resources. Five years later, in 2000, Oracle announced plans to build a new OLAP capability into its flagship Oracle 9i database. The ugly truth, of course, is that -- thanks to development delays -- Oracle continued to sell the Express technologies through 2005, according to OLAP market guru Nigel Pendse, a principal with the Business Application Research Center.

Now, industry watchers such as Haff have more questions than answers.

"To be sure, both Sun and Virtual Iron's virtualization portfolios are based on the open source Xen project, so they're at least potentially complementary. However, these capabilities would seen to overlap Sun's xVM Ops Center to at least a certain degree," he points out, adding that Oracle also didn't talk about Virtual Iron's putative role in its channel ecosystem.

In any event, Haff speculates, Sun and Virtual Iron targeted two very different market segments. "Sun, like Oracle itself, offers products that have very much an enterprise flavor. Virtual Iron, by contrast, has in recent years primarily focused on the mid-market -- smaller companies that didn't necessarily have the best fit with the sophistication (and complications) of product from the likes of VMware," he indicates. "There seems at least the potential here for Oracle to expand its reach down market -- perhaps in conjunction with parts of Sun's open source stack such as MySQL."

A Gambit to Seize the Xen Initiative?

Gartner Inc. had a considerably more pessimistic take on the acquisition.

"Oracle could take six months to a year to integrate a fragmented assortment of management tools consisting of technologies from the ongoing Sun acquisition [e.g., OpsCenter and xVM], Oracle's Enterprise Management [e.g., Oracle VM], and [Virtual Iron's] Live-branded products," write analysts Philip Dawson and George Weiss in a Gartner research blast.

"For example," the duo continues, "the [Virtual Iron] Live products are certified to the Red Hat, Novell and Microsoft operating systems, but not to OpenSolaris." Another rub, Dawson and Weiss point out, is that Oracle VM supports both Windows and Red Hat -- but not SUSE Linux -- guests. Oracle positions VM as the lynchpin of its application, middleware, and DBMS virtualization strategy. To that end, it has certified a laundry list of its products (e.g., the Oracle 10g database; Oracle Real Application Clusters, or RAC; the Oracle E-Business Suite; Oracle PeopleSoft Enterprise 8.4.x and 9.0; Oracle Siebel CRM 8.0; and Oracle Hyperion 9.3.1) to run in conjunction with Oracle VM.

For this reason, the acquisition drastically shakes up the Xen status quo, Dawson and Weiss argue, basically putting the Xen ball in Oracle's court. This could create two distinct Xen blocs: an Oracle/Sun/Virtual Iron axis on one side and another bloc aligned to the aegis of Xen creator Citrix Systems Inc.

"This acquisition … will consolidate the Xen community around Oracle/Sun/[Virtual Iron], Citrix and Novell. Management and tools for Xen will evolve from a highly fragmented and competing community to two consolidated camps based around Oracle and Citrix ," write Dawson and Weiss, citing the likelihood of management agreements with Novell and Microsoft.

"However, today these make up less than 20 percent of a market dominated by VMware on x86. We expect to continue to see the Xen hypervisor embedded with the x86 hardware. Intel will continue to benefit from the Xen consolidation, further threatening VMware's dominance, which once seemed insurmountable."

comments powered by Disqus