On Monday Oracle announced that it plans to purchase Sun for $7.4 billion. Assuming no unforeseen glitches crop up, the deal should close sometime this summer.

The acquisition potentially makes Oracle a much more formidable adversary to IBM, Microsoft and its smaller competitors. With Java, Solaris, and Sun hardware in tow, Oracle will control a set of technologies that should blend nicely with the company's software and services business. Not one to shy away from hyperbole, Oracle CEO Larry Ellison declared: "Oracle will be the only company that can engineer an integrated system -- applications to disk -- where all the pieces fit and work together so customers do not have to do it themselves."

I imagine IBM would take exception to that particular claim. In fact, if you listen to Monday's investor webcast on Oracle's Web site, you might get the impression that Oracle and Sun are about to invent system integration. Since most enterprise computing systems are built from commodity parts and open source software, the advantage of owning specific technologies is not as great as it used to be. If it was, Sun wouldn't be in the position it's in now.

The Oracle-Sun merger is getting mixed reviews from analysts and journalists. As Timothy Prickett Morgan writes at The Reg: "Sun and IBM mixed like oil and water. Oracle and Sun, the two original Silicon Valley IT startups (and no, Hewlett-Packard, oscilloscopes don't count as IT) and the two darlings of the dot-com boom, will mix something more like vinegar and oil -- you can whip it up into a colloid, at least."

I think that's a fairly accurate metaphor. On paper at least, Sun and Oracle are a decent fit. There is barely any product overlap (even Oracle's flagship Database 11G offering and Sun's MySQL may find a way to live peaceably with one another). Synergies between Sun's hardware, Solaris OS, and Java, along with Oracle's database software, are there to be had if Ellison can find a way to wrap all the pieces together. In general, the acquisition has better chemistry than the IBM-Sun deal.

The acquisition certainly has some critics. Computerworld's Steven J. Vaughan-Nichols doesn't think much good is going to come from this relationship, and starts this commentary as follows: "I'd thought about Oracle buying Sun. But, then I thought, 'Larry Ellison isn't that dumb.' Well, I was wrong. Ellison is that dumb. Oracle is buying Sun in what may be the most moronic technology acquisition of the 21st century."

Aw c'mon, this one isn't even close to some of the more ridiculous acquisitions we've seen over the last several years (think eBay and Skype). A lot of Oracle database software is run on Solaris and the company's Fusion middleware is based on Java. There's every reason to think that Oracle will be able to exploit Sun software and hardware to build higher value database appliances.

High performance computing may be a different story. Since Oracle hasn't divulged plans about how it will integrate the two companies, it's too soon to tell what will become of products like the Constellation blades, the Sun Grid Engine, Lustre storage, etc., but the signs are not particularly encouraging. From what has been publicly stated, the focus is clearly going to be on enterprise database systems. As of yet, no mention has been made of Sun's high-end computing aspirations.

In the latest TOP500 list, Sun claimed only seven systems, but, rightly or wrongly, the company saw HPC as a way to drive innovation in the rest of the company. And Sun's whole Redshift vision was about garnering computing at enormous scale and moving traditional applications into the cloud. As John West astutely pointed out at insideHPC, Ellison was trashing cloud computing just last year. I doubt if he's had an epiphany since then.

Once the deal closes, Sun's workforce and internal infrastructure are bound to get downsized. If Oracle's priorities are such that Sun's institutional knowledge in high-end technical computing are not retained, that will eventually mean there will be one less vendor in HPC.