Throughout every sort of economic turmoil in recent memory – the deep recession in the early 80s, business downturns in the early and mid-1990s, and even the dot-com crash in 2000, there’s been one constant that you could take to…

Most observers believed that Oracle’s main goal in buying Sun was to secure Java, MySQL, and other Sun software assets. However, Larry Ellison’s recent pronouncement that Oracle is “in the hardware business to stay” has rocked the industry. Ellison clearly indicated that Oracle will be looking to optimize Sun hardware (including SPARC, CMT, and x64 platforms) to better run Oracle products. The implications of this are profound. Oracle would be competing head-to-head with their system vendor partners Dell, HP, and IBM.

At this point, all we know is what Ellison said. We don’t know if they truly intend to keep all of Sun’s products or how long it will take Oracle and Sun engineers to come up with a new integrated server offering. Right now, I’m pretty sure that not even Oracle knows the answers to those questions. However, they will certainly be coming out with something, whether it’s a roadmap, a promise for new hardware, or actual new boxes. My industry contacts are almost universal in their belief that Oracle doesn’t really want anything to do with Sun hardware, and that the bad economy has thwarted their intentions to sell off Sun’s gear and associated businesses. There is certainly some support for this position, given that Oracle and HP were rumoured to be looking to tag-team Sun, with Oracle taking some SW and HP grabbing the gear. That deal, assumedly, fell apart and won’t be resurrected anytime soon.

Industry folks, particularly other server vendors, typically say that they don’t see Oracle sticking with Sun’s hardware; or, if they do, they see Oracle being successful enough to warrant the investment. I’m not so sure. I think there is a case to be made for what Oracle is talking about with integrated HW/SW bundles – as long as the integration provides some compelling value to customers.

In a big surprise to most industry watchers, Oracle has swooped in and made what looks to be a successful bid to purchase Sun Microsystems. In the press release, Oracle is offering a pretty big number ($7.4 billion), about $500 million more than what IBM was prepared to spend, and considerably more than Sun’s market cap just a few months ago. The premium over IBM’s bid startles me a bit; I would think that given IBM’s backing out of the deal (along with Sun’s current market cap), that Oracle could get a much better price – maybe as much as $1 billion better. Details on the story here.

Taking Larry Ellison’s comments at face value, it looks like Oracle is planning on holding onto most, if not all, of Sun’s stuff. We can assume that he definitely wants the software assets – Java, MySQL – but does Oracle really want to get into the hardware wars? Do they want to invest billions in supporting, developing, and pushing boxes? Most of the people I’m talking to right now believe that Oracle is going to keep some parts of Sun, but part out most of the rest of it to others. The stuff getting the heave ho, according to industry insiders, is most likely the proprietary HW assets – like SPARC64 and UltraSPARC-related stuff. They are also expecting Oracle to sell off the StorageTek stuff along with the rest of Sun storage. This might garner some quick bucks to defray the cost of the deal.

But I’m not so sure.

The latest from The Wall Street Journal today is that the IBM/Sun deal may be coming apart. IBM made their formal offer on Saturday and was rejected because the price was too low – $9.40 or below, according to WSJ…

For those of you ready for a fresh take on IT topics and trends, we can point you to the latest from our own Dan Olds, who will be hosting webcasts on The Register. Some of these will be talks…

Breaking news from The Wall Street Journal points to progress in the IBM/Sun deal: the share price has dropped a dollar or two to a range of $9 – $10 per share. In addition, our pal Timothy Prickett Morgan at The Register is reporting on rumors (just rumors, as he correctly emphasizes) that the deal may be announced as soon as tomorrow.

In any negotiation, you try not to give up something without getting something else. The guys driving the deal on the Sun side (Southeastern Asset Management and KKR) gave up purchase price in order to get a commitment from IBM that they’ll pursue the deal even in the face of intense regulatory scrutiny… or in the face of regulatory requirements that might cause IBM to disgorge part of the Sun purchase. As stated in the WSJ article, Sun’s negotiators want to limit IBM’s ability to exit a potential deal.

And there well may be some considerable scrutiny focused on this deal. It all depends on how the regulators slice and dice the various markets in which Sun and IBM compete. The most obvious problem area is in tape storage, where a combined IBM/Sun (via StorageTek) would control nearly half of all tape storage revenue and the standards for high-end tape drives.

In the server markets, it’s a bit cloudier; just how cloudy depends on just how you define a “computer”. On the Unix system front, the combined companies would earn 65% of all Unix market revenue, and assumedly at least that much in terms of units. For the server market overall, IBM/Sun would account for 42% of all revenue. But, again, what’s a business computer? Do mainframes really compete straight up with Unix and x86 servers? Mainframes are still a significant portion of IBM server revenue, but they aren’t exactly substitutes for Unix and x86 systems. The vast majority of customers don’t look at a server purchase and say, “Hey, should I buy a Dell 2-socket Intel box for our new web server application, or perhaps go with a mainframe?”

The purchase of SGI, although not yet a certainty, is a very good move for Rackable Systems. It gives them a great entrée into the HPC market and instant credibility with customers, plus access to those customers. And the price…

A combined IBM-Sun could present a whole new competitive challenge to HP and Dell. With the purchase, IBM would get a much larger footprint in data centers worldwide, into which they can sell the entire IBM portfolio of products. This is a major reason behind HP’s purchase of EDS last year. The company would also get the lion’s share of the Unix system footprint and, assumedly, an inside track on replacing all of these systems over time. IBM would also get an interesting x86 gambit with ownership of IBM’s Solaris x64 operating system and virtualization products. This may be persuasive to ISVs and convince them to port to the O/S in greater numbers. A blending of AIX and Solaris that would run on RISC and x86 processors could be a very compelling combination for customers who are looking to simplify their environments.

In the near term, I believe that IBM would continue support for all of Sun’s current server platforms. The key to getting the most out of an acquisition like this is retaining customer good will and preserving Sun’s revenue stream.

On the Unix side, I believe that the Niagara-based systems stand the best chance of being developed further in a combined IBM/Sun company. Those systems are unique and have a value proposition that isn’t duplicated by existing IBM systems. I see Sun’s standard SPARC-based platforms gradually being subsumed into IBM’s POWER-based gear. To do this, and to preserve Sun’s customer base, we’re talking about a blending of Solaris and AIX into a common code base. This would take quite a bit of work, but it’s certainly possible.

We’re going out on a limb and saying, “Probably”; we believe there’s a better than 50% chance that some sort of deal will be cut between IBM and Sun. According to published reports, HP, Dell, and Fujitsu have all turned…