Why an IBM purchase of Sun would make sense

This was originally posted at ZDNet's Between the Lines. Also, see update at bottom.

IBM is reportedly in talks to buy Sun Microsystems for $6.5 billion and the deal is long overdue. The companies mesh on the open-source software front, Sun is struggling, and IBM can consolidate some server market share.

First, the headlines. The Wall Street Journal is reporting that IBM could acquire Sun as early as this week. IBM would pay all cash for Sun. The Journal also reported that Sun has approached a number of large companies about an acquisition; a move that throws cold water on CEO Jonathan Schwartz's everything-is-fine video.

Behind the scenes here, we've frequently had chats about how IBM would take out Sun. The only real debate was whether Big Blue would acquire Sun in parts or as one sum. The working assumption was that Sun would be broken up and sold in parts, but the IBM deal also works nicely. Here's why the deal makes some sense:

IBM can acquire server and storage share. Sun still has a lot of hardware on the market in key verticals such as finance and telecommunications. The problem is that Sun is reliant on U.S. sales and that's not a fun place to be right now. With Cisco Systems entering the server market, the profit margins could be squeezed--especially if the server essentially becomes a storage and networking box too. By acquiring Sun, IBM gets more scale so it can endure the margin squeeze. That same argument holds for storage hardware too.

There are issues to be worked out on the hardware side though. IBM has refrained from commodity servers and Sun plays in that space. Meanwhile, servers run on different chips--Sun has Sparc and IBM has the Power architecture.

The time is now. IBM is a software and services company, but it needs hardware, which would be roughly a third of revenue with Sun, to sell its other offerings. Hardware is often the entry point for IBM's software and services. With a stronger hardware business it can fend off HP in the marketplace.

Sun is a powerhouse in Unix, which is still a key platform, but isn't exactly gaining in the market these days. IBM could acquire Sun and establish two key beachheads: Linux and Unix. The former will ultimately take over for the latter in the data center. IBM can play both and sell you the services to migrate while it's at it. Bonus for Big Blue: Sun would enable it to pressure HP's Unix-based businesses too.

One problem: IBM sells AIX Unix servers. Sun sells Solaris (hat tip to a reader pointing that out).

Sun has to do something. Sun is a company that has to transition a legacy hardware business to one modeled more on open-source software and services. That's a wrenching change that may not work out. An exit ramp makes sense right now. Besides, Sun was reportedly turned down by Hewlett-Packard and Dell was mum.

This is about HP--NOT Cisco. The initial headlines will paint IBM's move as a reaction to Cisco's entry into the data center. The reality is HP is the target. HP acquired EDS to directly take on IBM. IBM is returning the favor by squeezing HP on hardware.

Open-source software. Sun's future is offering an open-source stack of software led by MySQL at its core. IBM is all about open source. The two together make a lot of sense and IBM could pull MySQL, Lotus, OpenSolaris, and other parts together in a nice stack. In addition, Sun's open-source software needs distribution--it recently did a deal with HP. IBM has distribution galore.

One problem: There's a lot of software overlap here. In databases, IBM has DB2 and Sun has MySQL.

Java. Java is arguably Sun's best asset, but the platform was never monetized. Perhaps IBM will have better luck.

Cloud computing. Sun has some interesting ideas on cloud computing and its plans could work. Sun has also made targeted cloud-computing acquisitions. Meanwhile, Sun has a cloud-computing press conference Wednesday morning, a shindig that will be dominated by IBM talk now. A few folks seemed to buy that Sun-as-cloud-player marketing, but it really looks like a relabeling to me. IBM could absorb Sun's plans in its big cloud services offering.

It's about consolidating the data center. Data center managers will want fewer throats to choke and Cisco isn't going to make life easier. Sun can allow IBM to consolidate a data center rival and bring things back to equilibrium now that Cisco has entered the market.

What could go wrong? A few items. For starters, there are regulatory concerns. Who knows whether the Obama administration would approve this deal. IBM would have roughly 42 percent of the server market with Sun, according to IDC. HP, however, would be second with 29.5 percent. With Dell a strong No. 3 with 11.6 percent share the IBM-Sun deal should pass scrutiny. But it is a wild card.

Here's a look at the server market share standings:

And then there's culture. If you want to know the big cultural difference, look no farther than Schwartz's ponytail vs. IBM's typical look. IBM is all business and Sun is sort of business with a lot of Silicon Valley shtick. That said Sun's workers may find themselves relieved by the IBM deal. Sun's upcoming struggles are fairly obvious and IBM looks like a great option for those who choose to look ahead.

Update 7:04 a.m. PDT: Pondering the valuation: As of its most recent quarter, Sun had $2.6 billion in cash, equivalents, and short-term marketable debt securities. So the rest of the business is worth $4 billion roughly. One working theory is that IBM would potentially buy Sun and sell the hardware business to Fujitsu, a close partner of Sun. IBM could license Solaris, since it would own the code, to Fujitsu. Assuming IBM does sell some hardware assets to Fujitsu for $1 billion or so, the Sun acquisition doesn't look so large.

Meanwhile, analysts are mixed on the deal. UBS analyst Maynard Um says in a research note:

Our Sun thesis has been that there is a potential restructuring story given relative inefficiencies. We do think IBM would be in a position to take out cost more quickly, however, we expect Sun revs to continue to be pressured in FY09 & FY10 given its reliance on high end servers and limited ability to monetize its software. And while IBM is certainly financially capable of purchasing Sun, we would note that a $6.5bn acquisition would be large even by IBM's standards, and would also be a deviation from IBM's current strategy of tuck-in acquisitions.

Meanwhile, Sun shares had a predictable response to the news, soaring approximately 72 percent in premarket trading.