John L over at InsideHPC writes about an article that appears in the print version of the New York Times. In this article, John notes that IBM is scouring Sun Microsystems contracts … well, their lawyers are doing due diligence, in large part, to help figure out if there is something of value there.
John opines that there are many things of great value. As I see it, the market has indicated otherwise, by valuing Sun Micro stock where it is now.
Basically, the lawyers will have a more detailed understanding of the financial aspects, issues, liabilities and assets going forward.
Whether or not IBM makes a real offer for them remains to be seen. This would be as a function of the fit and synergy with the IBM product/offerings, the likely future value of the properties, and how much cost savings and additional revenue could IBM realize by such an acquisition.
Sun has quite a lot of baggage. They are not really an HPC company, HPC is rounding error to their overall business. They are the maker of an OS on the rapid decline, and several versions of hardware. The commodity version competes with everyone else (remember that not a good idea trying to out-Dell Dell comment I keep making? ), and the non-commodity/proprietary version isn’t a volume platform (so ISVs/IHVs are either dropping it or not porting to it unless Sun pays them to stay on it).
Sun has some customers that seem to have the cold dead fingers mentality … literally, they won’t consider alternatives, even if the alternatives have been proven (again and again) to be better, cheaper, faster than their preferred hardware. Such is life, we run into lots of this in the government space, and fewer these days in the financial services space, though I can say that times are changing.
You can’t paint a box [purple|some other color] and charge 3x the price for it. Nor can you offer boxes which underperform the competition, while asking a premium over the competition for these boxes.
Which goes to what Sun has to offer IBM.
John points to HPC stuff … GridEngine (the old Codine code from way back) which Sun acquired then open sourced … and competes with IBM load leveler, the Lustre file system which Sun acquired when they bought the assets of Cluster File Systems (and seem to be quite intent on turning into a ZFS based cluster file system, which means it won’t run on Linux) … which competes with IBM GPFS, MySQL which Sun acquired for way too much money, Sun’s support for Rocks which competes with Xcat2 from IBM …
In fact, looking over the list John provided, I am hard pressed to see the synergy. IBM developed lots of its own HPC IP in house. Sun acquired most of its HPC IP. Sun’s proprietary hardware is of dubious value … the market has spoken on this. Sun’s OS is of declining value, again, the market has emphatically spoken on this matter. In fact, looking over Sun’s business, the market (before the IBM announcement pushed things up for them) valued Sun at little more than double its cash minus its debt.
A different strategy would be IBM acquiring Sun, and looking at the support/maintenance contracts, specifically to take out a competitor. That is, buy them out, own their IP and support/service. Cease all development on competing hardware/software. Completely GPL Solaris. This gives IBM access to the aforementioned “cold-dead-fingers” people, as customers, and a chance to convert them to IBM hardware.
That could be valuable.
Then again, IBM could just want the cash. Tack the debt onto its existing balance sheet, grab the cash, shut down the rest, assume the service/support. All stock transaction, IBM gets $2.6B in cash, in exchange for stock. If done right, they take on the debt as convertable to equity later on.
The whole transaction costs them effectively nothing. And the get a competitor out of the way, ownership of some properties (of dubious value), and access to Sun’s existing customer base.
Did I mention the “cold-dead-fingers” folks? Yes? Oh.
Basically, if this goes through, then Sun’s days are likely numbered. The execs won’t spin it that way of course, scaring customers is not a good thing to do.
But I would expect huge swaths of Sun hardware to be EOLed. I’d imagine we’d see Sparc put out to pasture. OpenSolaris would continue as the natural evolution of the EOLed Solaris. And that would likely be GPLed.
Do remember, IBM is not an HPC company, they are a business machines and services company. The services part is profitable for them, and ways to grow it would be good for them. HPC is rounding error to their bottom line. As it is for Sun. IBM is a strong player in HPC due to some very cool internal IP they use, and their ability to leverage external IP without so much of an ego issue. Sun hasn’t quite mastered that last portion.
My bet is that IBM is looking at them from a services/market expansion/competitor take down scenario. Because that is where IMO, the value of Sun is … not in its hardware (insert the “its a bad idea to try to out-Dell Dell” comment here).
But I could be (badly) wrong on IBMs intentions.
I hope, someday, when we sell the day job, that it doesn’t get torn down by the buyers. This all depends upon our ability to prove that the market and revenue is larger by investing in the business than the value the buyer would get in tearing it down. Its a cost-benefit strategic analysis, and that is what IBM is doing vis-a-vie Sun. Which strategy will be most positively impactful, and minimize adding new liabilities?
[update] I should probably point out that, during my 8 month stint at IBM’s T.J.Watson research center in Yorktown, I saw a huge number of very cool tools, which never saw the light of (commercial) day. The e3 editor, which I used 20 years ago, is still, IMO, one of the best programmer editors out there. Many many other technologies are kept within IBM, and never reach outside.
What I am saying is that IBM has a huge wellspring of internal use only IP they can tap to bring products out. Adding externally generated IP to this stockpile may or may not be a good thing, but IBM certainly has strategic IP depth, and probably doesn’t really need Sun IP here …