Sun Microsystems: Too good to fail

Regardless of whether or not you use Sun hardware or software, you need Sun Microsystems. Maybe Cisco needs it, too...

Well, it all fell apart this weekend. IBM pulled its offer to buy Sun off the table, and the future of the deal is uncertain at best. My first reaction to the IBM/Sun deal was that it was a 50/50 deal. There was a 50 percent chance that the acquisition of Sun would be a good idea and 50 percent chance that it would do significant harm to both companies. Frankly, IT cannot afford to take that chance. IBM and Sun are both too important to be weakened by a bad deal. Let me explain.

I have no real love for Solaris. I've used it extensively, consider it to be miles ahead of other operating systems in some respects, and miles behind in others. OpenSolaris is a Very Good Thing that's changing my opinions considerably, but overall, I wouldn't choose Solaris over, say, FreeBSD or Linux in most cases. I'm also not a terribly big fan of Java, and I regard it as a necessary evil in some cases, horrifically overused and abused in others. That said, Sun has much, much more to offer IT in general than its two flagship products.

[ Special report: IBM in talks to buy Sun. ]

First off, it's currently selling some extremely well-designed and well-built x86 servers. Just about across the board, Sun Fire servers are simply doing the job elegantly and at a low cost. I was so impressed with the Sun Blade 6000 chassis I reviewed over a year ago that I spec'd them for two completely different purposes: a VMware ESX farm and a compile farm. I'm happy to say that they've been completely problem-free, with the minor exception of a single failed hard drive in the 18 months since they went into service. The Sun hardware I've had in my lab has been polished, sleek, and extremely capable. Mind you, this wasn't always the case, and Sun certainly produced some flops in the early part of the century, but it has definitely righted that particular ship.

Sun's storage products are also quite impressive, with its new Unified Storage Servers offering huge amounts of extremely fast storage based on low-cost, low-RPM SATA drives and the miracle that is ZFS (Zettabye File System, another Sun development, of course). And obviously, there's the large installed base of SPARC systems, including the impressive SPARC T2 and the myriad other products available from Sun. They have an unmistakeable presence in the IT world.

This was what concerned me with the IBM deal. From a hardware perspective, IBM competes with Sun in several markets, and there's no telling which hardware design would win that particular battle. To me, IBM's hardware is simply not as well designed and produced as Sun's offerings, and were IBM to buy Sun and kill off the x86 hardware line, a great disservice would have been done to the IT world in general. I recall feeling exactly this way when the HP/Compaq deal was announced -- Compaq server hardware was significantly better than HP's products. Luckily, HP knew that and killed off its own abysmal NetServer line in favor of Compaq's ProLiant server line. There's no guarantee that IBM would do the same. There are similar concerns over storage, too.

So that brings me to a postulation. I previously pooh-poohed Cisco's Unified Computing System as a premature technology. I, as well as many others, also questioned how Cisco could hope to compete in the server market without any prior experience. Maybe it can buy that experience. Maybe it approached Sun with a buyout offer as well, prompting the dissolution of the talks with IBM.

Now that might make some sense. Sun could bring several decades of server experience and stellar server designs into Cisco, along with all the other goodies like OpenSolaris, MySQL, Java, ZFS, and so on. It would make Cisco a simply massive player all of a sudden, with almost no overlap between the companies and products in almost every possible category. There would be no hard choices about which server line got the axe along with the thousands of employees that comprise the production, sales, marketing and support structure for that business line. There would have to be some restructuring, sure, but nowhere near the pain that would be caused by trying to glue Sun to the side of IBM. Then again, it might be nearly impossible to manage. Anything's possible.

I honestly don't know if Cisco has the dough or inclination to buy Sun, but I do know that IT wouldn't be nearly as far down the proverbial road without Sun, and losing the company completely would be a significant step backwards. If Sun can't make it by itself, then maybe a Cisco acquisition would be the next best thing.


Copyright © 2009 IDG Communications, Inc.