There was other news this week, to be sure. Just nothing that anyone wanted to talk about (with the possible exception of Sun’s cloud). No, all the talk this week was of the rumored acquisition of Sun by IBM: I’ll give Reuters the link because of their very, um, visual headline, “Web powerhouse Sun Micro poised at brink of IBM’s maw.”
Anyway, to own the truth, your guess is as good as mine regarding the accuracy of the articles in question. While I’m acquainted with the people that could presumably tell me, I also know better than to ask them to comment on material discussions. Even should the articles turn out to have the story right, I would not care to predict the outcome: sometimes these things happen, and sometimes they do not. Without better data, I would not care to speculate on the probability of either potential outcome. Though I will observe that this would be – by far – the largest acquisition IBM had undertaken in a while, and it’s frankly out of character for a firm that prefers the tactical to the strategic when it comes to acquisition.
What I do feel compelled to comment on, however, not least because people keep asking, is what a potential combination of the firms might look like. True, I would normally skip the exercise on the possibility that the effort would be wasted given the uncertainty involved, but in this case I think there are a few points slipping between the cracks that might merit a bit of attention, if only because they deserve it absent the current news.
As always, we’ll turn to the Q&A.
Q: Before we begin, any disclosures to make?
A: Yes indeed. Both of the parties at issue here, IBM and Sun, are RedMonk customers. As are several competitors of both, including Dell, HP, Microsoft and Red Hat. Others mentioned, such as Amazon and Google, are not. That should about cover it.
Q: First question: did the news surprise you?
A: Well, it was the day after St. Patrick’s day, so yes, I was a bit startled when I got onto the network Wednesday morning and my IM client lit up like a Christmas tree. Even in retrospect, however, the news was a bit surprising. True, rumors about Sun have circled for years involving various players – IBM included – but culturally they’re not a terrific fit, and there is substantial overlap in the product portfolio. So yes, I was at least mildly surprised.
All of that said, when Sun was briefly trading at or below its cash on hand, I briefly considered borrowing a few billion to buy it. Well, not really, because no one would loan me billions of dollars, but with value of the assets it has long seemed like a logical target.
Props to Peter Yared, by the way, for predicting such an acquisition back in 2004.
Q: Most of the reactions to date have been skeptical. Consider Stephen Shankland’s initial take:
IBM buying Sun no slam dunk. Java, multithreading expertise fits, but otherwise difficult. IBM needs fewer, not more, OS and server lines.
Do you agree with that assessment?
A: I’m certainly in agreement that it’s no slam dunk; with any deal this complex, there are always a lot of moving parts, and there’s also the aforementioned redundancy. But I respectfully disagree with the assessments, such as analyst colleague Dana Gardner‘s, that the deal makes “no sense.”
Q: Ok, let’s get to that in a minute. How about the aforementioned redundancy: what product lines overlap at the two systems vendors?
A: It would almost be faster to list those that don’t, like Sparc servers. But let’s see: Java middleware, identity, x86 systems, storage, operating system (albeit indirectly), database, office productivity, development tools, high end servers and so on and so on. Which sounds like the prescription for disastrous merger, true, but that ignores the fact that many of the pieces would either be complementary to current IBM product lines or potentially useful entrypoints. This focus on redundancy also tends to obscure a few of the truly useful assets that IBM would control via a Sun acquisition.
Q: So you think that a merger potentially makes sense?
A: Let’s say instead that I think Sun would have a lot to offer a potential acquirer, IBM or otherwise.
Q: Such as?
A: Far too much to easily list. Consider the following assets for which Sun owns – in its entirety, in most cases – the intellectual property for (and can, therefore, exclusively license or relicense), which are listed in no particular order and is not intended to be complete:
- One of the most important languages and platforms of the past few decades, in Java
- The most popular relational database on the planet, in MySQL
- A promising database for the cloud, in Drizzle (IP ownership here is not exclusive)
- The most popular office alternative, in OpenOffice.org
- A System V Unix license
- One of the most advanced filesystems currently available, in ZFS
- The best observability and tracing software currently available, in DTrace
- A revitalized open source development tool with good dynamic language capabilities, in NetBeans
- Virtualization software, both server (xVM) and desktop (VirtualBox)
- A growth Java middleware line, in Glassfish
- A popular and well regarded identity business
- Some of the more sophisticated storage systems currently available, in the Fishworks line
- Growth x86 server and storage (Thumper) lines
All without getting into the people resources, such as the chip engineers that are consistently cranking out more and more cores per chip: not a bad skill to have in a future that is likely to look highly concurrent, or the community involvements and relationships. I wonder, as an aside, whether or not the terms of Sun’s settlement with Microsoft, which include protections for the StarOffice product line, would pass to IBM.
Certainly the patent portfolio would, where in spite of not professing any affection for the process, Sun’s played the same game everyone else to protect themselves.
Now some of you may observe that IBM has made substantial money off of many of the above – most notably Java – without spending a dime on a Sun purchase. But there is, as far as I’m concerned, a lot of IP that has potential value here, which was why Dana’s comment, “Buying Sun IBM gains little other than some intellectual property and mySQL,” seemed to me unnecessarily dismissive.
Q: Can you provide an example of where the IP might benefit IBM?
A: Sure. Take something as mundane as the filesystem, ZFS. While it’s not perfect – I regularly field complaints about its file creation, for example – it is nonetheless the design target for the next generation Linux filesystems, which aim to meet or exceed both its performance and its functional capabilities (think snapshotting, pools, etc). The primary problem is that even the bridge filesystem, ext4, is not yet production ready, and the ultimate response, Btrfs, is probably not going to be ready for testing for a year, meaning it’s years, pluarl, from seeing production workloads. Owning the copyright to ZFS, however, would allow IBM – if it so chose – to relicense the asset under the GPL so as to be compatible with Linux, thus permitting a port.
Or consider the case of Drizzle, the cloud oriented fork of the MySQL project. While it’s true that in forgoing the usual copyright assignment typical to Sun led open source projects, IBM would not be purchasing exclusive rights to the codebase, it would nonetheless be acquiring significant IP, not to mention expertise, in building out a database for a new and potentially massive market.
The fact is that while traditional relational datastores can be bent to purpose in the cloud, they were not designed for the environment any more than Microsoft Exchange was designed for multi-tenant architectures. Drizzle, on the other hand, is being built with an entirely different set of assumptions: N>1, 64 bit, no business logic in the database, etc. That IP is worth something, even to a firm like IBM which has sprawling database assets of its own, including investments in people like Damien Katz and their projects like CouchDB.
What would it be worth, do you think, to have the inside track in developing the MySQL of cloud persistence workloads?
Q: So, assuming the valuation less cash on hand is in the neighborhood of $4 billion, are you arguing the IBM should pay $4 billion for a filesystem or a non-production cloud database?
A: Hardly. The point is, rather, that there are a great many such assets that Sun’s excellent engineers have built over the years that represent value. Moreover, the ownership of many of these assets – unlike the open source projects under the Apache, Eclipse or Linux banners that IBM typically particpates in – is exclusive. Which would give IBM a variety of options in terms of downstream usage, from relicensing to taking some of the assets proprietary.
That plus the opportunity to more heavily leverage some inefficiently and undersold Sun assets could justify the current valuations, depending on the model.
Q: But examples like ZFS aside, is it really plausible that the major product lines could be perceived as complementary? Is MySQL complementary, for example, to DB2?
A: Purely complementary? Perhaps not. But this certainly would not be the first time we would have seen IBM add – and then continue to support – a competitive product line in the database division. This will play better in some scenarios than others, of course: while it’s easily conceivable, I think, to see MySQL and DB2 sold alongside of one another, it’s more difficult to envision IBM actively pushing sales of Solaris at the expense of Linux. More likely is that the business would become more niche and transition focused for IBM.
Q: Speaking of Solaris and its open source foundation, OpenSolaris, what would be the future of the open source projects under IBM? While IBM has tactically embraced open source in specific areas, they haven’t made the same broad based commitments to open source their entire portfolio that Sun had, correct?
A: Correct, in that IBM has been very upfront about its commercial focus, and made no commitments whatsoever to open source the entirety of its software offerings.
But the situation is mitigated to a certain extent by open source. Say, for the sake of argument, that IBM decided to decommit or sunset Sun products that competed with their own offerings. This would be highly unusual for IBM, which generally supports products for eons, but assuming the worst case, the projects would be free to make their own way in the world with the communities that grown up around them (with the exception, obviously, of projects that remain encumbered), governed, in all likelihood, by the rights granted by their original license.
So if, as would seem likely, IBM intended to decommit from or at reduced its investments in Solaris/OpenSolaris, that community could continue to work on and develop the product under the original terms of the license. While many projects might suffer initially from negatively impacted commercially sponsored development, its certainly not out of the realm of possibility that new businesses could form around the other assets, per the terms of the license. This is happening in the MySQL community to some extent already, and it would seem the likely outcome for Solaris should an IBM acquisition proceed.
Q: Ok, let’s change tacks: if we assume that IBM genuinely has interest, what do you believe the motivation is from?
A: Well, there are opportunities across the portfolio, from expanding and accelerating the Java presence to trying to use MySQL to undermine their rival Oracle, but I think it’s generally about the breadth of intellectual property that Sun controls. If that IP was attractive in the past, it will perhaps never be more so, what with the economy generally depressing the value of potential targets and Sun, specifically, being significantly impacted.
That said, there are many who see this as an attempt to outflank competitors such as Cisco and HP. Others, such as Om, believe that Cisco might be a better landing spot for Sun, and I’d agree that Sun would make a lot of sense there.
Q: Speaking of Cisco, John Chambers has said in the past that culture is a huge part of acquisitions; how would you see IBM and Sun meshing?
A: Poorly, candidly, and that would be one of the most problematic aspects of a potential merger. Assuming, of course, that IBM intended to retain significant portions of the workforce. Not only are the corporate cultures different to the point of potentially being inimical – Sun being engineering led, IBM being profit oriented – even the engineering departments would be in conflict. There’s been bad blood in the Linux v Solaris camps for years, the Power folks might not appreciate the Sparc guys and vice versa, and so on.
Q: What about Forbes’ Ed Sperling’s contention that Sun doesn’t remain viable as an independent player?
A: Well, Sun’s performed poorly, clearly, in recent quarters, but to date has seen less investor unrest than might have been expected under similar circumstances, though rumors do circulate about that and related matters. So are they economically viable going forward? With a few billion cash on hand and a recently reduced workforce, the answer – at least in the short to medium term – has to be yes, I think. But beyond that, it depends on the firm’s ability to execute on several fronts: 1.) cloud, 2.) conversion of user to customer, 3.) increased sales process efficiency (after years of problems, this still remains a black eye for the firm).
Q: Anything else to add?
A: Not really. There are dozens more questions to ask and answer, of course, but it’s Friday afternoon. If I missed an important one that you’d like answered, you know where to find me.