Late yesterday afternoon I got a request from the folks at Sun for some time this morning to discuss a “material” announcement. The catch? It was before 7 in the morning, my time. Given that I generally don’t stir before 8:30 at the earliest barring emergency, I hoped it was big news. In spite of the unholy hour, I wasn’t disappointed.
Some thirty seconds after my presence status went from Away to Online this morning, I was lit up with half a dozen different pings either alerting me to the news or asking my opinion on it. While I have very few lucid opinions at 6:20 in the morning, this was clearly of interest.
Generally speaking, I prefer to let a day or two pass before writing up events of this magnitude – the whole “you heard it here last” tactic – to allow myself time to digest the news and the various reactions to it. Having spent the majority of my day doing little aside from reading such reactions to the news, however, I’m comfortable breaking that particular rule.
Thus, the Q&A.
Q: Before we begin, do you have anything to disclose?
A: Yes indeed. Both parties in this planned transaction are RedMonk customers, Sun as a Patron and MySQL as a Supporter. Additionally, several firms that compete with either or both entities, such as EnterpriseDB, IBM and Microsoft, are RedMonk customers. A couple of firms likely to be impacted by the deal, such as Oracle and Red Hat are not. Lastly, we are also MySQL users, as our main website along with our blogs and other properties are all MySQL based.
Q: Ok, dumb question: how is MySQL pronounced?
A: It’s usually pronounced either My-ess-queue-ell or My-see-kwell. Given that the MySQL folks I’ve spoken with in the past have tended to use the former, that’s how I say it. The latter seems to be more common, however.
Q: For the few people in the audience that haven’t heard the news, can you explain what was announced this morning?
A: Certainly. Sun CEO Jonathan Schwartz was not terribly coy about the news, with the first sentence in his blog this morning closing with “we’re acquiring MySQL AB.” The valuation is approximately $1B US.
Q: Isn’t that valuation a bit high? What was the multiple?
A: Given that MySQL was a private company, actual revenue figures were consigned to the realm of speculation. One analyst on this morning’s call estimated $60-80M, which obviously puts the acquisition multiple up there. I heard a lot of folks questioning the valuation this morning, and I’m sure that more will in the days ahead. My guess is that the acquisition price was as much a function of the potential returns of a public offering as it was the current revenue picture, but that’s all it is: a guess. All of that said, understand that I am not a capital markets expert, and my coverage doesn’t extend to that arena. Like Tim, I still find it counterintuitive – on a superficial level, anyhow – that it cost Oracle better than eight times that to acquire BEA.
Q: Yeah, that was quite a coincidence. Is there any possibility these two announcements had something to do with one another?
A: Not that I’m aware of, and it seems unlikely. The contrast between the two, of course, is worth noting. As James put it this morning – I’m paraphrasing so as to not put words in his mouth – BEA is very much the traditional software model, while MySQL is more the upstart, disruptor.
Q: From a strategic rather than tactical perspective, why did Sun – who’s been in the black but is still not setting the world on fire – place such a big bet on an entity that is known as much for its poor customer conversion percentage as for its ubiquity?
A: The slides accompanying the announcement hint at the answer, calling MySQL “The Platform for the Web Economy.”
Q: So does that make this “We’re the Dot in the Dot Com,” the sequel? With the commensurate crash to follow?
A: Some would contend that’s the case, yes. Particularly a couple of members of the media we’ve spoken with. And of course it is possible that the model propping up Google’s immense valuation and those similar to it will prove to be similarly illusory. But somehow I doubt it. The fact is that the Google’s of the world have made real what Sun itself could not: a network that is, in fact, the computer. And the Google’s of the world, far more often than not, run on MySQL. Via this single acquisition, Sun’s made itself a relevant vendor in a space that very few, if any, of the larger commercial systems suppliers can play in.
Whether you agree with the valuation or not, YouTube sold for $1.6 billion, and consumed virtually no software from any of the major vendors. If that acquisition was to take place today, they would have been buying from Sun.
Q: What do you mean?
A: Simply put, open source generally and MySQL specifically have been very disruptive to their respective markets. Speaking of the database market in particular, the incursions into the traditional high end accounts owned by the likes of IBM, Oracle, and Microsoft have been fairly limited. But the growth everywhere else has been nothing short of spectacular. Here’s how I’ve put it previously:
Consider the problem that faces virtually every major software infrastructure vendor in existence at the moment. The startups during the bubble – and I should know, as we worked with many of them – were almost universally deploying on top of quote unquote “enterprise” technologies on the advice of their VC advisers: EMC storage, Solaris for the OS, Oracle for the DB, and so on.
Today, the vast majority of the startups we speak with are running on one LAMP flavor or another, and occasionally Microsoft a la a MySpace. Assuming for the sake of argument that the folks who bought IBM and Oracle in the past will stick with that, but startups eschew those technologies in favor of free and open source alternatives, where does that leave us? The mature, sizable and fairly static market goes to the old guys, while the new and growing market goes to the new guys. Obviously it’s not all that simple, but that’s the worry – or should be – for most of the major software vendors on the planet. While their installed base has and will continue to make them money, it will not give them the growth that Wall Street tends to demand; only access to greenfield, growing markets can do that. And to date, those markets have traditionally been won by open source technologies.
We work with many larger firms on just these sorts of questions, and part of the challenge is identifying assets that will be of interest to growth accounts. Well, there can be no question that MySQL is just such an asset, one that Sun apparently felt was worth a billion dollars.
Q: Are you contending then that MySQL is restricted – either by strategy or technology – to these down market accounts? That they’ll never be a player for the high end relational database workloads?
A: Remember that never is a long time. On the MySQL side, they have steadily been adding the kinds of features higher end users consider a requirement – e.g. arrival of stored procedures and triggers in 5 – but I doubt that even the MySQL engineers would contend that it’s an easy swap out for DB2 or Oracle. That said, it is the basis for some truly massive workloads at Web 2.0ish customers, so it isn’t as simple as saying “MySQL doesn’t scale.”
In my opinion, however, the key realization from the MySQL side – and a fundamental component of their approach – was the understanding that high end database purchases are only partially about technology. MySQL could introduce tomorrow a complete equivalent to Oracle, completely free and open source, and I can assure you that migrations to it would be painfully slow. Part of that is resources, part of that is the trust earned by the larger database vendors over decades, part of it is the hassle, but any way you evaluate it the conclusion is simple: direct competition is difficult. Not impossible, as EnterpriseDB is demonstrating – more on them in a bit – but difficult.
So rather than play the head to head game, MySQL has cleverly jujitsued the disparaging Oracle comment of “MySQL is a Toyota Corolla to our 747″ by embracing it. Why? Ask yourself how many Toyota Corollas there are on the planet relative to the number of 747s, and their strategy becomes more clear.
Q: So if Sun gets from MySQL access to important new accounts, increased open source credibility, and a stake in the LAMP game, what does MySQL get from Sun?
A: In a word: resources. Accepting that Sun is nowhere near IBM size in either its services or support arms, the fact is that Sun has a vastly wider reach and resources than does MySQL. On a call this morning, the Sun folks were quick to bring up the 17,000 resources in sales and service from Sun that MySQL can now draw on. While it’s counterintuitive to think that MySQL needs help with its reach, given its widely agreed upon status as the most popular database (at least relational) on the planet, it is nonetheless true that larger, more conservative buyers would prefer to buy from a Sun over a MySQL. Big businesses have a difficult time, generally, trusting smaller businesses with anything but tactical pieces of technology. MySQL could have pursued the size required inorganically via the public offering mechanism, but apparently decided that this path would be quicker and at least as rewarding to its shareholders.
Q: What does Sun actually know about the database space? Aren’t they all Java, Solaris, and hardware?
A: Those are the highest profile product lines, yes. But Sun – unbeknownst to many – has actually been in the database space for a little while now. Although I don’t have revenue figures available to me, it is a minor concern compared to the properties you mentioned, but Sun has offered commercial support for both JavaDB (their flavor of the Apache Derby codebase) and PostgreSQL for some time now – both of which are open source relational databases.
Q: So they both compete with MySQL?
A: Well, JavaDB less so, but certainly PostgreSQL is very often viewed as an alternate to MySQL.
Q: Sun will be dropping support for PostgreSQL, then, correct?
A: No. That was the common assumption, but everyone I spoke with this morning assured me that support would continue even post-acquisition. Subsequent to that, Sun’s Josh Berkus – a PostgreSQL committer – not only confirmed this confirmed this, but was reportedly gracious in his welcoming of MySQL developer Brian Aker to the Sun family. The reasoning being from a Sun perspective that the popularity of the respective projects varies widely by geography; some regions prefer Postgres, some prefer MySQL. Sun intends to sell into both.
That said, it would be illogical to expect the commitments to the products from a resourcing or marketing perspective to be even roughly equivalent, given the billion dollar capital outlay and the retaining of the MySQL staff.
Q: Which brings me to my next question: what is to become of MySQL’s staff?
A: This was a key question for me, as having gotten to know the MySQL folks both before and after they were customers of ours, I’ve got immense respect for execs like Mårten Mickos or Zack Urlocker, marketing folks like Steve Curry, or those on the engineering and community side of the house like Brian or Kaj Arnö. MySQL is a disciplined and well run company – particularly on the engineering side, but more importantly one that has impeccable community relation abilities. When was the last time, for example, you heard anyone complain about the fact that MySQL signed a deal with SCO? Or that they effectively forked the project and include non-distributable code in their Enterprise edition? It’s probably been a while, because the MySQL folks have been up front and transparent with their thought processes, and this has stemmed all but the most strident of criticisms.
If Sun’s focus, then, was purely on the code and the account access, I think they would be setting themselves up for serious future problems. Fortunately, from what I’m told, the structure and people are to remain in place: in Zack’s words “Marten will continue to run MySQL inside of Sun. The founders and management team are on board. I’m personally excited about being a part of Sun and helping to take MySQL and open source to the next level.” Just as importantly, folks on both sides of the aisle seem – to put it mildly – enthusiastic about the deal. Initial reactions from Planet MySQL and blogs.sun.com seem to validate this.
Whether or not the cultures will mesh as well as everyone seems to think they will is obviously an important question, as there are always challenges in bringing even two positively disposed organizations together, but the early indications are good.
Q: Speaking of acquisitions, is Sun particularly good at them? Or is this to be another Cobalt?
A: The answer to the first question – as my colleague noted during a long ago Press Q&A to the chagrin of many of the PR folks in the room – is no. Sun’s track record for acquisitions is decidedly mixed, though certainly better than Google’s of late. The only firm that seems to do this consistently well is IBM, which is almost certainly related to the fact that they tend to acquire smaller companies that were previously close partners.
That said, I – unlike some commenter’s this morning – don’t see a great many parallels with the Cobalt situation apart from the size of the deal. Cobalt was obviously an ill fit for a firm whose engineering core was passionately committed to Solaris, while MySQL – the Postgres conflict aside – highly complementary to the existing portfolio. Moreover, this is not an unproven product or a gamble: MySQL is the M in LAMP, something that virtually every employee from the executive team down to the engineers has respect, if not affection for (ok, maybe not so much the Solaris guys vis a vis Linux, but you get the point). Although distrust of the firm itself continues, most folks in the open source world understand and appreciate that the vast majority of developers within Sun are pro-open source. Seriously pro-open source, in many cases. That contingent, therefore, understands just how important MySQL is to software generally and open source specifically. And as for the hardware side of the house, MySQL’s nothing but good for them.
In short, the high profile of MySQL publicly and the lack of serious internal conflicts make this a much better fit, I believe, than Cobalt ever could have been.
Q: Given Sun’s commitments to Solaris, isn’t it likely that we’ll see a decommittment from MySQL to Linux?
A: According to Mårten, the order of preference from a platform perspective for MySQL is Linux, Windows, Solaris, in that order. While actual percentages were not provided, Linux is apparently the leader “by far.” For that reason alone, it seems tremendously unlikely that Sun will jeopardize the marketshare and ubiquity of MySQL – its primary strength – by trying to push customers towards an operating system they aren’t looking for. Will they opportunistically poach customers? Sure. But force the issue? I doubt it.
Sun derives considerable revenues from hardware remember, and if there are questions about their agnosticism on that basis, it’s useful to remember that they OEM Windows. In his post commenting on today’s news, MySQL’s Arnö firmly emphasized the need for continued independence on both the operating system and programming language fronts, saying “I don’t expect that in any way to be at the cost of other popular operating systems (Linux, Windows, Mac OS/X, other Unixes etc.) or development environments (PHP, Ruby on Rails, Perl, Python, ODBC, C++, C#, VB etc.). MySQL grew with LAMP and MySQL without LAMP at its core is simply unimaginable.”
For Sun’s part, Jonathan said this morning, “We realize that MySQL is the M in LAMP, not the M in anything but that.”
Q: What about licensing fears? Is Sun somehow going to “take back” the MySQL code? Or make it closed source?
A: To answer the first question, the answer is that they cannot legally. MySQL is licensed under the GPL, and the important part of the license at this juncture is the reciprocal provision, which essentially ensures that the code currently available must carry with it the rights and privileges currently assigned to its users in perpetuity. What Sun could do, legally, if they chose to would be to close the source and keep future updates to itself, which would leave users and customers with access only to what’s currently available.
Given that this would be brand suicide for both MySQL and Sun and would likely provoke an immediate fork from one of Google, IBM, Oracle, or Yahoo, I think this unlikely in the extreme. Forget the fact that this is inconsistent with the general direction of Sun’s software which is towards open source, it simply makes no sense from a business perspective. Instead, I’d anticipate business as usual from a licensing perspective.
Q: Is it all smooth sailing, then, for MySQL?
A: Not quite. Though it hasn’t been as frequent a topic of inquiry as I expected, questions remain concerning the previous acquisitions of InnoDB (by Oracle) and Solid (by IBM), the vendors behind two of the more popular storage engines for MySQL. Inquiring on the subject today, I was told that Sun expects – as has been the case with MySQL previously – business as usual with their respective partners. Even prior to the acquisition, however, MySQL appeared to be hedging its bets with the development of its Falcon storage engine. In short, this is an area of potential concern, but the worst case scenarios here do not appear to be particularly dramatic, particularly given the open source nature of the code in question.
Q: Should MySQL and Sun have any other technical concerns?
A: Years ago, Adam Bosworth decried the limitations of the relational databases of the time, and the message is as relevant to MySQL today as it was then. The relational store is and will continue to be a fundamental piece of application infrastructures, but there’s a reason Amazon’s cloud based database SimpleDB is schemaless, Oracle bought Sleepycat, and that IBM hired Damien Katz to work on CouchDB, and it has something to do with the less is more concept. Not to mention Joe Gregorio’s MegaData.
This isn’t to say that MySQL will be caught flatfooted – Mark Atwood’s S3 storage engine for MySQL is an intriguing demonstration of what’s possible – but rather that my way too early prediction that relational stores will increasingly be competing with non-relational stores is beginning to come true.
Q: Apart from the primaries, who are the winners and losers from this?
A: I don’t know that either could be characterized as a loser, but both Oracle and Red Hat could be negatively impacted by this particular decision, albeit for different reasons. Oracle, for its part, has been eyeing MySQL’s rapid growth at the lower end and green field markets warily, to judge from the commentary we hear from customers. While not directly threatened – part of MySQL’s tactic, remember – neither were they thrilled by the rise in popularity for the open source database, and many contended this was the primary motivation for the acquisition of InnoDB. Thus a MySQL that now has the backing of a much larger entity, and its commensurately larger services and support arm, is unlikely to please them.
Red Hat, for its part, is impacted more indirectly as I see it. Matt contends that Sun is now a direct threat to Red Hat for the open source king of the hill title, but I think the more significant problem is that MySQL is now off the board as a potential weapon in its contentions with both Oracle and Sun. One potential beneficiary from that scenario? The aforementioned EnterpriseDB (their CEO Andy Astor comments here), because if Red Hat is seeking an open source database alternative to complete its stack and compete with the larger entities, they’re probably the best remaining candidate.
Q: What have the reactions to the news been so far?
A: Generally positive, from my vantage point. As noted some have questioned the valuation, and others are reasonably skeptical, and both Glyn Moody and Matthew Aslett over at the 451 Group were decidedly mixed on the news. But reactions within the two firms have been almost uniformly positive, and outsiders like Andi Gutmans (Zend), Mark Radcliffe (DLA Piper), and Matt Mullenweg (WordPress) are all reasonably enthusiastic.
There is some fear, uncertainty and doubt – as should be expected given the magnitude of the deal – but I think most are willing to keep an open mind. See, for example, Jeremy Cole or Rafe Colburn. If you’re MySQL and Sun, that’s all that you can reasonably ask for.
Q: And what is your take?
A: In keeping with the above, I’m cautiously optimistic about the deal. Notwithstanding the (assumed) high valuation and the inevitable challenges of integration, this looks on the surface like a mutually beneficial combination. MySQL gives Sun relevance in markets it absolutely must have to succeed, not to mention an additional infusion of open source DNA, while Sun provides MySQL with the resources and room to take itself to the next level. It’s all dependent on the integration, of course, including the retention of key personnel, not fumbling the integration of MySQL into the sales process, and so on, but these are solvable problems.
At the very least, I’m very comfortable saying that this is nothing short of the most important acquisition I’ve seen from Sun in my career. All for the low low price of an eighth of what Oracle paid for BEA and a fourth of the dollars Sun itself doled out for StorageTek.
The synergies, as the financial crowd likes to say, are there: now it’s all about execution on both the MySQL and Sun sides.
Q: Are there any other questions that I should have asked?
A: Just the most important one of the day: given that RedMonk’s 5th Birthday Party is on the last day of the Sun Analyst Conference, will the MySQL folks be able to attend? We’d love to have them, so if you’re a MySQLer and are co-opted into attending SAS, do not – under any circumstances – leave San Francisco at its conclusion. You’ve got beers to drink that night, and given the valuation, you’re buying.