When Android debuted in 2007, I couldn’t figure out how Google had managed to apply an Apache license to the project. Java, like Linux, was governed by the GPL and thus incompatible with the more permissive license Android was sporting. Stefano Mazzocchi subsequently answered the Java related questions: Google wasn’t using Sun’s VM, they’d built their own. As had Danger before them, from whence many of the Android team arrived. Called Dalvik, Google’s cleanroom reimplementation was, if not “Sun’s worst nightmare” as Mazzocchi put it, a clear fork-in-the-eye to the Java license holders. However brave a face they put on it at the time.
Whether Google decided to reimplement the JVM for financial reasons, technical reasons, or both, is unclear. Whatever the motivation, Dalvik allowed Google to bypass Sun en route to market. What Dalvik never did – never could have done – was protect Google from patent litigation.
In estimating the risks of such action, Google could have reasonably assumed that the probability of Sun suing them was near zero. Sun may have been unhappy, and may even have suspected that Google’s cleanroom reimplementation was anything but. The Vegas line would still have been decidedly tilted against Sun turning to legal action.
Maybe Sun’s reluctance to sue was financial. I’m personally skeptical of this claim – companies with failing financial fortunes in my experience are generally more inclined to seek legal remedies to their problems, not less – but Shankland’s sources are always good. Even if this were the case, however, Google couldn’t have assumed that would suffice as a shield. Instead, the search giant would have expected Sun to behave consistently with its past behavior and future interests. Besides the fact that Sun had effectively zero history using its patents offensively, as James Gosling acknowledges, there was the fact that attacking Google over patents would irreparably damage Sun’s then nascent efforts to repair its fractured relationships with developers. It would be tough to reconcile a suit with the public positions of its chief executives. Worse, it would have injected fear, uncertainty and doubt into the dominant enterprise software ecosystem at a time when Sun could ill afford it. While its pursuit of patents remained high even then – for what former chief counsel for Sun Mike Dillon characterized as defensive reasons – offensive use of its intellectual property was more or less unheard of, the Microsoft settlement notwithstanding.
Sun, in other words, was not going to sue Google. And Google knew it.
It’s safe to assume, however, that Google also knew that Sun was unlikely to be the permanent owner of its intellectual property given the firm’s financial trajectory. Which is why Google’s legal team probably started preparing for a suit the day the Oracle transaction closed. If not earlier. And it also explains why Google was prepared enough to fire back with something more than a “no comment,” dismissing the claims as baseless and cleverly reframing this as not just an attack on Google, but on open source Java as well. Oracle shares neither Sun’s old qualms nor its conscience, and Google knew that too. Finances are not an obstacle, and Oracle does not care about perceptions of the company, developer or otherwise. After yesterday, it would appear that they are similarly unconcerned about what the ecosystem thinks of their stewardship of Java.
The latter point is perhaps the most important. It’s the only real clue we have to answer the only real question here: what does Oracle want?
Because the answer to that is: not what they’ve asked for in the complaint. Oracle may indeed request recompense for “the damages sustained and will sustain” as well as “any gains, profits, and advantages obtained by Google as a result of Google’s acts of infringement and Google’s use and publication of the copied materials.” But you can be sure that that’s not all they want.
As Andy Updegrove covers, the obvious motivation is financial. Specifically, maximizing the return on the six and a half billion capital expense that bought Sun’s assets, the patents in question included. If Oracle realized the same return as Sun from the Microsoft settlement concerning Java, for example, the cost of Sun becomes four billion. Remaking what was arguably a bargain into a steal.
Purely financial justifications for this suit are less than satisfying, however.
To begin with, Oracle would effectively be trading long term ecosystem health for a short term cash windfall. Unless the settlement is historically immense – a difficult outcome to rely on from a planning perspective – it’s not clear that this would be a net win. For all of its sustained success in the application and database markets, Oracle remains as fundamentally dependent on the Java ecosystem as Sun was before it. Even for a company that’s sought and found growth through stack ownership and category dominance, the health of the ecosystem is and must remain a concern. While the original technology was technically groundbreaking and differentiated, the key to Java’s success lay outside its featurelist. What drove its ascension within enterprises was the reality that Java offered at least the potential for independence from vendors. That will not be surrendered lightly, whatever Oracle may believe. A Java ecosystem dominated by Oracle is a doomed ecosystem. While it’s far from clear that this action by itself would create that perception amongst current Java ecosystem participants, it, coupled with Oracle’s own aggressive history, would be unlikely to be beneficial from a participation standpoint. As Andy put it, “it’s less clear to me what the strategic value would be to Oracle to prevent Google for incorporating Java into Android, or to impede the marketplace generally from relying on Java.”
Nor is the outcome of this suit certain. Granted, Oracle’s lawyers will be belligerent and numerous. Given the nature of both the patent system generally and the patents at issue specifically, it’s not only possible but likely that Android infringes in some fashion. This is perhaps not surprising. News that the patent system as it pertains to software is broken is not, in fact, news (coverage). It is also true that the fact that Oracle is proceeding in the face of obvious, substantial costs both financial and not suggests a level of confidence in the merits of their suit. What will be interesting to observe, however, is how the suit is or is not complicated by the post-Bilski patent landscape as well as the public promises of senior Sun executives not to assert their intellectual property rights in this fashion. Groklaw discusses the Bilski implications in their coverage. The Sun executive blog entries, meanwhile, cannot supercede inherent intellectual property rights, of course. But might they be used to build a case challenging the intent of the rights holders? Perhaps. Can patents once regarded to be “open” be “closed?” Barring a settlement, we’re likely to find out. If they cannot be closed, and Oracle’s suit fails, the costs, in both dollars and damage, will be exorbitant. Even before we get to the inevitable countersuit which will result from the asinine mutually assured destruction game the current system forces upon us.
It can be argued, then, that this is a high risk exercise for Oracle. The only satisfactory return for high risk exercise is high reward. Based on past software settlements, it’s difficult to project this being material to Oracle financially over a multi-year timeframe. Which is why I suspect there’s more at stake here than royalties.
What that is is non-obvious. All that we know about what Oracle wants, realistically, is what they are prepared to surrender. Aside from bearing the hard costs of litigation, Oracle is willing to absorb soft costs in risk to reputation and participation rates in the Java ecosystem. We must expect then that Oracle’s expected return will be commensurate with these costs. Oracle is many things, but stupid generally isn’t one of them.
Perhaps, as Forbes speculates, this is a prelude to a cross-licensing arrangement. Though if that’s the case, I’m far less certain that this suit actually has anything to do with Android; might patents like this “Large-scale data processing in a distributed and parallel processing enviornment” or this “Information extraction from a database” be relevant to Oracle’s core businesses? Perhaps Google is already or plans to compete directly with Oracle in ways we are not aware of yet. Or maybe Oracle just wants Google to buy a bunch of database licenses.
Whatever the real reason, this is a surprising decision even for a firm as aggressive as Oracle. The only thing more surprising is how quickly it turned Google – excoriated around the web for their questionable net neutrality proposal with Verizon – back into the good guys. Even if you speculate about differences in Oracle’s evaluations of its own assets – that Oracle believes that Java has peaked in popularity, for example, and that this is a one time opportunity to cash in on an asset that must, inevitably, decline – the calculus of this move fails. Nothing in Oracle’s product roadmaps hint at such a realization. Nor would a one time windfall, however large, be sufficient to offset the costs of a significant decline in Oracle’s Java related products.
As for predictions, I’ll make only one: whoever wins will also lose. This suit is going to negatively impact – probably substantially – Java adoption. The enterprise technology landscape is more fragmented by the day, as it transitions from .NET or Java othodoxy to multi-language heterogeneity. Oracle’s suit will accelerate this process as it introduces for the first time legal uncertainty around the Java platform. Apple and Microsoft will be thrilled by this development, and scores of competitive languages and platforms are likely to see improved traction as a result of Java defections.
Add up these costs, and the only supportable conclusion is that Oracle’s ambitions here are big.
- The complaint
- A subset of the patents at issue: 6125447, 6192476, 6530080, 6910205, 6061520
- Andy Updegrove
- Charles Nutter
- Miguel de Icaza
- Stephen Shankland
Disclosure: Oracle has in the past been a RedMonk customer; Google has not. Of the other mentioned firms, Apple is not a RedMonk customer while Microsoft is.