I had the pleasure of speaking this week at the Open Source Business Conference in San Francisco (my enigmatic slides are here, if you’re interested). The conference was well timed, set against the backdrop of Oracle’s audacious rollup of open source firms. Larry Ellison was The Man Who Wasn’t There. Oracle’s acquisitions, not surprisingly, are met with mixed emotions in the open source world, stirring excitement, foreboding and befuddlement. They appear at once to validate the robustness of the open source model and to call its future into question. Marten Mickos, CEO of MySQL, the leading open source database company, had the funniest response to Ellison’s strategy. He said that Oracle was “trying to kill a dolphin by drinking the ocean.” (MySQL’s corporate emblem is a dolphin.) It was a good line, if not entirely convincing.
One of the main topics of conversation was whether open-source “communities” could survive in the clutches of a Big Software company like Oracle. The answers hinged on how people view the nature of those communities. Some believe they are essentially communitarian: what glues them together is a passion for big abstract concepts like “freedom” and a distrust of traditional software companies. They’re rooted in a shared ideology. Others believe that the communities are essentially utilitarian: what their members really want is good software cheap. They’re rooted in a shared business need.
I’m of the latter view. While open source may have had its origins in ideology, it’s now a pragmatic movement. The communities have left the commune. I think the success or failure of Oracle’s open source strategy will therefore turn not on how it treats communities as communities but on how it treats communities as customers. If it positions its open-source offerings as loss leaders, using them as lures to lock up customers and then pick their pockets, it may end up discovering that corporate software users aren’t as tractable as they once were. Another of Mickos’s comments at the conference, while not as witty as the dolphin crack, feels like a more accurate description of the risk Oracle is taking: “Their DNA is not one of commodity software, where prices are low and volumes are high. It’s about their culture and tradition. It’s a mindset question.”
I will be most disappointed if Jboss is sold itself to Oracle. As you said mindset and shared common values and vision are the core of an open source movement and that is also the core reason for users flocking to it.
But in this capitalistic world no one can stop anyone else.
Lesson learned is as a user of any of these open/closed technologies, we have to prepare ourselves to migrate from one another with ease.
For example, we are ok as long as we can switch from a hotmail to gmail to some Xmail as and when they show their real intent!
Nick, 2 points. I wrote about Oracle and SleepyCat – the fact thta more than technology, the fact that Sleepcycat has learned to implement “Dual Licensing” is far more interesting to Oracle.
Also thanks for sharing your slides. Your vision about Utility Computing is right on – but as I have written before – somewhat optimistic short term because the IT Utilities (IBM, EDS etc) have the scale today but cannot or will not price themselves to be attractive – or provide reliable enough performance
Vinnie, re: utility computing. I agree with you. The problem is that today’s leading outsourcers are not utilities. They simply maintain private IT plants for individual companies. Their profit model is based on the fragmentation and underutilization of IT assets. That model will take time to overturn, and it may well be a new set of companies that ends up pushing through the change. Nick
Your comparison between today’s IT industry and the electric power is appropriate only to a certain point. I believe it applies to the future consolidation/centralization of physical components of IT infrastructure — hardware, networking components, etc. Software is where the analogy breaks. For the consumer of electric power, the delivered ‘service’ can be characterized by a couple of parameters, voltage, frequency, max current, etc. While infrastructure software is getting commoditized, it is far from being characterizable in a utility-like fashion. Even if the world standardized on just one version of one OS (pick Windows or Linux), the combinatorial possibilities of configurations is huge. Note that ‘hosting’ software does not change this because the need for multiple configurations arises from the differing needs of applications/end users. The inherent nature of software does not lend itself asymptotically to a service with ‘n’ parameters where n is small. It may be more feasible to do this for a software service at the business level, but this is decades away. Two key points here: I think that the open source phenomena is independent of utility computing (i.e. would have a similar impact whether or not the IT industry moves to a predominantly utility-like model, though is coincidently helping the economics of this tranformation); you’re right on, on the hardware side of the equation, though it might be a few years before we refer to servers as CPE (taking an analogy from another utility-like industry) :).
Thanks for sharing your slides re open source and the move towards utility computing. I agree to a large extent, however I view that standardization of software interfaces through technologies such as XML is ultimately going to assist in the development of utility computing to a much greater extent. Note, I do not mean that XML itself will be achieving it, but rather industry specific flavours of XML – for example FpML (www.fpml.org).
To take it back to your analogy, it was the selection of specific voltages and frequencies that made electricity utilities possible (the ‘XML’ flavour), rather than the generators generating the electricity (the ‘software’, which could be open source or not).
Evolution of electricity into a reliable utility is a powerful illustration of direction towards which enterprise computing “infrastructure” has to move towards.
In terms of impact on enterprise IT function, I like to characterize this as the “doughnut effect” that will remove the infrastructure and common applications from within the enterprise. Beyond economies of scale in production, the other economic drivers are modular global outsourcing, need for innovation networks, consumer’s access to superior ubiqutious & pervasive computing resources in addition to technologies like SaaS, grid etc.
I would also agree with the earlier comment by (Raj Sehgal) that needs of users are too contextual, granular and hence it is only inherent “programmability” of software (i.e. applications) that can address this diversity of computing demand. A simple, universal interface to all computing needs is utopian simplification. Clearly this very thing that makes innovation opportunities truely boundless in IT.
I heard the buzz somewhere that the Oracle-JBoss deal closed last week for $485m…
I think Oracle is fully aware of the commoditization dynamic OSS is driving up the tech stack and the consequent blur its creating between software and service revenues. fyi – of licensed dB’s running on Linux, Oracle has an 80%+ share so they’re pretty warm to working with OSS. About 60% of their revenue comes from support services, so they know the value services drive. Also fyi – they recently hired a chief evangelist for OSS. They’re quickly sizing up the OSS force. I think their moves reflect a trade-off between the uncontrollable risk to their proprietary dB license revenue, versus the controllable risk of baking in OSS IP strategies into their business model.
From the community angle, deep technology firms tend to understand how their business needs to align with the efforts of world class engineering talent. The best developers today are passionate about working for OSS projects. What better way to win their favors than to compensate them through acquisitions for their labors, enable them the freedom to continue tinkering away in pursuit of their technical vision, and provide a global service infrastructure to allow them focus on sustained IP creation.
That’s where I think Oracle is heading. fyi – I work at Oracle, not on the Tech side of the business, but the Apps side. Am nowhere near privy to what Larry’s thinking right now, but am continually awed with his business sense, appetite for risk and ability to keep Oracle a step ahead in the game.
I had the urge to rebut mySQL chief Mickos’ view about Oracle cited above. Here’s my post on what I think Oracle is trying to do with its Open Source acquisitions. I do work at Oracle.
Oracle’s buying binge should have little long term impact on the open source community. In fact, the community seems more vibrant than ever to me. While promoting our new service I’ve been impressed with how quickly the community is adapting to the new business opportunities utility computing offers.