Sun's commitment to R&D, developers is too easily dismissed
Between the keynote "inquisition" of Sun's newly installed CEO Jonathan Schwartz at Gartner Symposium and the press conference that followed, I had a lot to think about on the flight home from San Francisco yesterday. To be fair, the Gartner analysts (Paul McGuckin and Darryl Plummer) who put Schwartz on the hotseat had every right to ask the new chief executive how he plans to put the company on a profitable track. After all, IT executives who are looking out for their own futures as well as those of their companies rightfully prefer to work with stable and solvent IT suppliers. Sun could very easily pare back its R&D and interest in developers and become more of a commodities player like a Dell or an HP. But to spend nearly 60 minutes of a 60 minute interview rephrasing the same profitability question over and over and over again in hopes of tripping the always on-message Schwartz up was not only embarrassing, it was disrespectful. Not of Schwartz, but of the IT buyers who come to Gartner in order to figure out their next move.
If we want to see Schwartz get pan-fried to Paul Prudhomme-blackened perfection, then we can tune-in to one of Sun's quarterly results conference calls. Wall Street analysts love to punish Sun. But, at an IT event for IT strategists, it would have been better if McGuckin and Plummer spent just a bit of time on the financial stuff and the rest on picking Schwartz's brains for the sort of details that are relevant to enterprise IT strategies. After all, when you've got an executive on stage that oversees a $2 billion per year hardcore IT R&D budget in front of a hardcore IT audience, then might be a good time to ask that executive to walk through the Top 10 highlights of that budget and why those highlights might be relevant to the way an IT strategist should be thinking. Would that be setting Schwartz up for a bit of shameless promotion? Sure. Are IT buyers saavy enough to mine relevant information from such a dialogue? Absolutely.
Like any media-trained on-message IT exec however, Schwartz did his best redirect Plummer and McGuckin's line of questioning. But, as happens with many inquisitors of Sun, the importance of R&D as well as the developer community that Sun passionately cultivates (as evidenced by the 20,000 or so attendees that were across the street at JavaOne) seems to get too casually dismissed.
Sun could very easily pare back its R&D and interest in developers and become more of a commodities player like a Dell or an HP. Not that those companies don't have R&D budgets of their own. But the two are in a blood bath for marketshare in what are clearly commodity markets. If you're like Dell or HP and have figured out how to squeeze a respectable profit out of such a low-margin business, then there's no shame in being in those wars. With its Opteron-based servers, Sun is actually in one of those wars. But the fact of the matter is that neither Moore's Law nor the Internet work in favor of these particular commodity markets. Lest you think I'm inventing some argument to defend Sun and Schwartz, I'm not. I'm stealing it from a source that Gartner knows very well: Itself.
At the same event in San Francisco, Gartner research chief and distinguished analyst Steve Prentice said:
Like breakfast cereals, you can now buy a computer in your local supermarket. From the way prices are falling, there will probably come a day not to far away where you'll even find the computer in the cereal box. Disk storage is rapidly becoming a commodity. I can buy a terabyte of storage on ATA drives for around $300 if I shop around. The prices are falling steadily…..Looking ahead, we can see that commoditization is climbing steadily up the entire technology stack.
The "toys" that come with today's cereals or Happy Meals are sometimes more sophisticated than the PCs we once used. I'm sure Sun would love to have Kellogg's or MacDonald's as customers. But not in that way.
The Internet is another counterveiling force to largely commodities-driven technology plays. In his own advice to Symposium/ITxpo attendees, Plummer was crystal clear that the service model of software delivery is the way of the future when he issued the recommendations "Don't own. Rent" and "Don't buy applications. Buy solutions based on service." As you can see from the other blog I posted today, I'm totally with Plummer on this and very much sold on rented solutions based on my own experience. But if that's truly the direction things are heading, then participants in cereal-box commoditized markets that are looking to make up the difference with customers who are owing apps and insourcing are in for another rude awakening.
Dell for example offers some of the best scale-out solutions and has often downplayed the importance of scale-up. Not surprisingly, a scale-out strategy plays to both Dell's strengths and weaknesses. In the category of strengths, Dell, like other sytems integrators (I'm sure Dell would rather not be called that) largely relies on Intel's R&D for the core server technologies such as processors, chipsets, multiprocessor architectures, and on-board RAID that are the gating factors in scaling up. You can't hit an R&D homerun like Java without striking out a bunch of times. And to be fair, there's plenty of business in the scale-out market to be had which is good for Dell. But, the minute enterprises and businesses begin to heed Plummer's advice en masse, the demand for scale-up systems to support the multi-tenant architectures that are run by application landlords like salesforce.com will go up, the number of IT providers who can service the needs of those landlords goes down, and the difference in who gets those deals and who doesn't will very likely depend on who has the better R&D. Salesforce, by the way, runs on Sun.
But as long as someone else's R&D is at the foundational extremes (and anywhere in-between) of a stack along which commoditization is creeping (as Gartner says) and as long as that R&D is equally available to the competition, differentiating on technology, cost, or both will become increasingly more difficult. In essentially saying that today's commoditizing forces are going drive today's systems into cereal boxes and recommending a turn to rented apps on the one hand, and being critical of a strategy that plays well into both trends on the other, I felt Gartner was being a bit disingenuous in its quest for a good stage show.
Is Sun's $2 billion per year too big? Is there some fat on the bones that needs trimming? In the press conference after the drubbing he took from McGuckin and Plummer, Schwartz admitted as much and said that, as chief executive, it was now one of his top priorities to work with CTO/R&D EVP Greg Papadopoulos to make sure that Sun's R&D investments are well aligned with the company's business strategy. That said, R&D is like venture capital. I don't know what the industry standard success/failure ratio for venture capital is, but I do know that you can't hit an R&D homerun like Java without striking out a bunch of times.
Even if choosing between an R&D-driven versus a commodity-driven strategy is like picking the lesser of two evils, can those who are hypercritical of Sun's devotion to R&D really believe that the company has any alternative? And it's not as if Sun is ignoring the commodity markets. The company claims to be the fastest growing x64 server manufacturer and is just as platform agnostic as HP, Dell, and IBM if not more so between its support of Windows, Linux, and Solaris (with Linux support launching for its Niagara-based systems this week).
The other oft-dismissed facet of Sun's strategy are the developer communities it cultivates. I'm not talking about just any old developer community. I'm talking about a developer community with so much passion that when you walk down Howard Street between the North and South Halls of the Moscone Center where JavaOne is being held, the buzz is palpable. Particularly when compared to the near state of sedation that was underwhelming Moscone West across the street where Gartner's event was taking place.
I'm not implying that other developer communities such as Microsoft's, RIM's, or Linux's are any less passionate. But here's a connection no one bothers to realize: it was Sun's R&D that ultimately resulted in that large community that's incredibly passionate about Sun's technology (Java). Not only could you feel the passion for Java oozing off the 20,000 (Schwartz-estimated) JavaOne attendees, you could sense the affection they had for Sun and some of its key people like James Gosling who, as the person that's widely recognized as the father-of-Java, is practically a cult figure.
Sure, between the hundreds of thousands of Java developers and the bazillions of servers, desktops, notebooks, handsets, gas pumps, smart cards, and car dashboards out there running Java, Sun has had suprising difficulty monetizing it (or, as my colleague Dan Farber so aptly puts it, "intercepting demand"). Not to pick on Dell, but where's the community of people that are willing to stick by Dell through thick and thin? The only people who are passionate about Dell's technologies the way Java developers are passionate about Sun's technologies are Dell's employees.
Which is better for Sun? Going the developer-based ecosystem route that cultivates loyal communities who are willing to stand by you or going the commodity route where there are fewer places to hide as commoditization creeps right through your stack? Let me put it another way. The only thing reason the PalmOS has hung on this long (despite the Palm hardware company now loading Windows Mobile onto its hardware) is its developer community. How much additional R&D value accrues to Sun by virtue of the innovation taking place in the developer communities it cultivates? I'm not saying that the PalmOS's pending death is what awaits Sun. But, what I am saying is that the industry's strongest and most vibrant vendors (including Intel) derive a significant amount of their strength from the developer communities that form around them.
Java was not the first developer community that Sun cultivated and it won't be the last. Somewhere in that $2 billion of R&D lurks another a developer community that could change Sun's fortunes. For example, this week, at JavaOne, Sun took the wraps off a skunkworks project called Phobos. Phobos, as a part of Sun's larger effort to jumpstart dynamic language support for scripting the Java Runtime Environment, is a technology that allows developers to script server side Java (eg: J2EE) with Javascript as opposed to Java. For those of you not following along, the JRE has long been an operating environment that was the exclusive domain of Java programmers. Much the same way .NET can be programmed with C# or Visual Basic, Sun is making it possible to program Java with languages other than Java. Think Python. Or Ruby. And Javascript (which, despite its name, had nothing to do with Java until now). [See: Sun turns Javascript into Java. Google turns Java into Javascript]
Server-side Javascript? I thought it was dead. Sun's director of Web technologies Tim Bray agreed that Phobos may not be a game changer. But it does allow developers who are already writing client-side Javascript (eg: AJAX) to work with one language. We've seen what AJAX can do on the client-side and at first blush, perhaps there's no application for it in server-land. But then again, there's some AJAX developer out there that was probably inspired by Phobos to do something that will blow our minds. Or maybe an entire community of "Local AJAX" developers will form around Sun's incarnation of Derby (known as JavaDB). These are just a couple of the many swings Sun is taking at the plate, any of which could eventually produce another community and perhaps Sun's next home run.
There's another aspect of these communities that Sun cultivates which goes unheralded too. If Sun is spending $2 billion per year on R&D, then how much additional R&D value accrues to Sun by virtue of the innovation taking place in the developer communities it cultivates. While coming back from a long day at Gartner, I decided to pop into NetBeans Day at the Argent Hotel (this was Monday, the day before JavaOne started). It was late in the afternoon and I didn't have any JavaOne or NetBeans Day credentials. Not that it mattered. The main room where they were talking about innovative technologies that don't even come from Sun (like the Apache Foundation's Maven) was so packed to the gills with developers I couldn't fit in.
I had flashbacks of taking the #4 subway from Grand Central Station to my job as a systems analyst at Swiss Bank Corporation at New York City's World Trade Center (back in the 80's) and how the subway cars were so packed with people, you had to use your feet to compress the person waiting on the platform in front of you into the subway car. A breakthrough technology -- either the killer Java developer tool or the killer Java application that the world is waiting for -- that will send Java even further into the stratosphere than it already has been sent was probably hiding in the head of one of those bodies that was filling the room. Or maybe not. But what are the chances without such a community. Who in Dell's community could end up producing the breakthrough innovation that changes Dell's fortunes? That was a rhetorical question. And that's the model some people want Sun to move to?
Sun absolutely must produce a profit at some point and deserves little slack for taking as long as it is taking to get things turned around. But to even subtly criticize it for its emphasis on R&D and developers is just easy pickens for bullies who don't understand the value of either. Gartner said it. Not me. The whole stack is getting commoditized. Perhaps Lenovo has it right. The only refuge at this point for a commodity systems provider is in a low cost Chinese workforce. Here in the US, the cost has been wrung out of everything else. The choices Sun is making may not ever produce the turnaround Sun and its stockholders are looking for. But I can promise you that Sun is nearly guaranteed to strategically fail if it chooses any other path than the one it's on.