I’m pleased to learn of the download rates and apparent adoption of Sun Microsystem’s Solaris 10 as a new alternative in the Open Source world, and even more pleased to read some of the terrific reviews Solaris 10 has garnered in the developer community.
However, I’m a bit baffled by Sun Microsystems President Jonathan Schwartz’s recent note about an Executive Advisory Council he facilitated with some key customers to find out the state of the world for IT developers and architects. In his weblog article, he waxes poetic about how so many of the attendees had been working with Solaris 10 and how there were multiple “developers who admitted their teams built, tested and qualified apps on Solaris, given far better tools and utilities, then ran them on Red Hat to appease those who’ve tied their reputations to it.”
Um, of course if you contact your customers and ask them about their development process, it’s quite likely that you’ll find they’re early adopters of Solaris 10 and are trying to figure out how it integrates into their existing development environment and architecture. Indeed, these very same customers might well already have Solaris 9 running on some SPARC hardware and Solaris 10 is just a logical extension. Certainly it seems a bit forward for Jonathan to suggest that it’s a harbinger of change in the industry and that “The bloom is off the rose with Red Hat.”
But what was most astonishing about Jonathan’s posting is his ending note where he highlighted how out of touch Sun Microsystems has been with its key customer constituent these last few years as the company has lost its way. In his article, he ends with this quote:
It seems so phenomenally obvious to me that if you’re selling a development platform, your key customer is always going to be your developers. Indeed, why bother switching from a proprietary operating system (Solaris 9) to an Open Source operating system (Solaris 10) in the first place, if not to appeal to the increasingly Open Source fanatical developer community that are the core of your customer base?
This is one of those obvious facts of business life — your customer is the center of your company and great companies are customer-centric — that it seems odd to see it presented as some sort of strategic epiphany. Don’t you think that Sun’s competitors have been very focused for the last few years on what their customers need, how to deliver it, and how to ensure that their customers are having great results? And maybe that’s why Sun’s market share keeps declining?
In some sense, this seems to highlight the danger lurking in a corporate blog too, doesn’t it? After all, if this message were part of an internal memorandum to senior management, everyone else would have no clue that prior to this meeting Sun was even less focused on its customers.
Thanks for your thoughts on Solaris. Before I moved over to the Solaris kernel organization last year, I used to write speeches for a couple of senior execs at Sun (I hated it). And I did a little work on the Executive Advisory Council. It was always the senior execs hanging with other senior execs from our biggest customers. Pretty special club. I’m away from that stuff now. 🙂 But I hear that this EAC was very different. Schwartz wants to talk directly to developers, and we’re really psyched about it because it brings our issues right to the very top. For many executives, it simply wasn’t this way, but Schwartz talks to people at multiple levels of our customer base — which is extremely encouraging for our developer programs. I’m not sure this means we were less focused on our customers previously. I think it’s more of a recognition that our conversations with customers need to be much more comprehensive.
The OpenSolaris program is growing rapidly and will represent a massive change for Sun. Solaris, after all, is really the guts of the company, and a great deal of growth is occurring. Some of it painful and complex, but it’s all for the good and we are starting to make real progress. I think we’ve learned a lot from our other open source projects, and I also think they’ll benefit from the trend OpenSolaris sets. If developers and customers don’t feel we’ve given them enough attention, OpenSolaris will provide a new platform and program for them to interact with us much more closely. This is already happening in the OpenSolaris Pilot Program we are running. The conversation has started, pilot participants are working with the code, and we are building the infrastructure and roadmap for co-development. Should be a wild ride moving a thousand Solaris engineers and 10 million lines of code (kernel, networking, libraries) to an open source model to directly engage with a market that we are simultaneously turning into a community.
PS: I love the kanji, by the way.
Thanks for your note, Jim. I appreciate your inside information about the Executive Advisory Council, but just like with any other source of competitive information, I can’t help thinking that having people who aren’t Sun customers would be invaluable at these events, rather than having them just be the insular ‘customer feedback session”. Or some hybrid, since some of the topics discussed are doubtless only relevant to customers, but if I were in Jonathan’s position, I’d be darn curious about two classes of people who aren’t represented at EAC events:
1. Companies who are former Sun customers but now run different hardware / software solutions, and
2. Companies that have never been Sun customers but match your target customer profile.
What’s their story?
I know that your team already does this, goes to industry events and talks with lots of management and IT professionals, but it sure seems to me that you’d gain much more insight from the people who say “no” than the people who say “yes”.
Oh, and one unsung advantage of Solaris 10 moving into the Open Source world that hasn’t been covered in the publications I read: more laptop options for geeky engineer types. 🙂
I thought I’d add my two cents as a long-time customer, a manager of both development and operations teams, and an attendee of the recent developer focused EAC. While I understand your interpretation of Jonathan’s statement, I can speak from personal experience as a customer when I say that my experience with Sun at many levels throughout the organization has always been one that represents their interest in what we’re doing, what we have to say, and how they might change or enhance their strategy to serve us, the customer, better.
Part of what is so interesting to me about the open source movement is the shift in the paradigm around decision makers within an organization. How many applications or frameworks exist today that have proliferated quite quickly through organizations that wouldn’t have been used otherwise (due to cost in a non open model and licensing issues) and probably would have been needed to be built in-house instead? Or would have been built in house out of convenience and bureaucracy avoidance that may have been necessary in the past (Last i checked, most developers aren’t generally allowed to sign checks). Developers are increasingly influencing decisions on software platforms (which eventually translates to some extent to hardware platforms as well – I hear Sun makes those too) and are more involved in the decision making process (or lack of process) in many organizations today. Sun’s change in direction may be due to increased focus, versus focus at all as you imply, on areas that have always been important to Sun but weren’t necessarily in their direct focus at the highest levels. The analogue between the old EACs and the new applies here. Instead of selling SPARC in general at a very high level (which I’m pretty sure still goes on as well), Sun seems to be trying making the integrated stack from a development AND production perspective far more attractive than it has been in the past. I frequently debate how helpful the era of the virtual machine is for education in the engineering space (as new comp. sci. grads seem to come out with less and less low level hardware knowledge), but from a business perspective we’re getting closer to being able to commoditize the hardware and the stack and everything in between which allows us to just worry about what matters most to many of us – building applications that best support our customers and our business.
Note: My comments are my own opinions and speculation and are not necessarily representative of the opinions of my current employer, or Sun, or anybody else. I hope you and your readers find this additional context useful.
Great note, Justin. Thanks for taking the time to stop by my weblog. Your comments about the commoditization of hardware are spot on with my own thinking about the evolution of the industry, but I can’t help but think that if you let your development teams make decisions based on what’s out in the Open Source community rather than what you actually need, strategically and tactically, to succeed in your business space, aren’t you just letting the inmates run the asylum?
I realize that’s not exactly what you’re saying, but as we business writers say, “paradigm shifts are a big deal”, and if you indeed see that it’s all changing, then aren’t you saying that in the past the head of IT or President would say “we need X. Make it so” and now they aren’t? That replacing that is the post-paradigm shift of “we need X, let’s just cobble it together, lads” of the software teams in the firm?
Perhaps you can clarify your thinking for us?
Not being stuck in airports means I don’t get much time to read/post. I’d love to clarify my thinking for you – here are some additional notes.
I’m definitely not seeing any shift away from “we need x. make it so” – but if you see that happening somewhere, let me know. What I am saying is that developers are increasingly more influential in the decision making process regarding overall platform and infrastructure, which i suspect just might drive direction when the whole stack (from hardware up through software) actually becomes commoditized. On a micro level, one example might be IDE selection . Does the head of IT or the president care about whether or not the developer is building custom workflow or using struts? If it affects delivery, then yes. Otherwise, unless the head of IT is very involved from a technical perspective, i think the answer will no and the question will just be how long?. A few questions need to be answered of course – does it scale? how is licensed/what does it cost? how long? etc… but with positive results here it becomes simpler. Fast forward to more commoditization now – and we have much lower barrier to entry, maybe you no longer need lots of high end expensive sparc hardware, you can run the Sun stack that better enables your developers to deliver cost-effective, scalable, and elegant code that is focused on your application and is rolled out on a much more aggressive timeframe. Mark Lucovsky writes on software delivery at http://mark-lucovsky.blogspot.com/2005/02/shipping-software.html This speaks to traditional process and how it changes at “internet speed” – and is extended further from a software development perspective when we can make more assumptions about the platform and the stack per my comments above. While it may be particularly relevant to service oriented businesses today, it will extend to create efficiencies in software development process in many areas.
Great stuff, Justin. Kinda sorry you’re not stuck in more airports. What you post begs the question: what happens to Sun hardware when Solaris 10 really takes off and becomes the de facto development environment for the Solaris community?