Jarrett House North

I spent the day today bouncing back and forth between sessions that were a little closer to our product’s core functionality (service desk, asset management) and actually doing product demo. I had a demo via Webex that ended just twenty minutes before I had to check out of my hotel room.

An aside to the Gartner folks: please try not to stack the ITXpo on the same week as JavaOne again. There weren’t any non-smoking rooms available in any hotels, and my congestion is killing me after three nights (and one two-hour demo) in the sour stale smoke in my room at the Renaissance.

Below is a partial transcript of new Sun Microsystems CEO Jonathan Schwarz’s Q and A at the Gartner Symposium.

Q: Let’s talk about StorageTek — loyal following but what’s next?

A: Talked to a board member who lost their tapes. I said, “What’s the big deal? The tapes are encrypted, right?” But tape data isn’t encrypted. So the new StorageTek products support encryption, tie into the public key infrastructure…that seems like a simple requirement, a great synergy.

Q: Thumper (codename) is the best and worst of Sun. There are so many platforms in Sun that do what Thumper does.

A: Three weeks ago we moved into a new set of opportunities. the thing about Thumper is it runs across all Sun’s R&D. and we want to see that happen over and over again–look at Southwest. They’re the most successful airline because among other things they run only one airplane, the 737.

Q: What about your other businesses? Sparc and x86 and … A: We have to meet our customers where they are.

Q: How do you characterize your customers? A: Developers don’t buy things, they join things. They love creative disruption. It’s like social experiment. IT has some of that too, but the difference between them and you is that you have money. Developers love free. You guys hear free and think “free puppy.”

Q: If developers love free, how does that affect your business? A: Developers are the leading indicator. Java started with developers. Now it’s in the enterprise.

Q: But how do you move from a billion handsets with Java and 5 million Solaris downloads to revenue? A: We will add value. The amount of value that Sun receives is directly tied to how many of you use our customers’ services.

Q: What’s the value of Java to Sun? A: What’s the value of a standard plug in the wall to the generator business? Q: Oh but what about investment in R&D? A: GE doesn’t have to worry about someone plugging in a vacuum cleaner that won’t let its electricity expand. And Java is fundamentally the Internet, which drives all our business. Java is about modular architecture on millions of devices. The more handsets with Java, the more the Internet is important and the more they need our support to build the architecture. It’s hard for the Street to understand, but revenue is a lagging indicator of the developer’s adoption of Java. Q: How lagging? A: Well, the day someone starts using Java, Sun is relevant to you. Q: But how much will they pay you for Java?

A: There is a division in the world–those for whom IT is a competitive weapon and those for whom IT is THE competitive weapon. We want the latter, and they will pay to support the infrastructure that underlies that competitive weapon. Q: What about your former customers, those who are running Java but not on Solaris or Sparc? What is attractive about that to you?

A: If you look at the numbers for middleware and servers, we are the leading performance, the best reference platforms for Java. We can deliver a better developer environment. Secondly, OS consolidation. Every large enterprise has cats and dogs. we can offer Solaris that runs on every major platform and is remarkably scalable. Third, we can offer the systems infrastructure (Opteron and Niagara) to affect the productivity and efficiency of their data centers.

Lots of argument about whether Sun is giving away too much value. Q: How do you translate the value from Java into people writing checks? How do you translate the technology lead into marketing? A: You have to understand that while I am deeply focused on financial performance–I need to be focused on the long run.

Q: But Sun lingers too long on Solaris, on Linux… A: Do I get to play the new guy card now or later? Let’s think about mainframes. I went to talk to a now-bankrupt large airline one time about their mainframe, and the guy said, “Son, we installed the mainframe before you were born and will retire it after you die.” Infrastructure sticks around.

Q: But how does Sun manage to exploit its innovation? A: You look at the airplane industry. The engine makers mint money, the airplane manufacturers sometimes make money, and the carriers never make money. Money accrues to fundamental R&D.

So ZFS is a 128 bit file system–do you need 128 bits? Maybe not but you might need the 65th bit. Solaris with ZFS eliminates the need for RAID controllers…

Q: So let’s talk about open sourcing Java. Why is open sourcing Java without forking so hard? A: Look at the history of standards. Volume is the primary driver and standards bodies are secondary. And Microsoft is the biggest platform out there and Java on Microsoft is rife with litigation. So we have to be concerned about forking if we open source. Interoperability is hugely important to our customers.

Q: Top three reasons that Sun will be more successful in two years? A: 1. Leading indicators–lots of new customer interest. 2. Market opportunity–the need for reliable scalable infrastructure and solid development platforms isn’t going away. 3. Management team. And we have boomerangs coming back all over the place. I talked to Ed Zander yesterday on stage at JavaOne and asked if he wanted to come back, and he said, “You got a job for me?” Now I don’t know if he was serious…

…across the street in the other part of the Moscone Center, to be more precise, Sun’s new CEO Jonathan Schwarz announced at the JavaOne conference that Sun will release the Java source code, and is asking developers for feedback on how to do that in the least disruptive way.

Wow. Cool. Guess I was at the wrong conference for Really Big News this morning.

Spent some time this afternoon looking at Vista and talking to the Microsoft team here (which includes some folks I know from my past life at Microsoft.com and my internship—hi Arvind! hi Peg!) about what’s coming down the pipe that our company needs to know about from an application development perspective. The guidance I got from the team there is that the major thing to pay attention to is the change in the privilege model—User Account Protection—and how that affects the installation and running of applications. Other than that, there are plenty of cool new features to take advantage of, of course. And the eye candy is impressive.

I also liked the built in RSS widget on the Sidebar. It does appear to be a little funky though—not in parsing the RSS, which is fine, but in loading it in chronological order rather than reverse-chronological order. I loaded my RSS feed for kicks on one of their Vista test machines and was surprised to see that the top entry was an old one—then surprised again to load the feed source in IE 7 and to see the same thing. Apparently the default XSLT+CSS orders the items oldest first. That kind of detracts from the usefulness for me. Maybe there’s a way to change it. I walked away with a beta CD and will check it out once I can install it on Virtual PC.

If you know me, you already know one of my points about a few of the sessions that in a well-intentioned and generally thorough way address the question of “consumer” technologies in the enterprise. That’s the C-word itself. Our employees are not gullets that crap cash, and talking about technology expectations set by “consumer-grade” services like Google, desktop search, and IM that are all actually free brings some real cognitive conflict. How can it be consumer technology if there is no money changing hands?

The Docs (Searls and Weinberger) et al have done us a tremendous service in making us aware of the naming problem that the “consumer” label brings. I’m not sure they’ve done a good job of identifying alternative labels. “Producer” has been floated in the case of bloggers and other user-authored media creators, but it doesn’t generalize well; “human being” and “citizen” have the opposite problem—they’re too general. From a technology perspective, is there a useful way to talk about technologies that stand in opposition to the enterprise central-control model that doesn’t use the C word?

This is, I think, an important question. Anyone who has been around “consumer” driven businesses knows it can be like pulling teeth to get them to acknowledge that consumers are the same people who are inside enterprises, just seen at different times. And some of it has to do with the label. If IT organizations are to take “consumer” technologies seriously, as the Gartner mavericks suggest that they should, maybe finding a different word is a good starting point. The session I just was in, led by David Smith and Tom Austin, indirectly suggested one alternative: open market technologies. But of course this suggests that enterprise technologies are not open market.

The question of what to call it (reluctantly) put aside, the idea that systematically bringing non-enterprise technologies inside the enterprise could drive real benefits—not just to the end user but to the enterprise as a whole—is I think worth taking seriously. The speakers cited a case study of a company (which I think was Ford) which hardened its internal systems, then piloted an approach in which rather than providing a corporate desktop it gave its users a stipend with which they could buy their own—shifting the burden of administering and supporting the desktop to the end user, but also allowing them to take advantage of rapidly shifting consumer technology. They also discuss possible worlds where the base OS on an office machine is a consumer OS, and that the standard corporate desktop runs as an image inside a virtualization environment. How silly it is, they point out, to be a consumer service like Google or Ebay and to say that you can’t do business with me unless you have an approved, hardened browser that I provide and can guarantee is secure.

What I found interesting about all of these scenarios, and what they pointed out toward the end of the discussion, is that these trends could open the door for players like Apple—not iPods but Macs—to be dragged into the enterprise by end users who are comfortable and productive with them and can do most if not all of their jobs on them, once the users are given the leeway to provide their own desktop machine. Now that’s interesting.

One of the pitfalls (or blessings, depending on your perspective) of being a small software company is that you get laser-like focus on your core business problem out of necessity. For me as a vendor, one of the real value points about the Gartner show is getting exposed to other market segments that touch ours that I might not run across otherwise. That’s the case with the session I just attended on BSM (Business Service Management).

BSM is essentially a live dependency map, integrated to monitoring tools, that escalates only the monitoring events that have a real impact on the business and presents them in a business-consumable format. This is a goal for a lot of IT organizations—I know the Microsoft.com operations team was trying to implement something like this using the Microsoft server product stack plus homegrown tools about five years ago before the BSM market was grown, and knowing them they now have a complete solution. What was interesting to me was how BSM seems to dovetail with the work that my company has done in the last year on the CMDB, which really is about creating and documenting the service dependency map that BSM needs as a starting point.

When you combine a good CMDB with robust change management, and then tie in a good monitoring API and logic about how component status rolls up (or doesn’t) to the status of a service, then all of a sudden the time and effort spent on building that CMDB has paid some unexpected dividends.