SAS07 – Trip Report, Suite Dreams, More than Java, Fine Glass Cups, and The Battle Against Punk IT

I’m not much of a vacation photo taker, so there’s little visual evidence of Kim and I’s stay in San Francisco last week. The first half and last day of our stay looked like this:

While the middle part was more along the lines of:

In truth, I wasn’t disappointed.

I was however disappointed in one of our rainy day adventures, watching The Messengers. (A zombie movie wasn’t to be found.) I was shocked to find out that it was the number #1 movie of the weekend; of course, it could be worse. As I told Kim, I’m placing a moratorium on American-made haunted house movies. They’re just all terrible. Asian film-makers, esp. the Japanese, seem to have cornered the market in quality ghost movies of late, even getting into some playful Lynchian feels. And, as The Messengers shows, you can’t just import it, sprinkle some “American” on it, and expect the movie to be good.

On another note, thanks goes out to all the folks who suggested restaurants both in the comments and in email. While we had tons of good food at all sorts of places, the best in my book was on the last day on the way to the airport:

For those wondering what “yuh!” means, there’s a picture of it ;>

But, enough with the pictures and glib-linking.

Sun’s Suite Dreams

Time and again, the most repeated theme in conversations I had was at the Sun Analyst Summit this year was a return and reiteration to what I call “suite dreams” thinking. The idea here — kept alive over the years by large vendors who can afford to engineer, maintain, and market it — is that buying your IT from one vendor will produce the most optimized experience. In the world of Sun, as one table buddy put it, imagine if you start all the way from a Niagra, go up the Java compiler and hot spot optimized around that architecture, developing in NetBeans, and then ran the resulting software on Sun servers and storage.

The idea of playing to the performance strengths of a “closed loop” environment are intuitively compelling: of course everything will be optimized to work together if it comes from the same vendor, the thinking goes. Who’d know better? There’s also a marketing appeal: you can hook the customer on the “thin edge” of software, or pull them in through hardware.

In that sense, as we’ve always felt at least, Java is Sun’s primary advertisement. They’ve always been miffed that people don’t appreciate Java that way and, instead, want to see the revenue breakout for Java itself. It’s like asking Absolut, Coke, or Toyota how much all those ads are bringing in.

Now, the suite dreams notion wasn’t a high-flying theme, but a sort of assumed discussion point. It works well from a cross-marketing stand-point as the high flyers in the suite dreams tool chain can pull in the “low flyers”: you need it all, best of breed be damned!

Open Loops

Ironically, as a Java developer, the idea of “suite dreams” rubs me the wrong way. As I told several people, the message I get is that I’ve got to lock myself into the Sun world if I want “the best” Java software. While I wouldn’t characterize that lock-in as Sun’s ostensible intention, it may be their revenue ambitions to “just happen” to be the stack that people choose. (This is the point where The Usual Java World Suspects start poking fun at Sun. Of course, they have stacks to sell as well.)

Along those lines, as I talk more and more with the Java and tools folks at Sun, the spirit of openness and expansion of the Java platform seems to get more and more buy-in. The vision and strategy being: welcoming, embracing, and helping non-Java languages and mind-sets flourish in the Sun ecosystem. My hopes and expectations are that over the next year we’ll see this expanded vision play out enough to make a judgement call on it’s viability. Keep your eyes on JavaOne.

Commidification and the Indirect Path to Black

Of course, Sun’s suite dreams are also a sly jiggling of the commidification lever, which is always a tricky dance partner for Sun, as it is all the elder companies. The grand vision is that commidification rises the IT tide and, of course, all boats will be lifted. That’s conceptually a good lemonade-from-lemons judo move. At the same time, there’s a sly “but Sun servers will be better than those Dell boxes, friend” once you’re on the mat.

While hearing about the “brutal efficency” of Sun hardware, I kept thinking of the nice, Norweigon tumblers and matching wine glasses I have. I could have just as easily gotten a bunch of plastic (or even glass!) cups from Wal-mart, but I sure do like those Norse beverage-vehicles better. Put another way, my high school economics teacher tried to give us some parting advice for our college years: “don’t buy a fancy car, it’s just four wheels to get you from point A to point B.” Little wonder that economics isn’t one of the sexier sciences…

Clearly, Sun’s recognized that that they can’t continue selling high-high end servers only, as, for example, the recent partnership with Intel and the continuing one with AMD shows. To me, the story for Sun hardware is: “sure, Dell is cheap, but we don’t fail.” To that end, I often wonder if it’s worth the time to turn up the “nasty dial” a bit about Dell and other “low-end” options for servers.

There’s a certain gentle-person’s code of honer when it comes to bad mouthing along those lines. Though, the spate of green energy ads from all sides are implicitly throwing mud at products that aren’t. Most people I talk with feel that Dell is the Wal-mart of servers, despite their enterprise efforts. They buy Dell because it’s cheap.

There’s also a cultural gap to speed-fill when it comes to quality expectations in hardware. Perhaps I’ve just been drinking a lot of Sun and IBM kool-aid of late, but it seems like most people are waking up to the notion that punk IT isn’t always the best choice when you have the money for some wing-tips. In my gut, I get the sense that most people in the punk IT area “don’t know better”: the idea of getting paged frequently to restart a server or service seems “normal.” While the anecdotes from getting involved in Sun’s Startup Essentials program evidence a move against punk IT, following in Google’s foot-steps and deploying “crap boxes” is still the “cool” way to do IT. Sun, and others, would probably offer up their first born to see a headline like: “Google Replaces White-Box Server Farms with…”

The afore mentioned x86 partnerships and offerings are Sun trying to meet that demand half-way, and as their profit being in the black shows, something seems to be working. As was pointed out, it’s a nice way to segment out the market, hopefully not too late. Perhaps it’s like Doctorow’s Law of IP: once you actually package your offerings how people want to consume them (“mid-end” x86 vs. high-end SPARC), they’ll spend where previously they didn’t. And don’t worry yourself about seconding guessing what the customers want: product management hubris has no place in hardware.

We’ll see how it looks next SAS.

Disclaimer: Sun is a client.

Technorati Tags: advertising, hardware, netbeans, punkit, sanfrancisco, sas2007, servers, suite, sunw