April 2007
Mon Tue Wed Thu Fri Sat Sun
« Mar   May »

Month April 2007

Ubuntu 7.04 “Feisty”

This morning I upgraded my “development” desktop, a Dell Core2Duo, from Ubuntu 6.10 “edgy” to 7.04 “feisty.” I chose a complete reinstall, since there were aspects of the old setup and disk partitioning I wasn’t happy with. The install just gets simpler all the time, with a graphical wizard that asks fewer questions than the old Windows XP install.

Most important from my perspective is the new VMI kernel support for virtualization, which replaces the need for a recompiled Xen kernel, and consequent difficulties in getting the NVIDIA drivers working properly. Video is working perfectly in 7.04 with the NVidia cards that Dell ships, and everything needed for virtualization was easily installable via apt-get. This afternoon I’ll install both a Linux and a Windows VM under QEMU and see how it works.

Ubuntu 7.04 also includes a package-installable Sun JVM, through the “multiverse” sources. I was able to directly add Java 6 with no problems and it also switched the default JVM for the system, which is nice.

I have to say, thus far I’m impressed by the strides Ubuntu continues to make in creating a powerful yet incredibly friendly Linux distribution. Word on the street is that Michael Dell is even trying 7.04 on his personal laptop…

My project at Microsoft: EMC Smarts and Operations Manager Integrated

Last year I was at Microsoft before moving north and heading back to graduate school, and much of what I worked on hadn’t yet been announced (some still hasn’t been). But recently one of the projects was announced by both parties, so I can mention my involvement. Microsoft has licensed EMC Smarts network management and root-cause analysis technology to integrate with System Center Operations Manager 2007 (formerly known as “MOM”). I worked on Vij Rajarajan’s team and was part of a 3-4 person task force negotiating the deal with EMC. I’m very excited to see it announced and moving forward.

I’m particularly excited because I think both technologies are strong and are architecturally an excellent match. SCOM 2007 is very flexible model-based management technology, and EMC Smarts is the best of the “new generation” network management systems that inherently model root-cause and network causality. The merger of the two, along with injections of SDM/SML model technology and additional vendor-specific models written by equipment vendors will create a very strong management solution that finally will do more than “ring the alarm” when a symptom occurs. Model-based management has been in the works for years, and in our own way at Network Clarity we attempted to do MBM for network configuration and optimization, but the latter is an even harder challenge to do well in some ways.

So congratulations to Rich, Vij, Shawn, Christian, Ashvin, and others on announcing this deal and making the strategy a reality!



An Evening at Coupage in Seattle

I had a terrific evening last night at Coupage, here in Seattle. The restaurant, located in Madrona close by the Hi-Spot, blends Korean and classical French cuisine, and is the first Seattle effort of Portland chef Tom Hurley, along with chef Rachel Yang. I recommend it very highly; last night’s meal was perhaps the best food experience I’ve had in Seattle in a long time — possibly since my first revelatory evening combing the menu at Lark.

Walking up to the restaurant along 34th, I could smell grilling meat a block away. Getting closer, it turns out Hurley had a Weber kettle out on the sidewalk and was grilling Kobe beef and some chickens as specials. His plan is to add more grill capacity, both here at Coupage and his new upcoming restaurant downtown. When he does, make reservations immediately because this man can grill.

I dined with Marc and Bill, a couple of friends from our tasting group and both aficionados of white Burgundy. They took care of white wines, with a “starter” Coche-Dury 1996 Meursault, followed by two Leflaive Chevalier-Montrachets: 1985 and 1988. All three wines were stunning, but for me the 1988 was absolutely a standout: dense, creamy, spicy, lush, but still possessed of a crisp minerality and good acid. Pretty darned near a perfect glass of wine. The whites were accompanied by the mache salad, dressed with a nice truffle vinaigrette, grilled maitake mushrooms and marinated bamboo shoot (the latter was savory and tasty and my favorite part of the dish). We also tried the wild mushroom Bi Bim Bop, a variant on the Korean classic and very tasty. We finished off the first course with the crispy pork belly — not my favorite of the three but still excellent.

As a “mid” course to finish up the whites we had the duck paparadelle pasta, which I thought was excellent. Throughout the meal at various times, Hurley came out and told us about the food, his philosophy for preparing food and running a kitchen, and I recommend talking to him. He’s led an interesting life and has the energy and passion for food you see in a rare few.

The “main” course was a family-style platter of the night’s special — grilled Kobe beef and grilled chicken. Both were superb, especially the chicken breast and the crispy end pieces of the Kobe. I served the Henri Bonneau 1990 Chateauneuf du Pape “Marie Beurrier,” which although a “second” cuvee for Bonneau (alongside the Cuvee Celestins), was a masterpiece. Deep, sweet, yet beefy and herbal, it reminded me strongly of the best bottles of 1981 Beaucastel — “Mourvedre cotton candy” was Parker’s descriptor for the latter, and although Bonneau uses very little besides Grenache in his blends, it fits. The man makes pure Grenache taste deep, dark, and complex like Mourvedre. Naturally, my stock of these wines is tiny, given availability and price, so this isn’t a wine I’ll taste again for quite awhile but I’m amazed at the experience. Marc opened a 1970 Jaboulet La Chapelle as well, but the bottle seemed to be a bit tired — clearly La Chapelle underneath a slight soy sauce layer.

We had a selection of desserts, but what stood out for me was one of the ice creams in my sampler: sweet chili ice cream. Just the faintest hint of a sriracha-like chili, which went well with the 1989 Von Hovel auslese that Bill brought.

In all, the evening was terrific — good friends, great food, and spectacular wines. I can’t recommend Coupage highly enough.

TransmissionLab Update

Yesterday I posted TransmissionLab version 1.4, a fairly major reworking of the model class core. I was dissatisfied with the way that RepastJ models, by default, seemed to tightly couple the main model class to all of the other classes I’d written for data collection, transmission rules, and population construction. My goal with TransmissionLab is really a framework for building models to study cultural communication and transmission, not just writing one giant model and bolting new stuff on.

A paper by Railsbeck et al., in the September 2006 issue of Simulation, is right on the money in saying that the original Objective-C version of the Swarm toolkit is a strong “framework,” as opposed to the “library” style of successor toolkits like RepastJ and MASON. Swarm definitely forced a style of organization onto your simulation models, via the concept of nested “swarms” of agents, observers, etc. I suspect this is much like Ascape, but the latter doesn’t seem to be an active development project any longer (at least given the website – leave a comment if this incorrect). Whereas Repast provides a ton of infrastructure but simulation models themselves seem to be fairly unstructured, as I read various examples and models folks have posted online.