Wednesday, January 27, 2010

The Oracle Has Spoken


The Oracle-Sun merger press conference just ended. Some trends made themselves pretty evident and a few questions do remain.

What does it all mean for ComputeSpace? It it time to get weaned from Sun, SPARC, Solaris, and Java?

No. At least, not yet.

Here's my take on what was said today:

Oracle is primarily concerned with vertical integration. They want to consolidate ownership of the entire computing stack from applications at the top to microchips at the bottom. They want to be what IBM was about the time I was getting into computing 40 years ago.

Whether that turns out to be a viable strategy or not I can't say. But I will say that, having been involved with computing from mainframes to minicomputers to distributed systems, to server/client PCs to clusters to grids to clouds...

I have to ask what were the circumstances that made developers move away to more diverse, distributed, open standards-based approaches in the first place? And what's changed and what hasn't?

Oracle's president made it clear that they view Sun as a value-added Oracle platform for the high-end, high-value enterprise.

That leaves thousands of small, innovative groups like ComputeSpace, below their radar in a way that simply wasn't true with Sun. We're not likely to see anything like the Sun Startup Program over in Redwood Shores.

In the 21st century they want to turn relatively complex machine architectures into pre-configured machine templates. A model being borrowed from virtualization and, frankly, a technique we are emulating.

Our test deployment will be small for the next 6 months but is being designed to scale as large as necessary. By perfecting the deployment architecture, we'll "freeze dry" that configuration image to deploy on additional servers in the cloud as we expand. That gives us standardization and ease of maintenance and fewer production bugs.

The road map for Java is a little more interesting. You'd expect this from a software-oriented company. Lots of good things in development. Also they're moving Java functionality down into the SPARC hypervisor environment. That means java machines running apps without an intervening OS layer.

Our app server is Glassfish v3 and Oracle is keeping it for "the low end" (departmental services). The flagship will be Weblogic.

They're betting a lot of JavaFX (as are we) and integrating Java, JavaFX, and JavaScipt to work as one. Putting fine-grained security on web pages so we can secure the individual javascript or JSP code blocks in the page.

OpenOffice will get a release as Oracle Cloud Office which was an ongoing Sun project. As it competes with Google's offering we're not likely to take advantage of it until the day we move out of our Google Site.

Finally, a solid port of Oracle VM to SPARC could be useful for us as we start to get more traffic. We'll probably be scaling into a private cloud of some sort and that's the place to deploy our freeze-dried images in SPARC VM containers. It would allow us to use our current equipment for a while before making new purchases.

So, the Oracle has spoken, the Sun has set, and Night has fallen and the jungle seems menacing. What's next?

I'm lighting my Promethean torch and moving forward. Everybody follow me.

Monday, January 25, 2010

Brave New World


As the ComputeSpace Project eases into the final stretch before deploying our first version of the application, I'm looking back to where we came from and also forward to the challenge of actually changing the way teachers and students, administrators and school boards, concerned parents and concerned investors see the future of science education.

About half the team (Werner, Jeri, and myself) have spent a lot of time in enterprise computing, which is a much different environment than consumer-facing, retail computing. We're used to making software for businesses or science institutions to function, not apps to entertain you or make your desktop useable or allow you to personalize your experience.

The Good Thing: Because of that enterprise background, we've paid careful attention to making ComputeSpace work. Not kinda work, but really work. Work dependably, reliably, efficiently, effectively.

Don't get me wrong. We faced a fair amount of pressure to just "get the thing kicked out into the street" for schools to start using. That would be nice because I can't wait to see it running in school districts all over rural America, but the two things we didn't want to happen --making schools do our QA for us (yes, Microsoft), and getting drowned in a torrent of bug-fix requests-- we think we've avoided by holding firm to the idea that, like Paul Masson wines, we will sell no app before its time.

This is a complex software application being presented as SaaS (Software-as-a-Service) and, while I don't expect it to get out to the masses bug-free, I'd be happy if the ratio between bug-fixes and feature requests in our Bugzilla were 2-to-1 in favor of the latter.

And, in looking to the future, ComputeSpace has a real opportunity to change the whole ballgame if we do it right and we're in the right place at the right time. There are lots of features in it that I can't talk about publicly yet, but also a lot of thought about execution. There are lots of features that are in a lot of other software that's out there: Starry Night, WorldWideTelescope, Google Sky, etc. With respect, we don't believe they're doing it right. And that's our aim.

A close friend who's an engineer recently noted that, by clinging to this madness of an elongated development and testing schedule, I'll eventually have something to really be proud of and he hopes I enjoy all the peanut butter sandwiches I'll be living off of.

He may be correct. But, having been a teacher, if teachers see it and say "This is really engineered with my needs in mind and not just to make a buck off me"; having been a parent, other parents can say " I have real faith in its ability to engage and excite my child about learning astronomy"; having been a student, other students can say " This is crazy intriguing!"...

...then that sacrifice is worthwhile. And my response is "O, brave new world, that has such sandwiches in it!"

Subscribe Now: Feed Icon

About Me

My photo

If you're looking for the Court Cannick who's architecting ComputeSpace for education or the Court Cannick who built NetworkMathematics for systems management or the Court Cannick who engineered the storage systems for SETI@Home or the Court Cannick who produced Physical Science Journal for Storer Cable or the Court Cannick who lectured on Space Colonization at CMSI, then you probably have found the right Court Cannick.  Otherwise, keep looking...