I decided to surf email for a bit and re-found the link to this article my lead (what some companies might call a "boss" or "manager") sent around to the team. I'm pretty sure my lead is addicted to CNET-news. But I suppose there are worse things to be addicted to.

The article has a pretty decent explanation about our approach to management with the Dynamic Systems Initiative (DSI). However, I mostly disagree with the issues presented under the "Potential Hitch" heading. If the System Definition Model (SDM) serves as the "blueprint, or description, for how software and hardware components can be controlled" and system administrators are given that "blueprint" then system administrators will have some pretty "detailed information on the inner workings of a system". The SDM has the potential of transforming what used to be "black box" components thrown over the wall by developers into "glass box" components that can be deeply managed by system management systems.

I also would have liked to see some more discussion about the differences between utility computing (that term seems to be synonymous with "grid computing" and "on-demand computing" but I could be missing some finer details) and the Dynamic Systems Initiative. For example, if you reach the end of this paper written by Jim Gray (who has a really fun writing style and is really flipping smart) and agree with his postulations then it seems there are a limited number of applications that are "mobile" (to use Jim's term) enough to play nice in "grid computers". Interestingly this article, also on CNET-news and also sent around by my lead, starts off with the exact same example application that Jim uses in his paper: movie special-effects render farms. All other example applications provided in the article are theoretical implementations that are not complete.

Anyway, I independently came to the same conclusion Jim did a while ago and I guess that's why I work for Microsoft's DSI and not IBM's "On-Demand" or Sun's "N1 software". <grin/>