One of the best posts I’ve read in over a year on “Web 2.0″ and the challenges of managing large-scale services comes from Tim O’Reilly as a follow up to a meeting he had with Debra Chrapaty from Windows Live. There are several very interesting ideas that are surfaced in the post:

1. The “developer platform” of the future will include not only the APIs, but also the operational infrastructure of the platform itself. In this world, the developer not only builds on the platform but entrusts the hosting of his application to the “platform vendor” such as Windows Live, Google, or Yahoo.
2. Competitive advantages will not only come from operational technologies, but also from the ability of the operators to drive those innovations back into the core offerings for people still hosting their own systems. As an example, O’Reilly cites the enterprise features driven into Windows Server by the Windows Live group based on operational knowledge and their own development efforts.
3. Small players need not apply. In other words, only the Microsofts, Googles, Yahoos and perhaps Amazons will be able to support the kind of scale required for the platform vendors of the future. Whereas the barriers to entry for traditional software vendors has been moderately low, the barrier to entry for online platform vendors will be extremely high.

I highly recommend this piece for anyone interested in building scalable systems or in thinking about how services are changing the software industry. Also, I recommend reading ALL of the comments. I don’t agree with everything written there (much of it is anti-MS in terms of technology) but there are many interesting points of view to consider.