While you wait for the Indigo show to start, here is some stuff to look at and consider (again).
The links at the bottom of this post point to five slide decks that I have been using for presentations throughout this year. All of them are, indeed, very relevant to the Indigo story you will be hearing at PDC 03.
This spring, I’ve been on the road together with my good friend Steve Swartz, who is one of the Architects and Program Managers at Microsoft’s Indigo Team. On this tour, we have presented lots of ideas around scalable applications in seven cities all over Europe. And of course, we knew at the time that Indigo was coming … ;)
The “DistSys” ZIP files below contain the four decks we have been using on that tour. “Layers” is about layering, tiers and services (pay attention to “dialogs”), “Processes” is about implementation aspects such as process models, state and sessions, “Transactions” is about taking thinking about transactions beyond the database and “Scaling” highlights several essential ideas around scalability.
My DEV357 talk at TechEd Dallas, which is in part an aggregate of the talks from this tour, may even be more important, because it actually contains outspoken, concrete guidance for how to build applications on today’s technology stack in order to be ready for Indigo. To summarize the core message of that deck in terms of appropriate use of the existing technology stack for distributed systems:
· .NET Remoting: Use for “local”, on-machine, cross-app-domain communication.
(In clear words: Remoting calls don’t leave the machine!)
· Enterprise Services: Use for “near”, cross-process, cross-machine communication
· ASMX: Use for “near” or “far”, cross-process, cross-machine communication. Prefer over Enterprise Services, unless you need the features or have pressing performance problems.
Read. Understand. Absorb.