During the last month I spent a lot of time going into more depth into UMDF. I wanted to make a more in-depth post about UMDF, but I just saw the UMDF talks (audio+slides) from WinHEC 2006 have been posted at the UMDF website. Since some of the topics that I wanted to analyze are covered by the presentations (especially by the 2nd one), I’ll postpone my post for now.
The first talk (audio+slides) can be found here and the presentation slides can be found here. It covers UMDF’s business model: where it can be applied, its advantages, how it performs compared to KMDF and WDM, etc.
The seconds talk (audio+slides) can be found here and the presentation slides can be found here. This technical synopsis covers many important aspects UMDF’s architecture. You’ll hear about how the device stack of a user-mode (UMDF) driver differs from a kernel-mode driver’s (KMDF or WDM), how a UMDF driver is loaded, what the UMDF components are, what are the basic programming objects, etc. This is a very interesting talk for everybody, who wants to learn how UMDF works behind the scenes and how everything is bundled together.
Finally, I would like to suggest once more the presentation found here. It is the most detailed UMDF presentation until now (It just includes the presentation slides. There is no audio or video). I suggest that you should also take a look at the first part (pages 1-43) of this presentation, after listening to the second WinHEC talk.