Brian Bershad (Univ. of Washington)
David Cheriton (Stanford University)
Frans Kaashoek (MIT)
Steven Lucco (CMU)
Larry Peterson (Univ. of Arizona)
Why are there so many major new kernel efforts underway today -- more
than at any time in recent memory? What is spawning this upsurge in
kernel development activity? Is it the information highway? ATM?
Wireless? Do RISC processors make microkernels too slow? Is OS
extensibility really required to cope with changing system
requirements? Do we really need to insert untrusted application code
into the kernel to get performance? In this panel,
five leading OS researchers will
explain what trends in computer systems make
their new kernel necessary, and why their approach is better than the
other four.
The panelists initially presented in reverse alphabetical order:
Audio: part 1 (20MB, 45 minutes),
part 2 (20.2MB, 44 minutes).