Linux research topic ideas

[Cri]

Greetings all,

This semester in my graduate operating systems class I (and 2 others) have
to do a research paper/presentation. I would like to do something with Linux,
but I'm not exactly sure what. Any ideas? The prof. is pretty much open to
anything, but it has to be fairly technical since it is an OS class. I've thought of a few topics, but if you have any other thoughts, I would love to
hear them.

Here are my initial brainstorms...rough, but a start:

-Some sort of simple experiment that benchmarks Linux and other OSs

-Research on the kernel

-Something with C...

-??


Thanks to all who reply.

Chris

[Gaccm]

are there any requirements besides just being "technical"? If you really want to dig deep, you can compare 2.4.9- with Rik's memory manger and 2.4.10+ with Andrea's memory manager.

[camelrider]

I agree with Gaccm. This was the first subject that jumped to mind. You can get into an interesting aspect of the OS and maybe contribute some useful info to the project in the process!

[Cri]

Really, there are no requirements. However, my prof is being picky. I emailed him 5 ideas and he shot them all down, except 1, but he said I'd need access to proprietary UNIX code to do it. Obviously, that is not happening...

I am interested in this memory topic. Thanks for the idea. Excuse me for not following the kernel mailing list, but what is the situation you are describing? One memory management scheme was used in kernels 2.4.9- and a new one in 2.4.10+?
How do you think we could go about testing these algorithms/schemes? Do you think that enough of a behavior analysis could be done to get a 15 page paper out of it (probably since memory management affects everything :)? What would you like to see researched and what would be useful to the community? It would be nice to write something useful for outside the classroom...

Thanks again to all who reply.

[ 23 January 2002: Message edited by: beatMiner ]