Possible projects for CS330 class

advertisement

Possible projects for CS330 class

This is a suggested list of topics only. Your actual topic may be quite different. Check with the instructor before finalizing on your topic of interest.

A. Operating Systems topics

1.

Explore alternative CPU scheduling schemes not discussed in the class. They must be variants of Multiple-level feedback queues with preemptive round robin scheduling with parameterized time quanta, for example. Investigate plausible algorithms using simulations.

2.

Explore multiprocessor scheduling with a view to minimizing wait time, given that (a) the jobs are independent, or (b) the jobs are related. Use simulations to explore various options.

3.

Explore schemes to optimize medium term scheduler (job swapper). When, and how often, should this be invoked? Is it possible to make it adaptive to load densities? How?

4.

Explore an alternative way to design a memory manager. Suppose it were to queue non-trivial tasks in its memory queue but allow the trivial ones to join the

CPU queue directly, each of which would be given sufficient memory to work with. Will that increase throughput? Is it possible to make it more adaptive to system demands? Simulate such environments.

5.

Read and survey the topic on "Threads: the lightweight processes. Find out all you can from the journals. How is it implemented in the SunOS? How do they function in OS/2?

6.

Same thing as above but now with the objects What about object oriented systems?

7.

Explore relative merits of various real storage placement strategies. Your program should measure storage utilization, and average job residence time for the various options. Assume a real memory with 1 MByte capacity. New jobs arrive having job sizes between 50K and 300K in multiple of 10K, and the durations range from

5 to 60 minutes. Simulate the environment.

8.

Do the same (with less options) but now with the garbage collection scheme. Vary the frequency of garbage collection and decide when to carry this out. Does it improve your average storage utilization?

9.

Do the problem (7) (i.e. the memory placement problem) but now with storage swapping. Explore improvements.

10.

Simulate the operation of a pure paging system. Assume a very small page size and a small number of page frames per process in real storage. Explore some of the page replacement strategies. How do they compare, say, with the optimum replacement scheme?

11.

Caching and Pre-fetching.

12.

Simulate a specific page replacement scheme but now with the following options.

In option one, replacement must be strictly local. In option two, replacement could be global. In option three, you are going to allow some sort of mixed domain strategy, say, it is global but only among a specific class the process belongs to.

13.

Simulate a paging environment where page replacement is predicated by some kind of page fault frequency scheme. Explore this option against LRU scheme.

14.

Simulate a disk operation where the objective is to optimize the seek strategy.

Explore SCAN, C-SCAN, N-step scans. How do the throughput, mean response time per job and variance of response times appear in these schemes?

15.

File System. Short-term storage of all versions of files and long-term storage of selected versions of files. Using adaptive heuristics to decide in between cases.

See Elephant: The file system that never forgets.

16.

The Google File System. Anything similar to it? Look into its page-ranking policy. http://www.cs.rochester.edu/sosp2003/papers/p125-ghemawat.pdf

17.

Decentralized user interaction in global file systems.

B. Networking topics

1.

VOIP system. How does it work? Problems and challenges? Opportunities?

Security problem?

2.

VPN system. How should it work? Problems and challenges? Opportunities?

Security issues?

3.

Peer-to-peer system. How should it work? Where and when is it vulnerable? How do you reduce its vulnerability?

4.

Coopnetworking : Expanding client-server computing by augmenting cooperation among peer processes. See the following: The Case for Cooperative Networking

V. N. Padmanabhan and K. Sripanidkulchai, Proceedings of the First International Workshop on

Peer-to-Peer Systems (IPTPS), Cambridge, MA, USA. Also,

( http://research.microsoft.com/projects/coopnet/papers/iptps02-with-addendum.pdf

5.

Distributed streaming media content using Coopnet: Look at http://research.microsoft.com/projects/coopnet/papers/msr-tr-2002-37.pdf

http://research.microsoft.com/projects/coopnet/papers/nossdav2002.pdf

6.

Resilient Peer-to-Peer Streaming V. N. Padmanabhan, H. J. Wang, and P. A. Chou

,

Microsoft Research Technical Report MSR-TR-2003-11, March 2003 http://research.microsoft.com/projects/coopnet/papers/msr-tr-2003-11.pdf

7.

Denial-of-service intrusion in Routers. How do we guard against such problems?

8.

Intrusion-detection systems in a network of routers.

Download