![]() Princeton University
|
Computer Science 518
|
|
Graduate Coordinator: Melissa Lawson
Discussions: | 20% |
Presentation: | 20% |
Critiques: | 20% |
Project: | 40% |
September 2004
S M Tu W Th F
S
1 2 3
4
5 6 7 8 9 10 11
12 13 14 15 16
17 18
19 20 21 22 23
24 25
26 27 28 29 30
09 - | Organizational meeting |
14 - | Vivek Pai Flash & Harvest |
16 - | pres #1 - discussing CDN and CHash |
21 - | pres #2 - discussing L4 and exokernel |
23 - | pres #3 - Byzantine and RAID |
28 - | (bumped) PC meeting |
30 - |
(bumped) PlanetLab meeting |
05 - | pres #4 - Aquinas presents ESX and DISCO |
07 - | pres #5 - Lottery and Clock |
12 - | pres #6 - Seshadri presents Metal |
14 - | pres #7 - Soft-updates and LFS |
19 - | pres #8 - SEDA and Scheduler Activations |
21 - | pres #9 - Dtrace and Debox |
26 - | Fall recess |
28 - | Fall recess |
November 2004
S M Tu W Th F S
1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25
26 27
28 29 30
02 - | project proposals |
04 - | pres #10 - XFS & FFS |
09 - | pres #11 - kqueue & select |
11 - | pres #12 - DSM & RPC, Haakon Larsen |
16 - | (bumped) HotNets workshop |
18 - | pres #13 - LARD & Network Dispatcher, Seshadri |
23 - | pres #14 - GMS & LOCUS, Aquinas |
25 - | Thanksgiving |
30 - | pres #15 - Resource & Scout, Amit Agarwal |
December 2004
S M Tu W Th F S
1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30 31
02 - | pres #16 - Xen and Denali, Mike |
07 - | (bumped) OSDI conference |
10 - | pres #17 - SPIN & VINO, Aquinas |
04 - | Project presentations |
06 - | Project presentations |
07 - |
Project presentations |
11 - | Dean's Date |
Byzantine & RAID - making systems fault tolerant
CFS & NFS - distributed filesystems, LAN and WAN
CDN & Chash - content distribution networks and consistent hashing
DSM & RPC - distributed shared memory and remote procedure calls
Energy & DVS - energy efficiency and dynamic voltage scaling
ESX & DISCO - virtual machines
Flash & Harvest - Fast servers - web and proxy
IO-Lite & fbufs - buffering systems inside the OS
kqueue & select - fast event notification systems
L4 & Exokernel - stripped down kernels
LARD & Network Dispatcher - cluster request distribution
Metal & Synthesis - error detection, dynamic code generation
SEDA & Scheduler Activations - managing parallelism
Soft-Updates & LFS - file system restructuring
UVM & Mach - virtual memory restructuring
XFS & FFS - advanced file systems
Worldwide timing - clock skew becomes an issue when running distributed applications. Design a system to provide a more coherent global clock.
Scalability - test some aspect of OS scalability - threads, zero copy, virtual memory design, etc. Fix it if possible.
Interface - the select system call does not like disk file descriptors. Make it useful by having it begin prefetching when presented with a disk fd.
Anomaly detection - by observing the behaviors of large numbers of somewhat similar programs, can you determine when a program has gone out of control?
geolocation - can you tell where someone is coming from, and redirect them to the closest instance of a service?
Page faults - can you lay out code or data to reduce fault rates? Does it matter?
Virtualization costs - can you reduce the overhead of virtualization by making changes at the program level?
Do class-based resource management - see Marc Fiuczynski's project 3 for a more complete description.