Jonathan
Eisler (Johny)Department of Computer Science
University of Toronto
10 King's College Road
Toronto, Ontario, Canada M5S 3G4
My username is jeisler.
My email address is [username]@cs.toronto.edu.
Welcome
Jonathan
Eisler (Johny)
Department of Computer Science
University of Toronto
10 King's College Road
Toronto, Ontario, Canada M5S 3G4
My username is jeisler.
My email address is [username]@cs.toronto.edu.
I'm a Master's student here at the University of Toronto. My supervisor is Sam Toueg; I've also worked closely with Vassos Hadzilacos. My research focuses on fault-tolerant algorithms for solving problems in distributed computing using failure detectors.
Movies and cookies! A survival guide. Social events. Really nice people.
Holy free, licensed Microsoft software, Batman! Thanks to Victor for spotting this. The authorization form referred to on the site needs to be signed by Lloyd Smith or any other CDF staff member. MSDNAA is only available to CS, Eng and IS students; however, there's a link on the MSDNAA page to a list of all available licensed software, generously discounted.
Seminars
Computer
Science and Math WebNotice listings for the current week;
Student Seminar
in Complexity Theory;
Theory Student
Seminar, currently hosted by Mark Braverman.
SGS Office of English Language and Writing Support
Courses, workshops and appointments. They offer classes throughout the year.
Stewards
I'm a CUPE 3902 Steward for CS.
Weather, traffic & school closings
The Engineering and Computer Science Library is the first stop for CS grad students. Learn about helpful workshops at the ILU. What is the ILU? I don't know, but it has something to do with library and internet skills.
Research
Solving Consensus in Shared-Memory with $\Omega$
This is just the extended abstract, and leaves out some proofs, but the algorithms and discussion appear to be mostly intact. They solve consensus with other failure detectors as well.
Sam Toueg's Failure Detector page at Cornell
This is a great first stop for failure detector papers. Read the classic [CT96] and [CHT96] results. The first provides a discussion and numerous algorithms dealing with fault-tolerant algorithms using failure detectors; the second proves that $\Omega$ is necessary and sufficient (it is the weakest FD) to solve Consensus with a majority of correct processes.
Raynal and Mostefaoui's Leader-based Consensus
Read the abstract. This algorithm has many desirable properties: decision in one round in the stable leader phase, as well as easy randomization hybridization.
Rachid Guerraoui's publications
R. Guerraoui's 2003 publications include a number of papers on failure detectors for implementing registers, consensus and NBAC. (There is no direct link to the 2003 papers.)
These cover many of the distributed computing topics I'm interested in. Many important reports and papers are hard to track down, so this listing is very convenient. Of particular interest are On Interprocess Communication, which defines three classes of registers, Paxos, as well as numerous notes and papers on mutual exclusion. And let's not forget the Byzantine Generals problem.
Random Papers
Read-Copy Update
Tom mentioned this paper on an alternative technique to traditional locking (e.g.: Herlihy's consensus objects). A background paper is here.
Shameless
Yes, I have a domain. Not much there, but it's still good to know it exists.
Ryan's daily comic, which is most adamantly still daily. Sorry, Ryan.
vocabulary test
Try your hand at comparing 200 pairs of words. A friend of mine correctly compared 169. There is a link to a page listing the top 100 performances up to a few years ago. The cut off is around 174. A link to other (even more) insanely difficult tests of all kinds can be found at the bottom of the page.