These
figures show, in order, (1) a peer to peer system (Gnutella), (2) a sensor
network, (3) the Internet ISP topology, and (4) a food web in a biological
ecosystem.
Instructor: Dr. Indranil Gupta, indy at cs dawt uiuc dawt edu, 2112 SC
Class meets Tu, Th 11.00 am - 12.15 pm in 1214 SC
Office hours are Tu, Th 3.30 pm - 5 pm (class days only) in 2112 SC
Class Newsgroup is class.cs598ig. Instructions on accessing newsgroups.
This is a course on distributed systems. It brings together research centered in theory, on peer-to-peer systems and on sensor networks. The course reviews classical work in these three scattered areas of distributed computing research, and explores overlaps across them.
One of the themes for Fall 2004 is design methodologies for distributed protocols. We will ask questions like: What do protocols in peer to peer systems, protocols in sensor networks, and the dynamics of natural systems such as biological ecosystems, all have in common? Can we design protocols rigorously from natural phenomena (including epidemics and biological ecosystems)? Can we generate protocol code automatically?
Best 5 Final Project Papers in class (in no particular order)
James Newell, "Increasing the reliability of DHTs using MultiRouting".
Vartika Bhandari, "PriorityCast: Gossip-based dissemination of prioritized information in wireless ad-hoc networks".
Yookyung Jo, "Distributed computing algorithms and protocols inspired by the collective behavior of honeybees".
Pradeep Kyasanur and Romit Roy Choudhury, "Adaptive Gossiping in Sensor Networks".
Boris Capitanu and Ellick Chan, ["An Operating System Framework for running Regular Applications on Mote Networks"].
Other Notable Mentions among Final Project Papers (in no particular order)
Charles Yang, "Characteristics of flash crowds and a comparison of contemporary solutions".
William Conner, "Neighbor Filtering: preventing application-layer DOS attacks in structured P2P systems".
Jin Liang, "On-demand overlay network for distributed system management".
Jungmin So and Jintae Kim, "An adaptive protocol for reliable aggregation in wireless sensor networks".
If you would like to see copies of these papers, please drop by my office.
12/24: This course has been completed. For all questions about grades, projects, reviews etc., please email Indy after January 5, 2005. Happy holidays.
[Average, Standard Deviation, Lowest, Highest] in Class for Project (survey+midterm+final, totalled out of 50) is [38.47, 4.68, 30.5, 44.5].
12/24: All your paper reviews (with comments) were handed back in class.
12/19: All projects have been graded, and you should have received email (dated 12/19) with review comments on your final project report.
Final grades have been assigned. If you would like to ask/chat/discuss about the project or course grading, please drop by my office by the end of Tuesday December 21, 2004.
12/13: The final project report is due December 17, 11.59 PM. By this time, please submit (i) softcopy of your final report to indy at ad dawt uiuc dot edu (Word or pdf or ps); and (ii) turn in a hardcopy of your project report (only) into 2112 SC (slip it under the door if closed). Previous survey and midterm reports are not required. Your final project report must be limited to 12 pages with at least 12 pt font.
12/7: Please attend the last lecture on Thursday, Dec 9th. The lecture will bring closure to some of the questions and discussion points we had opened right at the beginning of the class. It is a "closing parenthesis", if you will.
11/30: The final project report is due December 17, 11.59 PM.
11/30: Midterm reports have been graded and returned. Grading criteria:
The final project report will be "graded" from the viewpoint of a conference reviewer/PC. Important criteria include: Importance of problem, Novelty of solution, Evaluation of solution, Clarity of Presentation, Nits (grammar, references, etc.).
The highest total (survey + midterm) so far is 20.5 / 25.0.
11/18: At the end of the semester, three best papers will be selected from among all the final project reports. Information about these best papers will be put up on the course website. In addition, the instructor (Indy) will continue working with the student authors on improving these best papers, with the aim of submitting them to reputed conferences in early Spring.
As an indiciation of how things usually turn out, all best papers from Fall 2003's version of this course have either been accepted at conferences or are currently under review.
11/16: Midterm deadline extended to Wednesday, Nov 24, 11.59 PM. By this time, please email a softcopy of your midterm report to indy at ad dot uiuc dawt edu with the subject line "598IG Midterm Report". Also, please slip a hardcopy of your graded review paper under the door of Indy's office 2112 SC anytime before this deadline. A hardcopy of the midterm report is not required.
10/26: Survey papers were handed back today (with grades and comments). Your midterm report (due Nov 19) should contain experimental data (requires an implementation of course!), and a compressed version of the related work section from this survey. Preserve the hardcopy handed back to you today and resubmit it along with the midterm report.
10/22: Project Survey Report due today, 11.59 pm [12pt font, single-sided, 3 pages max]. By this time, please (i) email a softcopy to indy at ad dot uiuc dawt edu with the subject line "598IG Survey Report" AND (ii) slip a hardcopy under the door of Indy's office 2112 SC (the hardcopy is REQUIRED if you want your survey report to be graded).
10/19: Unless mentioned otherwise, you are expected to write reviews to only TWO of the papers in each session from now onwards.
10/19: Unless there are extenuating circumstances (e.g., medical), NO extensions will be given for the survey report deadline.
10/15: For the classes of 10/19 and 10/21 ONLY, you are required to read and write reviews for only ONE of the three papers in each of these sessions. This relaxation is on account of the project survey deadline.
10/8: Survey papers are due on Oct 22 at 11.59 PM. By this time, please email a softcopy to indy at ad dot uiuc dawt edu with the subject line "598IG Survey Report" AND slip a hardcopy under the door of Indy's office 2112 SC (the hardcopy is REQUIRED if you want your survey report to be graded).
Survey report is expected to be at most 3 pages, 12pt font, single sided (two column ok). A possible split of the 3 pages is as follows (feel free to deviate). 0.5 pages to references (typical number is 15-20; select judiciously). 2 pages to the survey/description of the references (write cleverly; don't just summarize the referenced papers). 0.5 pages to your project statement and an approximate plan of activity. Remember that the midterm report is due in the middle of November, and a final project is due at the end of the semester.
All these deadlines appear on the Course Schedule webpage. The survey report counts for approximately 10% of the course grade.
9/13: Note: Please email Indy first for an appointment to show your presentation (the earlier the better, but not later than 5 pm the day before your presentation)! Faculty schedules are usually busy and you may find him working and unable to chat if you just decide to walk in to his office!
Email your reviews to indy at ad dawt uiuc dawt edu mail NOT to indy at cs dawt uiuc dawt edu.
9/12: Final Presentation schedule is up. Please double-check the date of your presentation (some dates have changed).
9/2: Project tips, and links to Fall 2003 Scattered Systems techreports that became conference papers: here.
9/2: Please sign up for a presentation slot by September 10 (Friday). The only slots available now are 9/16-11/16. Some of these slots are already taken (see "Papers Presentations and Reviews" link above), so hurry!
8/27: Please fill out the "Student Survey" sheet (above link) and return to me during office hours or next Tuesday's class. Please read "Gnutella" paper for Tuesday's class.
Class newsgroup is up class.cs598ig - please use this for discussions. Instructions on accessing newsgroups.
8/26: First lecture.
8/9: Course web page online.
Policies on Ethics, Attribution and Cheating
Reviews: You may discuss papers with others in the class, but reviews should be written independently. In other words, if two reviews are found to have been copied from each other, both students will be punished. The best way to ensure that you are safe is by mentioning right at the top of your review that "This review was written after discussion with Mr. John Doe and Ms. Jane Joe".
Presentation: Your work (presentations and reviews) should be original and independent, although you are allowed to cooperate with your partner for your session. Reusing slides from someone else's presentation should be done only if it is absolutely necessary, but should be acknowledged right at the beginning of the presentation.
The standard university policies on original work, cheating and attribution apply to all work in the course. Violation of these may result in either lowering of course grade by one letter, or failing the course, or a different final decision left to the instructor.