This proposed project is to support the third annual programming contest to be held at SIGMOD 2011 conference in Greece. The funding is to sponsor travel for the contest participants to the conference and staffing for running the contest at MIT. The contest is open to graduate and undergraduate teams from degree-granting institutions. This year's contest task is to build a system that stores (key, value) pairs sorted in key order, durably on a Flash disk. As with previous years, the goals of the contest are to: (a) stimulate student interest in the DBMS field, (b) present a research challenge, (c) foster the creation of sharable open source code modules, (d) foster Computer Science education by creating a forum where students can excel, and (e) present a forum where excellent students can get noticed without writing papers. This contest will stimulate the database community to develop robust open source implementations of challenging research software. Such implementations have intellectual merit in that they provide a base upon which other researchers can build tools and students can learn how others have architected large data management systems. The contest will attract young, eager, students with enthusiasm for building real systems to the database community. It will emphasize that the community values real implementations in addition to academic papers, and will potentially lead to open source software that will be used by researchers outside of the community. To promote participation from under-represented groups, participants from such groups will be especially encouraged.
2011 SIGMOD Programming Contest Outcomes Report In 2011 we ran a third annual programming contest at SIGMOD 2011 in Greece. NSF funding was used to support travel for students who were finalists in the contest to travel to the conference. The contest was be open to graduate and undergraduate teams from degree-granting institutions. This year’s task was to build a specialized data structure called an index for providing very fast access to data stored on a Flash disks, which are rapidly becoming the stanard for efficient durable storage on modern computer systems. The overall goals of the contest were to: a) Stimulate student interest in computer science generally and database research specifically b) Present a research challenge c) Foster the creation of sharable open source code modules d) Foster Computer Science education by creating a forum where students can excel. e) Present a forum where excellent students can get noticed without writing papers. The contest was conduted in Spring 2011, with a bakeoff among the finalists at the conference. A $5,000 prize was donated by Microsoft, and awarded to the winning team, from TU Dresden in Germany. Finalists also included teams from University of Minnesota, Stony Brook University, Saudi Arabia, and South Korea. For more information about the contests, the finalists, and the winning submission (as well as links to soruce code), see the contest website, http://dbweb.enst.fr/events/sigmod11contest/. The primary intellectual merit of the contest was to stimulate the database community to develop robust open source implementations of challenging research software. Such implementations have intellectual merit in that they provide a base upon which other researchers can build tools and students can learn how others have architected large data management systems. The broader impact of the conference is that it attracts young, eager, students with enthusiasm for building real systems to the database community. It emphasizes that the community values real implementations in addition to academic papers, and has lead to open source software that can be used by researchers outside of the community.
