COMM

There were 255 submissions to the technical program at this year's SIGCOMM. The program committee and external reviewers together generated over 1,100 reviews comprising of 45,000 lines of text. The program committee meeting was held at USC's Information Sciences Institute in Los Angeles on 18 April, 2005. The location was very appropriate, being a hub of much of the early development of the Internet. After 9 hours of discussion, we selected the 27 papers seen here in this program, which we hope you will find stimulating and, in the long term, of some importance in the development of the field of networking.We made two changes in the way the technical program was organized this year. First, we reduced the size of the program committee to twenty in an attempt to increase the amount of informed discussion during the program committee meeting. This greatly increased the workload on the members of the program committee. We are very grateful to them for their magnificent response and to the high quality of their reviews. Please join us in thanking them for rendering this valuable service to the community.Second, for the past several years the SIGCOMM Call for Papers has included an explicit call for a category of papers called "position papers". This was an attempt to broaden the vision of what a "SIGCOMM paper" might be, to bring in papers with a vision of a future not necessarily constrained by what is possible in today's environment or with today's prejudices. This year we reverted to the prior practice of not making explicit a call for position papers. We still received submissions of papers that might be considered position papers, and we wait to see whether the resulting program meets the needs of the community in terms of a broader selection of papers.One result of the smaller size of the program committee was an increased reliance on members of the community to provide external reviews of the papers. The names of these external reviewers are listed elsewhere in the proceedings. We, and the network research community at large, owe them a great debt, and we thank them.In terms of geographic diversity, this year's SIGCOMM has not made much improvement over previous years' conferences. Although we received papers from across the globe, the bulk of the program is from the U.S. (23), with three from the U.K. and one from France. The program committee included one member from Japan, one from the U.K., and one from Germany, in addition to 17 U.S.-based members.This year, program committee member Anja Feldmann from T.U. Munich ran an experimental "shadow program committee", with members from Europe and Asia, in an attempt to give younger colleagues an introduction to the functioning of program committees. Anja has written up her experiences running this shadow PC in an article that will appear in the ACM Computer Communications Review.As a community, we impose a large load on the SIGCOMM reviewing process by the large number of papers that are submitted. Of course, the submitted papers, whether accepted or rejected, are the lifeblood of the conference. But, as some have noted in the past, we, as a community, are not as selective as other communities in terms of what papers to submit to our conference. It's unclear how to resolve this conundrum, as stifling the flow of submissions could result in a much poorer conference. However, as the field grows, we hope that potential authors will learn to gauge which papers are more appropriate to SIGCOMM.SIGCOMM has become more than just the technical conference, but focusing on the technical program has been our job. Luckily, Roch Guerin and the other members of the executive committee have freed the program chairs from having to worry about the myriad of organizational issues surrounding SIGCOMM. These issues define the SIGCOMM experience for attendees, and we thank Roch and the executive committee for their excellent work.