Proceedings of the 4th ACM European conference on Computer systems

It is our great pleasure to welcome you to the 4th ACM European Conference on Computer Systems -- EuroSys'09. Previous years have established a reputation for high quality, a broad and enthusiastic attendance, and a wide ranging remit. We're delighted that this year is no exception, with an excellent programme that is being presented in a historic city. The topics cover a wide variety of areas, ranging from storage systems to trusted computing; from scheduling to fault tolerance; and from modeling to measurement. In addition, the programme includes a poster session to encourage detailed technical interactions as well as a work in progress session where researchers can present initial results and directions. This year's call for papers attracted 224 expressions of interest, and 149 of these turned into submissions. They came from all over the world: Asia, Australia, Europe, North and South America, and the Middle East. Two papers were rejected for failing to meet the formatting guidelines, so we had 147 papers to choose from: 21 short papers, 126 long ones. The programme committee (PC) used a two-round reviewing process that winnowed the full set down to 64 papers for the second round, and then selected 25 papers (2 short, 23 long) for the final programme. All papers received at least 3 reviews in the first round; all but two second-round papers received 6 or 7 reviews. A total of 704 reviews were written, containing 320k words of content. All this in less than 2 months! The majority of the reviews were written by the PC members themselves; where we received help from external reviewers, the PC members were still responsible for providing the numeric scores. Our scoring system was designed to elicit more information than the usual 4-point up/down scale: each programme committee member was encouraged to use a scale from 0 to 100, and their scores were normalized so that their sum added up to 100N, where N was the number of papers they reviewed. A paper's overall score was the mean of the scores it received, and correlated well with the final outcome.