SOSP

Welcome to the Proceedings of the 23rd ACM Symposium on Operating Systems Principles (SOSP 2011), held in Cascais, Portugal. This year's program includes 28 papers, and touches on a wide range of computer systems topics, from data center computing, storage systems and geo-replication to operating system architecture, virtualization, concurrency, security, and mobile platforms. The program committee made every effort to identify and include some of the most creative and thought-provoking ideas in computer systems today. Each accepted paper was shepherded by a program committee member to make sure the papers are as readable and complete as possible. We hope you will enjoy the program as much as we did in selecting it.

This year's proceedings are, for the first time, published digitally on a USB key with no paper copy distributed at the conference. The cost to the environment of so many reams of printed paper, plus the difficulty of shipping printed material to the conference site, made this an easy decision. The workshop proceedings appear on the conference USB key as well. You will find two copies of each of this year's SOSP papers: a traditional 2- column version designed for printing, and a one-column version intended for reading on a computer. In addition, the USB key contains a copy of every SOSP paper from each of the previous 22 instances of the conference, starting in 1967. The very nature of publishing is changing as we speak. We look forward to your feedback about the appropriate form and format for future SOSP proceedings.

We are most grateful to the authors of the 157 papers who chose to submit their work to SOSP (five papers were subsequently withdrawn by the authors). Their ideas and efforts are the basis of the conference's success. Selecting the program out of so many quality submissions was a difficult task. A program committee consisting of 28 leading scholars in the broad area of computer systems conducted a three-stage reviewing process and online discussion; final selections were made during a physical meeting in Boston, MA, which was attended by a core of 13 PC members. Each submission received at least three PC reviews, with a maximum of eight. All in all, 719 reviews were produced. The PC made every effort to provide detailed and constructive feedback, which should help authors to further improve their work. Author anonymity was maintained throughout the reviewing and selection process; PC members were removed from the deliberations of any paper with which they had a conflict of interest (co-author, same institution, recent collaborator, former student/adviser).