Abstracts of the 2018 ACM International Conference on Measurement and Modeling of Computer Systems

Abstract

This special issue of Performance Evaluation Review (PER) contains extended abstracts of the papers presented at the ACM SIGMETRICS Conference. SIGMETRICS is the flagship conference of the ACM special interest group for the computer systems performance evaluation community. This year marks the forty-fifth anniversary of the initial SIGMETRICS conference, which started as the "First National SIGME Symposium on Measurement and Evaluation" in 1973. The past four decades have seen enormous changes in the field of computer science, but the importance of measurement, modeling, and performance evaluation remains as critical as ever, and this year's program reflects the sustained ability of the conference to attract high quality submissions while evolving with the broad interests of the performance evaluation community. This year's conference program includes papers on topics that have been mainstays since the founding of our SIG, including load balancing, scheduling, resource allocation, and performance measurement. However, the program also highlights the emergence of a variety of new application areas that have become prominent in recent years, such as the sharing economy, cryptocurrencies, network science, and machine learning. This year marks the completion of a multi-year transition of SIGMETRICS to a so-called"jourference" model, where papers are reviewed using a journal procedure (including a one-shot revision), there are multiple submission deadlines throughout the year, and papers are accepted to appear in the ACM Proceedings on Measurement and Analysis of Computer Systems (POMACS) journal. Papers were submitted to Summer, Fall, and Winter deadlines, and the full archival versions of accepted papers appear across three consecutive issues of POMACS. The process was a resounding success and this year's program includes 54 papers, which is the largest ever to appear at a SIGMETRICS conference. All accepted papers from the three submission deadlines are presented in the conference, and extended abstracts of them appear in this issue. In total, we received 270 submissions (58 in the summer, 92 in the fall, and 120 in the winter) and accepted 54 papers across the three deadlines, 20 of which were accepted after a one-shot-revision. This gives a highly competitive acceptance rate of 20%, which is not far from the historical acceptance rate at SIGMETRICS. Note that the one-shot-revision process worked effectively to allow "exciting but flawed" papers to address concerns from the reviewers and then be accepted into the conference, and thus the acceptance rate is slightly higher than in recent years. As in prior years, we performed reviews in two rounds during each deadline period. In the first round, each paper was assigned to three reviewers. In the second round, an additional two reviewers were assigned for papers with fewer than three completed reviews, papers with divergent review opinions, and papers with fewer than two high-confidence reviews. All reviews were completed by the 65 members of our technical program committee (TPC), except for a handful of papers for which we sought an external expert opinion. Decisions during the Summer and Fall deadlines were finalized via a virtual program committee meeting, and decisions during the Winter deadline were made during a physical program committee meeting in New York. Best paper awards were selected by a small subcommittee of the TPC and will be announced at the conference.