On April 1, 2023 we stopped monitoring the traffic on our campus Internet link, nearly 20 years t... more On April 1, 2023 we stopped monitoring the traffic on our campus Internet link, nearly 20 years to the day since we first started doing so. During these two decades, we faced a vast array of issues that affected the collection, storage, analysis and backup of our monitoring data. In this paper we share some of our experiences, so that future networking researchers have an opportunity to learn from our successes as well as our many mistakes and misfortunes.
This paper discusses the design and evaluation of CATNIP, a Context-Aware Transport/Network Inter... more This paper discusses the design and evaluation of CATNIP, a Context-Aware Transport/Network Internet Protocol for the Web. This integrated protocol uses application-layer knowledge (i.e., Web document size) to provide explicit context information to the TCP and IP protocols. While this approach violates the traditional layered Internet protocol architecture, it enables informed decision-making, both at network endpoints and at network routers, regarding flow control, congestion control, and packet discard decisions.The ns-2 network simulator is used to evaluate the performance of the context-aware TCP/IP approach, using a simple network topology and a synthetic Web workload. Simulation results indicate a 10-20% reduction in TCP packet loss using simple endpoint control mechanisms. More importantly, using CATNIP context information at IP routers can produce 20-80% reductions in the mean Web page retrieval times, and 60-90% reductions in the standard deviation of retrieval times.
High Performance Distributed Computing, Jun 10, 2009
It is our great pleasure to welcome you to the first Workshop on Large-scale System and Applicati... more It is our great pleasure to welcome you to the first Workshop on Large-scale System and Application Performance -- LSAP2009. We have initiated this workshop as we think it addresses a very important and timely subject. Over the last decade, computer systems and applications in everyday use have grown to unprecedented scales. Large clusters serving millions of search requests per day, grids executing large workflows and parameter sweeps consisting of thousands of jobs, and supercomputers running complex e-science applications, have now hundreds of thousands of processing cores. In addition, clouds are quickly emerging as a large-scale computing infrastructure. Peer-to-peer systems and centralized video distribution systems that dominate the internet and complicated internet applications such as massive multiplayer online games are used by millions of people every day. In view of this tremendous growth, understanding the performance of large-scale computer systems and applications has become vital to institutional, commercial, and private interests. This workshop intends to be a venue for papers on performance evaluation methods, tools, and studies focusing on the challenges of large scale, such as decentralization, predictable performance, reliability, and scalability. It aims to bring together system designers and researchers involved with the modeling and performance evaluation of large-scale systems and applications. The call for papers of the workshop attracted 7 submissions, out of which the program committee accepted 4 papers that cover a variety of topics, ranging from modular data centers and grid meta-brokers to the MapReduce programming model and peer-to-peer systems. In addition, the workshop program features a keynote presentation by Douglas Thain of the University of Notre Dame, USA, on scaling up data-intensive scientific applications.
... 4In the ensuing discussion, we will generalize the counting formula to two dimensions, using ... more ... 4In the ensuing discussion, we will generalize the counting formula to two dimensions, using the form C(i; j; r; c). For absolute clarity, one should use C1(i; N) and C2(i; j; r; c) to distinguish the one-dimensional case from the two-dimensional case. ...
Measurement and Modeling of Computer Systems, Jun 6, 2005
We are pleased to present the program of the SIGMETRICS 2005 conference. The program covers a wid... more We are pleased to present the program of the SIGMETRICS 2005 conference. The program covers a wide spectrum of performance-related topics, including peer-to-peer networks, traffic measurement and classification, bandwidth sharing and scheduling policies, network performance measurements, caching and file systems, traffic estimation and topology inference, wireless networks, network and server performance evaluation. Overall, the program features a broad mix of methodological contributions and modeling studies, balanced with a diverse range of applications of performance evaluation models and methods. The paper selection process was highly competitive. A total of 237 submissions were received, of which 31 papers were accepted for presentation at the conference, while an additional 20 papers were selected for poster presentation. The papers were selected after a rigorous review process based on their originality, technical quality and relevance to the conference. For all papers, at least three independent reviews were sought from the Program Committee, whose ranks were filled with 57 established experts representing 12 countries and a variety of backgrounds in academia, industry and government institutions. For many papers, additional reviews were obtained from external specialists, resulting in a total of over 800 reviews. The final paper selection took place at the Program Committee meeting which was held at Columbia University on January 21 and 22, 2005. Throughout the entire process, every effort was made to avoid any (potential) conflict-of-interest situations, in particular during the discussions at the Program Committee meeting. A special subcommittee decided on the following two awards: "Coupon Replication Systems" by Laurent Massoulie, Milan Vojnovic (Microsoft Research) as the conference's best paper, and "A Network Service Curve Approach for the Stochastic Analysis of Networks" by Florin Ciucu, Almut Burchard, Jorg Liebeherr (University of Virginia) as the best student paper. Jussara Almeida and Thomas Bonald did an excellent job publicizing the conference. We thank Martin Arlitt, Anirban Mahanti, and Camille Sinanan for their highly capable work on finance, proceedings, and registration/local arrangements, respectively. Leana Golubchik, SIGMETRICS Chair, gave valuable advice and support at crucial moments. We received able and professional assistance from ACM Headquarters staff members, including Caryn Chan, Irene Frawley, Adrienne Griscti, Maritza Nichols, and Jessica Wilmers. Lisa Tolles, the Proceedings Coordinator at Sheridan Printing, provided dedicated and careful work on the proceedings production. Sue Williamson designed conference promotional material, with photos courtesy of the Banff/Lake Louise Tourism Bureau. The Banff Park Lodge personnel, especially Valerie Hunter-Prenger and Ann Haagaard, assisted with conference planning, and our student volunteers helped make it all happen. The main credit for the program obviously goes to the authors for contributing their finest work, as well as the tutorial speakers and the workshop organizers for their excellent contributions. We also wish to express our deep gratitude to Urs Hoelzle from Google for agreeing to give the keynote address. The Program Committee members and external reviewers made further vital contributions, lending their expertise, spending many diligent hours reviewing papers, and braving a January blizzard to attend the Program Committee meeting in New York City. We sincerely thank them for their tremendous dedication and efforts in ensuring a high-quality program. We particularly owe the members of the awards committee Thomas Bonald, Ed Coffman (chair), Arif Merchant, Philippe Nain, and Y.C. Tay for their selfless exertions in performing such a demanding task. We further thank the General Co-Chairs Derek Eager and Carey Williamson for their constant guidance, support and confidence. We are grateful to Leana Golubchik (Chair of SIGMETRICS) and several previous Program Chairs for sharing their experiences and insights. We are indebted to Sophie Majewski, Vishal Misra, Erich Nahum, and Dan Rubenstein for making the logistic arrangements for the Program Committee meeting. We applaud the Tutorial Co-Chairs Kimberly Keeton and Vishal Misra for setting up an excellent tutorial program, the Proceedings Chair Anirban Mahanti for assembling the proceedings, and the Publicity Co-Chairs Thomas Bonald and Jussara Almeida for their great job in publicizing the conference. Special thanks to Matthew Andrews, Bruce Shepherd, and Phil Whiting for contributing code to the paper assignment algorithm. Finally, big thanks to H.M. Law who served as webmaster for the conference website and administered the submission management system. Without his tireless efforts, it would have been impossible to handle hundreds of papers and reviews.
On April 1, 2023 we stopped monitoring the traffic on our campus Internet link, nearly 20 years t... more On April 1, 2023 we stopped monitoring the traffic on our campus Internet link, nearly 20 years to the day since we first started doing so. During these two decades, we faced a vast array of issues that affected the collection, storage, analysis and backup of our monitoring data. In this paper we share some of our experiences, so that future networking researchers have an opportunity to learn from our successes as well as our many mistakes and misfortunes.
This paper discusses the design and evaluation of CATNIP, a Context-Aware Transport/Network Inter... more This paper discusses the design and evaluation of CATNIP, a Context-Aware Transport/Network Internet Protocol for the Web. This integrated protocol uses application-layer knowledge (i.e., Web document size) to provide explicit context information to the TCP and IP protocols. While this approach violates the traditional layered Internet protocol architecture, it enables informed decision-making, both at network endpoints and at network routers, regarding flow control, congestion control, and packet discard decisions.The ns-2 network simulator is used to evaluate the performance of the context-aware TCP/IP approach, using a simple network topology and a synthetic Web workload. Simulation results indicate a 10-20% reduction in TCP packet loss using simple endpoint control mechanisms. More importantly, using CATNIP context information at IP routers can produce 20-80% reductions in the mean Web page retrieval times, and 60-90% reductions in the standard deviation of retrieval times.
High Performance Distributed Computing, Jun 10, 2009
It is our great pleasure to welcome you to the first Workshop on Large-scale System and Applicati... more It is our great pleasure to welcome you to the first Workshop on Large-scale System and Application Performance -- LSAP2009. We have initiated this workshop as we think it addresses a very important and timely subject. Over the last decade, computer systems and applications in everyday use have grown to unprecedented scales. Large clusters serving millions of search requests per day, grids executing large workflows and parameter sweeps consisting of thousands of jobs, and supercomputers running complex e-science applications, have now hundreds of thousands of processing cores. In addition, clouds are quickly emerging as a large-scale computing infrastructure. Peer-to-peer systems and centralized video distribution systems that dominate the internet and complicated internet applications such as massive multiplayer online games are used by millions of people every day. In view of this tremendous growth, understanding the performance of large-scale computer systems and applications has become vital to institutional, commercial, and private interests. This workshop intends to be a venue for papers on performance evaluation methods, tools, and studies focusing on the challenges of large scale, such as decentralization, predictable performance, reliability, and scalability. It aims to bring together system designers and researchers involved with the modeling and performance evaluation of large-scale systems and applications. The call for papers of the workshop attracted 7 submissions, out of which the program committee accepted 4 papers that cover a variety of topics, ranging from modular data centers and grid meta-brokers to the MapReduce programming model and peer-to-peer systems. In addition, the workshop program features a keynote presentation by Douglas Thain of the University of Notre Dame, USA, on scaling up data-intensive scientific applications.
... 4In the ensuing discussion, we will generalize the counting formula to two dimensions, using ... more ... 4In the ensuing discussion, we will generalize the counting formula to two dimensions, using the form C(i; j; r; c). For absolute clarity, one should use C1(i; N) and C2(i; j; r; c) to distinguish the one-dimensional case from the two-dimensional case. ...
Measurement and Modeling of Computer Systems, Jun 6, 2005
We are pleased to present the program of the SIGMETRICS 2005 conference. The program covers a wid... more We are pleased to present the program of the SIGMETRICS 2005 conference. The program covers a wide spectrum of performance-related topics, including peer-to-peer networks, traffic measurement and classification, bandwidth sharing and scheduling policies, network performance measurements, caching and file systems, traffic estimation and topology inference, wireless networks, network and server performance evaluation. Overall, the program features a broad mix of methodological contributions and modeling studies, balanced with a diverse range of applications of performance evaluation models and methods. The paper selection process was highly competitive. A total of 237 submissions were received, of which 31 papers were accepted for presentation at the conference, while an additional 20 papers were selected for poster presentation. The papers were selected after a rigorous review process based on their originality, technical quality and relevance to the conference. For all papers, at least three independent reviews were sought from the Program Committee, whose ranks were filled with 57 established experts representing 12 countries and a variety of backgrounds in academia, industry and government institutions. For many papers, additional reviews were obtained from external specialists, resulting in a total of over 800 reviews. The final paper selection took place at the Program Committee meeting which was held at Columbia University on January 21 and 22, 2005. Throughout the entire process, every effort was made to avoid any (potential) conflict-of-interest situations, in particular during the discussions at the Program Committee meeting. A special subcommittee decided on the following two awards: "Coupon Replication Systems" by Laurent Massoulie, Milan Vojnovic (Microsoft Research) as the conference's best paper, and "A Network Service Curve Approach for the Stochastic Analysis of Networks" by Florin Ciucu, Almut Burchard, Jorg Liebeherr (University of Virginia) as the best student paper. Jussara Almeida and Thomas Bonald did an excellent job publicizing the conference. We thank Martin Arlitt, Anirban Mahanti, and Camille Sinanan for their highly capable work on finance, proceedings, and registration/local arrangements, respectively. Leana Golubchik, SIGMETRICS Chair, gave valuable advice and support at crucial moments. We received able and professional assistance from ACM Headquarters staff members, including Caryn Chan, Irene Frawley, Adrienne Griscti, Maritza Nichols, and Jessica Wilmers. Lisa Tolles, the Proceedings Coordinator at Sheridan Printing, provided dedicated and careful work on the proceedings production. Sue Williamson designed conference promotional material, with photos courtesy of the Banff/Lake Louise Tourism Bureau. The Banff Park Lodge personnel, especially Valerie Hunter-Prenger and Ann Haagaard, assisted with conference planning, and our student volunteers helped make it all happen. The main credit for the program obviously goes to the authors for contributing their finest work, as well as the tutorial speakers and the workshop organizers for their excellent contributions. We also wish to express our deep gratitude to Urs Hoelzle from Google for agreeing to give the keynote address. The Program Committee members and external reviewers made further vital contributions, lending their expertise, spending many diligent hours reviewing papers, and braving a January blizzard to attend the Program Committee meeting in New York City. We sincerely thank them for their tremendous dedication and efforts in ensuring a high-quality program. We particularly owe the members of the awards committee Thomas Bonald, Ed Coffman (chair), Arif Merchant, Philippe Nain, and Y.C. Tay for their selfless exertions in performing such a demanding task. We further thank the General Co-Chairs Derek Eager and Carey Williamson for their constant guidance, support and confidence. We are grateful to Leana Golubchik (Chair of SIGMETRICS) and several previous Program Chairs for sharing their experiences and insights. We are indebted to Sophie Majewski, Vishal Misra, Erich Nahum, and Dan Rubenstein for making the logistic arrangements for the Program Committee meeting. We applaud the Tutorial Co-Chairs Kimberly Keeton and Vishal Misra for setting up an excellent tutorial program, the Proceedings Chair Anirban Mahanti for assembling the proceedings, and the Publicity Co-Chairs Thomas Bonald and Jussara Almeida for their great job in publicizing the conference. Special thanks to Matthew Andrews, Bruce Shepherd, and Phil Whiting for contributing code to the paper assignment algorithm. Finally, big thanks to H.M. Law who served as webmaster for the conference website and administered the submission management system. Without his tireless efforts, it would have been impossible to handle hundreds of papers and reviews.
Uploads
Papers by Carey Williamson