The primary goal of Internet content delivery is to improve Website performance, scalability, and availability by serving content from a large deployment of servers at the edge of the Internet, thereby avoiding Internet performance bottlenecks such as congested peering points while providing peak crowd protection. Deployment and mapping are the most publicized aspects of content delivery, yet they are by no means the most important. At least equally crucial are many challenging systems issues, including local and wide-area load management, fault tolerance, real-time distributed data collection, security, remote administration of software and configuration, efficient allocation of distributed storage resources, and the collection and processing of terabytes of log files daily. This talk will describe these problems in more detail and explain how some are addressed in the Akamai Content Delivery Network.

Daniel Lewin founded Akamai in September 1998, together with Tom Leighton and a leading group of MIT scientists and business professionals. As Chief Technology Officer, Lewin drives Akamai's research and development strategy.

Systems as Languages
Benjamin Chelf, Andy Chou, Seth Hallem, and Dawson Engler, Stanford University

Devil: An IDL for Hardware Programming
Fabrice Mérillon, Laurent Réveillère, Charles Consel, Renaud Marlet, and Gilles Muller, IRISA/INRIA, Campus de Beaulieu

Taming the Memory Hogs: Using Compiler-Inserted Releases to Manage Physical Memory Intelligently
Angela Demke Brown and Todd C. Mowry, Carnegie Mellon University

Surplus Fair Scheduling: A Proportional-Share CPU Scheduling Algorithm for Symmetric Multiprocessors
Abhishek Chandra and Micah Adler, University of Massachusetts, Amherst; Pawan Goyal, Ensim Corporation; and Prashant Shenoy, University of Massachusetts, Amherst

Performance-Driven Processor Allocation
Julita Corbalán, Xavier Martorell, and Jesús Labarta, Universitat Politècnica de Catalunya

Policies for Dynamic Clock Scheduling
Dirk Grunwald and Philip Levis, University of Colorado; Keith I. Farkas, Compaq Western Research Laboratory; Charles B. Morrey III and Michael Neufeld, University of Colorado

Towards Higher Disk Head Utilization: Extracting "Free" Bandwidth from Busy Disk Drives
Christopher Lumb, Jiri Schindler, Gregory R. Ganger, Carnegie Mellon University; Erik Riedel, Hewlett-Packard Labs; and David F. Nagle, Carnegie Mellon University

Latency Management in Storage Systems
Rodney Van Meter, Quantum Corporation, and Minxi Gao, University of California at Berkeley

A Low-Overhead, High-Performance Unified Buffer Management Scheme That Exploits Sequential and Looping References
Jong Min Kim, Jongmoo Choi, Jesung Kim, Seoul National University; Sam H. Noh, Hong-Ik University; and Sang Lyul Min, Yookun Cho, and Chong Sang Kim, Seoul National University

How to Build a Trusted Database System on Untrusted Storage
Umesh Maheshwari, Radek Vingralek, and Bill Shapiro STAR Lab, InterTrust Technologies Corp.

End-to-End Authorization
Jon Howell and David Kotz, Dartmouth College

Design and Implementation of a Self-Securing Storage Device
John D. Strunk, Garth R. Goodson, Michael L. Scheinholtz, Craig A. N. Soules, and Gregory R. Ganger, Carnegie Mellon University

Fast and Secure Distributed Read-Only File System
Kevin Fu and M. Frans Kaashoek, Massachusetts Institute of Technology; and David Mazières, New York University

Overcast: Reliable Multicasting with an Overlay Network
John Jannotti, David K. Gifford, Kirk L. Johnson, M. Frans Kaashoek, and James W. O'Toole, Jr., Cisco Systems

System Support for Bandwidth Management and Content Adaptation in Internet Applications
David Andersen, Deepak Bansal, and Dorothy Curtis, Massachusetts Institute of Technology; Srinivasan Seshan, Carnegie Mellon University; and Hari Balakrishnan, Massachusetts Institute of Technology

Operating System Management of MEMS-based Storage Devices
John Linwood Griffin, Steven W. Schlosser, Gregory R. Ganger, and David F. Nagle, Carnegie Mellon University

Trading Capacity for Performance in a Disk Array
Xiang Yu, Benjamin Gum, Yuqun Chen, Randolph Y. Wang, and Kai Li, Princeton University; Arvind Krishnamurthy, Yale University; and Thomas E. Anderson, University of Washington

Interposed Request Routing for Scalable Network Storage
Darrell Anderson, Jeff Chase, and Amin Vahdat, Duke University

Do you have interesting work you would like to share, or a cool idea that is not yet ready to be published? Symposium attendees provide valuable discussion and feedback. Short, pithy, and fun, these Work-in-Progress Reports (WiPs) introduce interesting new or ongoing work. We are particularly interested in presentation of student work. People interested in presenting their work at this session should follow the submission instructions.

Would you like an opportunity in an informal setting to present a poster describing your current work, an aspect of your work, or work you are considering to the OSDI attendees? If so, the Poster Session is for you. People interested in presenting their work at this session should follow the submission instructions.

Proactive Recovery in a Byzantine-Fault-Tolerant System
Miguel Castro and Barbara Liskov, Massachusetts Institute of Technology

Exploring Failure Transparency and the Limits of Generic Recovery
David E. Lowell, Compaq Western Research Laboratory; and Subhachandra Chandra and Peter Chen, University of Michigan

Design and Evaluation of a Continuous Consistency Model for Replicated Services
Haifeng Yu and Amin Vahdat, Duke University

Scalable, Distributed Data Structures for Internet Service Construction
Steven D. Gribble, Eric A. Brewer, Joseph M. Hellerstein, and David Culler, University of California at Berkeley

Processes in KaffeOS: Isolation, Resource Management, and Sharing in Java
Godmar Back, Wilson H. Hsieh, and Jay Lepreau, University of Utah

CpU: Component Composition for Systems Software
Alastair Reid, Matthew Flatt, Leigh Stoller, Jay Lepreau, and Eric Eide, University of Utah