DATE: Thursday , November 15, 2001
TIME: Noon - 1 pm
PLACE: Wean Hall 8220

SPEAKER:
Matt Welsh
UC Berkeley Computer Science Division

TITLE:
SEDA: An Architecture for Well-Conditioned, Scalable Internet Services

ABSTRACT:
Modern Internet services are increasingly dynamic, with massive concurrency demands as well as enormous variations in load. The well-documented "Slashdot effect" demonstrates what happens to an Internet service when faced with a spontaneous load many times that of the average: typically, the service buckles under the overload, giving the impression that it has crashed. On September 11th, the majority of news Web sites worldwide were unavailable for several hours due to an unprecedented spike in traffic.

The dynamic nature of Internet services makes it difficult to engineer this problem away. Simply overprovisioning resources is infeasible when load spikes are orders of magnitude greater than the average. The load that a given service places on its underlying resources is also unpredictable, given that most services involve a combination of static content, dynamic scripting, database access, and other code. Traditional thread-based models of concurrency are designed for timesharing and do not scale to many thousands of simultaneous tasks.

We present a general-purpose software architecture that addresses the problem of supporting massive concurrency and load conditioning in a highly dynamic environment. We describe the Staged Event-Driven Architecture, or SEDA, in which applications consist of a network of event-driven stages connected by explicit queues. This architecture allows services to be well-conditioned to load, preventing resources from being overcommitted when demand exceeds service capacity. In addition, SEDA makes use of a set of dynamic resource controllers to keep stages within their operating regime despite large fluctuations in load.

This talk presents the SEDA design as well as an implementation of an Internet services platform based on this architecture. We discuss control mechanisms for automatic tuning and load conditioning, including thread pool sizing, event batching, and adaptive load shedding. Finally, we present several applications constructed using the SEDA framework, including a high-performance HTTP server that outperforms both Apache and Flash; a packet router for the Gnutella peer-to-peer file sharing network; and a Web-based email service. Our results show that SEDA applications exhibit higher performance than traditional service designs, and are robust to huge variations in load.

BIO:
Matt Welsh is a PhD student at UC Berkeley working with Professors David Culler and Eric Brewer, and expects to graduate in May 2002. His research interests include Internet service architectures, distributed systems, language and compiler support for building robust systems, workstation clusters, and system area networks. Matt received his MS from UC Berkeley and his BS from Cornell University, and has spent time at Fujitsu Labs, University of Cambridge, University of Glasgow, and the Vrije Universitat. He is a longtime Linux developer and evangelist, and is the coauthor of "Running Linux", published by O'Reilly and Associates.

SDI / LCS Seminar Questions?
Karen Lindenfelser, 86716, or visit www.pdl.cmu.edu/SDI/