Go to Home Page
Scientific Program
Docs by topics
Social events
Conference location
 • Grid WShop & Tutorial
 • Grid Program
 • Visiting Padova
 • INFN Padova
 • University of Padova
 • CHEP: '94 '95 '97 '98

Last update: Apr 1, 2000

to first abs  to previous absby abs number to next abs  to last abs


to first abs on this KT  to previous abs on this KTon same keytopic to next abs on this KT  to last abs on this KT


Using NetLogger for Performance Analysis of the BABAR Data Analysis System

Brian Tierney1, Dan Gunter1, Bob Jacobsen1, David Quarie1, Jacek Becla2
  1. Lawrence Berkeley National Lab
  2. Stanford Linear Accelerator (SLAC)

Speaker: Brian Tierney

  Developers and users of high-performance distributed systems often observe performance problems, the reasons for which can be manifold and are frequently not obvious. Bottlenecks can occur in any of the components along the paths through which the data flow: the applications, the operating systems, the hosts, or the network. We have developed a methodology, known as NetLogger, for detailed, end-to-end, top-to-bottom monitoring and analysis of all components. Distributed application components are instrumented to perform precision timestamping and event logging every critical point (e.g.: all I/O and any significant computation). The events are correlated with the host and network monitoring to characterize the performance of all aspects of the system in detail during actual operation. NetLogger includes a set of tools which make it easy for distributed applications to log interesting events , and includes tools for host and network monitoring. The approach is novel in that it combines network, host, and application-level monitoring to provide a complete view of the entire system. The NetLogger Toolkit also includes tools that allow for exploratory, interactive, real-time analysis of the log data.
  NetLogger instrumentation is being added to the computing systems that are used at SLAC for the BaBar Project. The computing environment for the BaBar data analysis is very complex, and consists of a large number of components. At the time of this writing, the performance of the BaBar data analysis system is very inconsistent: the amount of time to perform the same analysis varies by up to a factor of 10. NetLogger instrumentation is being added to the BaBar environment, and we hope to use NetLogger to diagnose the causes of these performance variations.

Presentation:  Adobe Acrobat pdf  PowerPoint Short Paper:  Adobe Acrobat pdf 

  | Top | Home | Bulletins | Committees | Scientific Program | Docs by topics | Social Event | Conference Location | Secretariat |