Stardust
Authors:
Open Access
- 26 June 2006
- journal article
- Published by Association for Computing Machinery (ACM) in ACM SIGMETRICS Performance Evaluation Review
- Vol. 34 (1) , 3-14
Abstract
Performance monitoring in most distributed systems provides minimal guidance for tuning, problem diagnosis, and decision making. Stardust is a monitoring infrastructure that replaces traditional performance counters with end-to-end traces of requests and allows for efficient querying of performance metrics. Such traces better inform key administrative performance challenges by enabling, for example, extraction of per-workload, per-resource demand information and per-workload latency graphs. This paper reports on our experience building and using end-to-end tracing as an on-line monitoring tool in a distributed storage system. Using diverse system workloads and scenarios, we show that such fine-grained tracing can be made efficient (less than 6% overhead) and is useful for on- and off-line analysis of system behavior. These experiences make a case for having other systems incorporate such an instrumentation framework.Keywords
This publication has 10 references indexed in Scilit:
- Informed data distribution selection in a self-predicting storage system2006
- Hibernator2005
- Request extraction in Magpie2004
- The ganglia distributed monitoring system: design, implementation, and experienceParallel Computing, 2004
- High Resolution Forward And Inverse Earthquake Modeling on Terascale Computers2003
- Performance debugging for distributed systems of black boxes2003
- Pinpoint: problem determination in large, dynamic Internet services2003
- Trace-based mobile network emulation1997
- Mache: no-loss trace compactionACM SIGMETRICS Performance Evaluation Review, 1989
- A trace-driven analysis of the UNIX 4.2 BSD file systemACM SIGOPS Operating Systems Review, 1985