EVMS Performance Evaluation for Linux

Resources

Status
Results
Scripts

Introduction

The purpose of this study is to evaluate and improve the performance and scalability of EVMS for Linux.  The effort described here is currently maintained by AndrewTheurer .  This is an open project, and I encourage others to join my efforts.  The tests conducted for this evaluation currently takeplace at IBM's Linux Technology Center.  Check our status page for latest news.

Project Process

The general idea of this project is to cycle through a process of measurement, analysis, and improvement of performance.
 
Measurement
In this stage we run the benchmark as specified by run rules and verify the results. Currently we use IOzone, dd, and dbench for our tests.

Analysis
In this stage we use get a closer look at what's going on during the tests. We do kernel profiling and look for possible areas of improvement.

Improvement
We target code in the EVMS runtime, located in ./drivers/evms in Linux kernels (patched with EVMS). We generally try to improve code path efficiency and reduce disk latency.


This process should be repeated until we reach a target level of performance and scalability.   To find out what we're workinig on now, check out the project status.

Test Environment

So far all tests have been run on an IBM x220 server, which has 2 800 MHz PIII processors, 768 MB memory, and 3 18 GB, 15k RPM SCSI disks.