SLAMD provides a system for performing thorough benchmarks of network applications. Not only does it provide a means of measuring the performance of the application being tested, but it also offers capabilities to measure aspects of the system in which it is running, including CPU utilization, disk I/O, and network throughput. These system metrics can be collected using resource monitor clients, and just as SLAMD provides the ability to develop custom jobs to define new kinds of workloads, it also provides a means of developing custom resource monitors to capture a variety of system metrics.

As with all components of SLAMD, resource monitors are written in Java. There are many similarities between developing jobs and developing resource monitors, and as such it is expected that anyone interested in developing custom resource monitors has read the SLAMD Job Developer’s Guide.

All SLAMD resource monitor implementations must be subclasses of the com.sun.slamd.resourcemonitor.ResourceMonitor superclass. This is an class that defines a number of abstract methods that must be defined by any actual resource monitor implementation, as well as a number of helper methods that may be called by monitors. This section documents the abstract method stubs that must be implemented by resource monitors. The helper methods will be covered in the next section.

Download