Previous topic |
Next topic |
Contents |
Contact z/OS |
Library |
PDF
How is performance characterized? z/OS Network File System Guide and Reference SC23-6883-00 |
|||||||||||||||
Performance is the manner in which a process, system, processor, network, or device behaves for a particular load or unit of work. To measure, or quantify, performance, we monitor the length of time for a unit of work to complete. If units of work are being shared, we monitor the amount of time waiting for a resource to be available to perform a unit of work. A unit of work is a specific activity or action that we expect a process, system, processor, network, or device to perform. This could be something as granular as an I/O request, sending or receiving a buffer of data over a network, or processing an NFS request. We are frequently interested in the performance of a particular set of work activities which we will refer to as a load of work, or workload. When a particular workload has been identified for performance measurement, we can determine the performance metrics, or units of measurement, that are relevant to that workload. Some examples of the following performance metrics that might be used in reference to performance tuning for z/OS NFS are:
Measuring performance metrics can be as simple as using your watch to time the execution of a particular command or as complex as using specialized hardware and software tools to monitor and extract a diversity of performance metrics. Evaluating z/OS NFS performance will address some methods that can be used to evaluate the performance of z/OS NFS. The method selected will depend on the complexity of the workload and the monitoring tools available at your installation. |
Copyright IBM Corporation 1990, 2014
|