Skip to content
CA Spectrum - 10.3
Documentation powered by DocOps

How CA Spectrum Calculates CPU and Memory Utilization

Last update April 28, 2017

The MIB objects that are read to calculate the source of the CPU and memory utilization can be customized for an individual device model or set as the default for a given model type. This is only necessary if CA Spectrum is unable to identify a source for CPU and memory utilization, or if you prefer to use a different source.

Do one of the following to customize CPU and memory utilization:

  • Modify the order and the types of sources to be tested.
  • Modify the source attributes used by the attribute redirection type.

Attribute redirection is the process of using well-known attributes as a pointer, or redirectors, to proprietary attributes. For example, the well-known attribute NRM_CPUUtilAttr (0x12e2d) on the AirespaceSw model type holds the attribute ID of the Airespace proprietary CPU attribute, agentCurrentCPUUtilization (0x4b605ae).

With attribute redirection, it is possible for generic code to reference unique attributes for each device model or model type. Additionally, attribute redirection lets you change the source attributes without having to restart the SpectroSERVER. By default, for most model types, attribute redirection is the first source type that is tested. Therefore, in most circumstances, you only have to modify the attributes that attribute redirection is using.

Note: Only certain attribute types can be used for attribute redirection.

Before you make changes to the normalized CPU and memory intelligence, understand how CA Spectrum calculates CPU and memory utilization.

CA Spectrum does the following to calculate CPU and memory utilization:

  1. CA Spectrum identifies the source that is used to calculate the CPU and memory utilization. CA Spectrum identifies the source any time the device is reconfigured.
    A list of the possible sources is provided in a new preference attribute. Each source is tested in order. If a valid source is found, the source intelligence, the attributes that are used, and the attribute's model type are stored for reuse during the utilization calculation. This helps ensure that the source does not change until the source is reconfigured. If the list of sources is empty, or a valid, functioning source is not identified, the source type is set to “None” and no further reads are made to the device.
    In general, the sources are tested in the following order:
    • Attribute redirection
    • CA proprietary intelligence
    • Standard intelligence (RFC 2790 and Net-SNMP)
  2. CA Spectrum performs the actual calculation of the utilization using the correct source, attribute IDs, and model type handles that were identified. The running attribute IDs list and the running model handle list are passed into the calculation method each time, to help ensure that the same attributes are read.

This section contains information about the following topics:

Was this helpful?

Please log in to post comments.