Unsupported content

 

This version of the product is no longer supported. However, the documentation is available for your convenience. You will not be able to leave comments.

Network usage tuning

BMC Discovery, through its interaction with the IT infrastructure, generates network traffic. The amount of traffic is entirely dependent on the amount of work that BMC Discovery is doing. Some factors that could influence variations in peak load include:

Some factors that could influence variations in peak load include:

  • Custom patterns—Most of the network load is generated by BMC Discovery scanning the target environment. In addition to the core discovery required to build Host nodes, discovery requests resulting from patterns will generate some network traffic. BMC builds the patterns that are shipped in the Technology Knowledge Updates (TKU) to limit the amount of data that is requested. Similar care should be taken when developing custom patterns for your environment. For example, if a pattern were to retrieve the contents of a very large file that is common in the target environment, this could result in a larger consumption of network bandwidth.
  • Consolidation—Consolidation results in the transfer of data between appliances. Consequently, some additional network load will be added.
  • Moving appliance backups—It is possible to move appliance backups to or from BMC Discovery appliances. However, higher network traffic will occur during the transfer of these potentially large files.
  • Clusters—With Cluster deployment, BMC Discovery might consume more bandwidth than a standalone setup for the following reasons:
    • Management overhead—The coordinator bandwidth for group coordination and additional traffic between cluster members is negligible in a normal situation. However, while the cluster is recovering from failure of one or more members, the bandwidth consumption increases significantly.
    • Concurrent discovery—As the cluster size grows, the bandwidth consumed by the cluster members increases with the rate of discovery, as the cluster members share the discovery workload.
    For better performance, you should co-locate the cluster members on a switched network with low latency. 

Was this page helpful? Yes No Submitting... Thank you

Comments