Server Performance Tips

The following information may help you configure your Spectrum environment and your servers to optimize their performance.

Configuring for High Availability with Distributed Services

If you are planning a Spectrum environment that will include more than one Spectrum Application Server, it is recommended that you configure your environment to support distributed services. In a Spectrum environment with distributed services, all Spectrum Application Servers are configured to interact as peers within the Spectrum environment, all accessing the same Spectrum database.

In such an environment, you use distributed services to manage which server or servers are providing Spectrum functionality at any time. In Spectrum, you can configure distributed services for high availability or for failover only. It is recommended that you configure for high availability of distributed services in Spectrum. This approach involves configuring Spectrum to load balance those services that support load balancing so that they are run concurrently on multiple Spectrum Application Servers. Services that do not support load balancing should be configured to fail over among Spectrum Application Servers so that if one or more servers go offline, a standby server or servers begin running the associated services to minimize any interruption for users.

For information about how to install Spectrum to support distributed services, see the Loftware Spectrum Installation and Configuration Guide. For information about configuring distributed services, see High Availability with Distributed Services.

Optimizing Performance

The following tips may help you to optimize the performance of your Spectrum configuration.

  • Consult your database administrator for guidance about database tuning.
  • The more memory that you have configured for the Spectrum Application Server, the greater the number of devices, label templates, and users that it can support.
  • Increasing the amount of memory available to the database server typically improves performance. If you can make significantly more memory available, you can dramatically increase performance because all or a significant portion of the label and configuration data can be cached in memory.
  • The amount of space on the database server required for recent history data is typically greater and more variable than the space required for label and configuration data. However, if you can make the amount of memory required for recent history data available on the database server, you can dramatically increase the performance of reporting.
  • Sufficient network bandwidth must be available to support your expected printing throughput.