Click here to view and discuss this page in DocCommentXchange. In the future, you will be sent there automatically.

SQL Anywhere 10.0.1 » SQL Anywhere Server - SQL Usage » Monitoring and Improving Performance » Advanced application profiling using diagnostic tracing » Analyzing diagnostic tracing information

Detecting slow statements Next Page

Detecting when hardware resources are a limiting factor

As larger and larger workloads are placed on a database system, performance is eventually limited by one or more hardware resources. Typically one of three resources is exhausted: CPU cycles, memory space, or disk I/O bandwidth. When this happens, it may be because there are inefficiencies in the application or in how the database server is performing. If no inefficiencies are detected, you may need to add additional hardware resources to allow the database server handle larger workloads, and to improve performance on the existing workload. To view a list of common inefficiencies and recommendations on how to solve them, see Troubleshooting performance problems.

Adding resources may not resolve all scalability problems. Furthermore, even when resources are added, the capability of the computer is rarely improved in a linear fashion. For example, if it appears that the database server is fully using the allotted CPUs, it indicates that more CPU resources must be assigned in order to handle a larger workload. However, doubling the number of CPUs available to the database server likely will not double the amount of work the database server can perform in the same amount of time.

You can examine several database statistics using the Statistics tab in the Application Profiling Details area to detect whether hardware resources are a limiting factor for performance.

See also