IBM Tivoli NetView for z/OS, Version 6.2

Identifying Data-Processing Requirements

To support overall business goals, data-processing departments typically have requirements of their own. The requirements might be objectives like these:

Increase system availability by 10 percent over the next 2 years.
Accommodate 12 percent growth capacity (in millions of instructions per second, or MIPS) and network resources over the next 2 years with no increase in operation staff.
Improve system performance by 15 percent each year for the next 5 years.

Data-processing requirements typically fall into two classes: system-oriented and user-oriented. System-oriented requirements measure the amount of information that your systems process. These requirements include:

Expected batch throughputs
Workloads on each system
Interactive transaction rates
The number of concurrent users that you can support

By contrast, user-oriented requirements measure the impact of data-processing services on the user. Examples are expected response times for interactive work and expected turnaround times for batch work.

Service-level agreements reflect these expectations of performance. A service-level agreement resembles a contract between the data-processing department and that department's users. A service-level agreement might specify the services you provide, the hours you provide them, and various agreed measures of availability and performance. Whereas other requirements often represent goals that you want to accomplish, service-level agreements state minimums that you must accomplish.

If your organization does not use service-level agreements, or if your service-level agreements do not accurately reflect your goals, consider establishing agreements that are based on your goals. Service-level agreements can help you measure the improvements in service to users that automation provides. They can also help you identify problem areas of your operation that might benefit from automation.

Feedback