Capacity Management and ProvisioningWhen should IT organizations practice capacity management by conducting a Capacity Planning study? When should they Model the Performance of their Environment?

With the constant changes that IT organizations face today, the answer is that most of these organizations should be modeling performance on a daily basis and conducting capacity planning studies recurring throughout the year. In a strong economy when businesses are experiencing rapid growth in their infrastructures, capacity planning allows IT organizations to understand their resource needs and accurately plan for anticipated growth, ensuring that the infrastructure will always be able to support the growing business. In a weaker economy, Capacity Planning allows IT organizations to optimize existing resources, identify and eliminate unused capacity through consolidation or workload balancing, justify new acquisitions, and conduct "purchase timing analysis" to determine exactly when and where critical resources will be needed.

Why Predict?
What makes  Predict the high value, unique solution that no one else in the market can offer? BMC Software has built a "Predict engine" that uses advanced, analytical queuing theory to measure and calculate relationships between Utilization, Response-Time and Throughput. These are essential metrics in determining the breaking point of a transaction processing system. Response-time and throughput for systems and workloads are determined by analyzing observed performance metrics from multiple system resources like memory, CPU and I/O and creating predictive queuing models. Predict can accurately, quickly, and precisely determine where and when future system bottlenecks and failures will occur. Predict enables primary end users to quickly answer "What-if?" questions about adding users, work or resources, in terms of the impact on response time, utilization and throughput.

Capacity Planning versus Trending
There are products claiming to be "Capacity Planning" solutions, but they’re simply graphs of utilization metrics over time, estimating the breaking point of a system based upon extrapolation of single resource trend lines. Furthermore, their trend line is always an irrelevant linear straight line. Why isn’t this real Capacity Planning? Why isn’t this at all accurate? Response-time is non-linear with respect to utilization. Response time and throughput are never straight - computers and their application, being queuing-based transaction processing systems, behave like non-linear queued systems. Trending a single resource when trying to determine the overall performance of a system does not accurately reflect reality because applications use and share multiple resources (memory, CPU, I/O). To understand and predict the combined effects of multiple non-linear resources on application response time requires Predict - the only analytic queuing theory-oriented solution available. Predict offers accurate and proven capacity planning and modeling technology that allows primary end-users to confidently make system tuning, upgrade and consolidation decisions for their IT environment.