SIMALYTIC MODELING: A HYBRID TECHNIQUE FOR CLIENT/SERVER CAPACITY PLANNING
Tim R. Norton · 1997
The Simalytic^TM Modeling Technique (from Simulation/Analytic) is a hybrid technique that addresses modeling and predicting the capacity requirements of computer systems in complex enterprise-wide client/server multiple-platform applications. This technique uses a general purpose simulation tool as an underlying framework and an analytic tool to represent application response times at individual nodes. The bridge between the two techniques is a transform function that will adjust the service time for a given server depending on the load at that server. It combines both platform-centric tools (limited features but detailed platform information) and general purpose tools (rich low level features) to address today's large heterogeneous enterprises. This methodology takes advantage of features in the different techniques (simulation vs. analytic queuing theory) as well as features in the different tools (platform-centric vs. general purpose) by defining the interface between the simulation framework and queuing theory node models. The benefits of using a hybrid technique are discussed and results are presented to show the validity of the Simalytic technique.