SES/Strategizer

Simulate your environment for testing

If you plan to deploy a new client/server application, you might want to know how the deployment will affect your systems and network. SES/strategizer 2.1 is the tool you need to predict how such a deployment will affect your network.

Strategizer is a modeling and simulation tool. It lets you build a model of your environment, including your applications, network, computers, databases, and user behavior. After you define these elements, you can run simulations on the model and measure a variety of factors, such as network utilization and CPU utilization. You can use these measurements to evaluate your environment.

Strategizer can make adjustments in your model's configuration, rerun the simulation, and compare the results with results from previous simulations. For example, you can change the operational characteristics of your model's server and rerun the simulation to find out how adding 128MB of RAM will affect your database server's response to user requests. Strategizer lets you test various configurations without putting your network at risk.

By running different simulations, you can analyze the effects that deploying new systems will have on your environment. This simulation process can save you time and money by helping you avoid purchasing improperly sized equipment and making piecemeal attempts to correct performance pitfalls.

I installed Strategizer on a Pentium II machine running Windows NT Workstation 4.0. I inserted the installation CD-ROM and answered questions about my username, company name, and installation location. Installation took 5 minutes.

To test Strategizer, I decided to examine how a new public safety incident-reporting system would affect my network and how the application would increase CPU utilization on my database server. I started building a model of my environment using the drag-and-drop-enabled GUI that Screen 1 shows. I selected icons to represent the components I wanted to include in my model and specified the components' relationships.

You must include your systems' processes and services, define user behavior, and add instances of the behavior to your model. For my testing, I defined several components, including my network topology, my database server and tables, workstations that access the database, and information about how users access the database.

After I defined all my network's components, I started running simulations. In the Simulation menu, I selected the statistics I wanted to examine. Then, I clicked the Run simulation icon. I received several Application Definition Notation (ADN) errors. (ADN is the internal mechanism that defines how components operate and interact.) I had to spend several hours reading the software's Windows Help file and employing trial-and-error methods to adjust my components' parameters so that the simulation would run. When I corrected the ADN errors, the simulation ran successfully.

Although Strategizer's simulation capabilities are helpful, they are also complicated to use. Strategizer attempts to simplify the process with a GUI, but the product requires a steep learning curve. In addition, the product documentation isn't good—the only documentation is the Windows Help file that the program includes.

When you use Strategizer, your simulation results will be only as good as the model you create. If your model is flawed, the results will not be accurate. You must spend a significant amount of time making sure the data you put into Strategizer is accurate.

The suggested list price for Strategizer is $25,000, which includes first-year maintenance and one seat in an SES-sponsored Strategizer training class. Its cost elevates this tool to the corporate enterprise level. Small or midsized businesses deploying new client/server applications might have to rely on other ways of determining the applications' potential effect on their environment.