Pegasystems provides hardware sizing service and it gives a thick PDF file that describes the hardware resources such as RAM, storage size, and CPU. I usually use this at early phase of a project to get a rough idea of what kind of hardware client has to purchase. However, I never thought it as the solid resources that client could rely on but it has to be always validated by actual load test because no one call tell how much of memory would be consumed depending on how developer codes.
However some client believes in this sizing output as it is very solid, reliable metrics that would not really change after development. I just wanted to get some insight upon how actually people see this hardware sizing output?
Hardware sizing output is as good as the answers we provide to questionnaire.
There are lot of questions asked in that which we may not able to give an accurate answer in the early phases of the project. My usual practice is get one at the beginning of the project and another during the construction once all design elements and NFRs are clear.
In spite of all this, if you are using a framework i see that discrepancies are more. Sometimes Pega wants to follow the questionnaire format and it will be difficult to provide exact answer. Sometimes we had to have a call and explain the situation. We had deviations with respect to how we display our customer 360 and how we manage an interaction case.
So, in my experience i have always used hardware sizing pdfs as one of the inputs for capacity planning, but i wouldn't call it as rock solid. I am interested to see other inputs.