Resource contention is your biggest issue. Do you have enough threads/nodes handling concurrent requests? Do you have beefy hardware to support your system? You must look at the whole system from the GUI down to the hardware.
posted 14 years ago
Thank you very much, J!
Does this mean that I should calculate thread and connection pools sizes on the Application server. If that the case where should I put this information in the Assumptions section? Or my deployment diagram must reflect this info?
By the way, what is the good approach to calculate all this? Or just by looking on the specs of the Database and Application Servers.
As the delivery is just uml and assumptions, isn't the only real option of answering the concurrent user requirement to design it scalable. So that it's easy to add servers as the number of users increase? Prehaps set up a quantity of servers in the deployment diagram?