Delay estimation method for N-tier architecture

These days, the majority of large systems serving large numbers of users prefer N-tier architecture because of its scalability and flexibility. To maintain the quality of service in these systems, we need to understand how much delay is generated in each tier. However, the structure and the behavior of this architecture is so complicated that it is difficult to analyze these delays without installing special software or hardware - improvements that might change the server's behavior or result in significant additional costs. To solve these problems, we developed a practical method for estimating the delays generated in each tier of N-tier architecture that uses only easily obtainable parameters. In this paper, we first discuss what these easily obtainable parameters are. We then construct a performance model of the N-tier architecture using these parameters and analyze the delays in the model. Finally, we describe the experiments we conducted to evaluate our approach.