Capacity Based Load Balancing Scheme for Fair Request Dispatching

摘要

The load balancing scheme can achieve high performance for the web server. Most of the load balancing architectures are based on homogeneous web servers. If the backend servers in the system are different hardware specifications, the load balancing system must have a strategy to evenly dispatch the load to the backend servers. In this study, we derive a formula to define the capacities for heterogeneous web servers. In the experimental results, the maximum connection number with certain drop rate can be the capacities, but it can not get fair response time for each client requests. We must consider the capacities not only depending on drop rate but also on the response time. Under this definition, the response time for all client requests will nearly be the same.