Optimized Load balancing System for Large Scale Applications

Abstract

Distributed and parallel computing is an important trend today. Scaling, performance, response time, and server load are considered very important when it comes to large scale applications. Most of the clients need concurrent access for their web services. Large number of requests come to the application and database servers at a time. Requests need to be handled in a very short time period. Therefore response time is a very important fact. Thus a good system should give responses to all the requests and also should maintain a short response time. Providing a solution including both of these factors is a well-known problem. Load balancing and clustering play important roles in research. A single server is not enough for managing a huge number of requests. A proper load balancing technique should be implemented to overcome this problem. There are different kinds of load balancers for different purposes. But in the end the ultimate goal is to serve all the requests with a good response time. In this project, we have developed a new load balancing technique. We consider the CPU usages of the server nodes as the load. The technique is implemented using python programming language. The results of our technique are very promising as we managed to archive a response time speed up for a large scale application having different weighted requests.