A Randomized Algorithm for Load Balancing in Containerized Cloud

摘要

Cloud computing is one of the highly discussed topics in the field of Internet and communication technology. It is responsible for the on-demand provision of computing resources, mainly data and computing power to the end-users. More than one server works together in a cloud network. So, the incoming request for resources must be distributed among all servers in the network for better performance. The process of efficiently distributing incoming tasks and sharing the workload among a group of servers is called load balancing. In this paper, we propose a randomized algorithm for load balancing in a containerized cloud. The approach we have used is called Balls into Bins via Local Search. In our algorithm, we have considered tasks as balls and servers as bins. First, we construct a fully connected undirected graph of nodes(server) and then convert it to a minimum edge weight graph to reduce the network cost. Our experimental result shows that the load is distributed among all servers almost equally. The difference between the highest and lowest workload in the network is minimized.