Successful implementation of L3B: Low level load balancer

Cloud computing paradigm offers instantiating and deactivating the virtual machine instances on demand according to the clients requirements. When some customer´s application or a service needs more resources than a physical server can supply, the cloud provider offers a certain load balancing technique to distribute the load among several servers that will host the analyzed application or service. Additionally, the cloud provider should offer a resource broker to make the application scalable and elastic. In this paper we present a new solution for a low level load balancer, working on a network level. Our load balancer maps the memory addresses of the balancer and the target physical servers (or virtual machines on the cloud) and thus balances the load. The experiments showed that it adds small latency of several milliseconds and thus it slightly reduces the performance when the distributed system is underutilized. However, there is a region of client requests where the system achieves a superlinear speedup (speedup greater than the number of scaled resources). Our case study with doubled resources achieves up to 6.5 speedup (maximum expected linear speedup is 2).