Laboratory for Internet and Innovative Technologies

Cloud service providers (CSPs) aim to preserve infinite scalable and elastic computing resources. In addition, user requirements for flexible cloud computing resources impact the overall performance of the cloud. From the other side, virtualization adds additional layer in cloud stack and therefore decreases the service performance compared to onpremise traditional hosting. It seems that the price to pay for scalability and flexibility is not so small. Therefore, in order to improve the overall cloud performance and make it sustainable there is need for real-time resource orchestration. Considering the heterogeneous architecture of the cloud, CSPs must develop appropriate strategy to achieve maximum performance minimizing the utilization of hardware resources and energy consumption. In this paper, we propose a novel Low Level Load balancing (L3B) approach applied to Infrastructure as a Service (IaaS). Our approach preserves the cloud’s elasticity by dynamic activation of cloud resources and load balancing the traffic over the resources on low network level. In addition, our L3B balancer tends to load the instances in the region where they provide maximum performance. Introducing L3B improves the overall cloud performance, reduces power consumption and customers’ cost for renting cloud resources, etc.


Monika Simjanoska, Sasko Ristov, Goran Velkoski, and Marjan Gusev


Cloud Computing, IP, Networking, Performance

Full Paper

The paper is published in EUROCON 2013, Proceedings of the 8th Int. Conference on computer as a tool, Zagreb, Croatia, IEEE, ISBN: 978-1-4673-2231-7, 2013