Efficient Management of Telecom Components in Hybrid Cloud : Case SMSC
No Thumbnail Available
School of Science | Master's thesis
 + 51
AbstractTelecom operators use dedicated, large and expensive computing platforms for delivering, managing and billing the services. Decrease in Average Revenue Per User (ARPU) and increase in the operational expenditure (OPEX) forced operators to look for efficient solutions to decrease amount of operating and infrastructure expenses while guaranteeing the SLA for the provided services. Rapid growth in the area of virtualization and emergence of efficient virtual machine managers promises telcos to operate at minimum expense by offering large computing capacity and storage services at lower price by enforcing the pay per use principle. In order to demonstrate the cost benefits of cloud computing, Short Message Service (SMS) is selected as the use case. Since the deployment of SMS Center (SMSC) systems is done based on the traffic trend analysis of a particular location and due to its static nature the desired Quality of Experience (QoE) is not achieved during unexpected message peaks. Hence Telecom operators have a challenging task to optimize their infrastructure both technically and economically. Scaling the resources on demand, however, using the Infrastructure as a Service (IaaS) delivery model helps operators in designing the most feasible and cost efficient solution. This thesis presents an implementation and performance evaluation of an automated elastic framework built on hybrid cloud deployment. Applying the hybrid cloud approach to meet the requirements of SMSC key attributes is discussed as well. Message traces from a Mobile Network Operator (MNO) are used as input to verify the performance of our proposed framework under different deployment models. Compared with public and private cloud deployments, the results prove that under a variable work load our solution can process the requests at low cost while maintaining the acceptable QoE.
Thesis advisorRaivio, Yrjö
message systems, cloud computing, hybrid intelligent systems, hybrid architecture, load modelling, SLA