Many cloud technologies available today support dynamically scaling out and back computing services. The predominantly session-oriented nature and the carrier-grade requirements of telco services (such as SIP services) complicate the successful adoption of dynamic scaling in a telco cloud. This paper investigates how to enable dynamic scaling of these telco services in an effective manner, focusing in particular on call-stateful SIP services. First, we present and evaluate two protocols to transparently migrate ongoing sessions between call-stateful SIP servers. These allow to quickly shutdown call-stateful SIP servers in response to a scale back request, removing the need to wait until their ongoing calls have finished. Second, instead of responding to load changes in a reactive manner, this paper explores the potential value of pro-active resource provisioning based on call load forecasting. We propose a self-adaptive Kalman filter to implement short-term call load predictions and combine this with history-based predictions to anticipate future call load changes. We believe that session migration and call load forecasting are two important elements to safely reduce the operational expenditure (OpEx) of a cloudified SIP service. © 2012 IFIP International Federation for Information Processing.
CITATION STYLE
Janssens, N., An, X., Daenen, K., & Forlivesi, C. (2012). Dynamic scaling of call-stateful SIP services in the cloud. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7289 LNCS, pp. 175–189). https://doi.org/10.1007/978-3-642-30045-5_14
Mendeley helps you to discover research relevant for your work.