ABSTRACT: Application Service Providers (ASPs) may obtain resources from a number of cloud service providers (CSPs) in an attempt to improve latency and minimize operational expenses (OpEx). The CSPs may use management and control platforms, such as OpenStack and EC2 and the network service providers (NSPs) may use network management platforms, such as, OpenDaylight. However, today the ASPs do not have a common management and control platform that would present to them a converged view of all the cloud and network resources. OpenADN being developed at Washington University in Saint Louis aims to allow the ASPs dynamic and real time control of virtual resources across multiple clouds and networks to provide efficient application delivery. The OpenADN platform itself is a complex distributed and multi-threaded system. Performance evaluation and assessment of need for optimization of such a complex platform requires precise and fine-grained behavioral data. In this paper we establish the need for profiling OpenADN like platforms so that the ASPs can optimize its behavior and control their cost, performance (latency) and energy consumption.
Complete paper in Adobe Acrobat format.
The conference presentation slides are available at: http://www.cse.wustl.edu/~jain/talks/profilep.htm.