Open Cirrus Summit 2011

Monitoring as a Service in the Cloud
-- Sicong Meng, Georgia Tech

Abstract:
System monitoring is an indispensable capability of cloud computing for both cloud service users and providers. With the massive scale of cloud systems and the complexity of cloud applications, providing efficient and scalable monitoring functionalities to cloud administrators and users remains a grand challenge. Delivering monitoring as a service (MaaS) in the cloud not only makes monitoring functionalities easily accessible and reduces application maintenance cost, but also provides unique service and workload consolidation opportunities to further enhance the efficiency, scalability, reliability and customizability of state monitoring. We conduct our research by identifying the problems and opportunities in system and application state monitoring domains, developing practical and scalable solutions, and evaluating our solutions by providing experimental and analytical comparison with existing approaches. We present a suite of key functional requirements in MaaS, including instantaneous state monitoring and window-based monitoring, periodic and violation-likelihood based monitoring, single tenant and multi-tenant state monitoring. Furthermore, these monitoring functions should also meet quality-of-service demands on pursuing monitoring accuracy, enhancing monitoring scalability, and improving monitoring effectiveness through consolidation and isolation.

Bio:
Shicong Meng is currently a final year PhD student in the College of Computing at Georgia Tech. He is affiliated with the Center for Experimental Research in Computer Systems (CERCS) where he works with Professor Ling Liu. His research focuses on performance, scalability and security issues in large-scale distributed systems such as cloud datacenters. He has recently been working on several projects on cloud datacenter monitoring and management, with a strong emphasis on cloud monitoring services. He has worked as an intern in IBM T.J. Watson research center and VMware. He is also a recipient of IBM PhD fellowship award (2011-2012).