Imperial College > Eva Kalyvianaki

I am a Post-Doc Researcher working with Dr. Peter Pietzuch on the DISSP project.

Before this, I finished my Ph.D. at the Computer Laboratory in the SRG netos group in Cambridge University under the supervision of Dr. Steven Hand and Dr. Ian Pratt.

Prior to this I was a student in the Computer Science Department, University of Crete, Greece, where I obtained a M.Sc. and a B.Sc. degree and was supervised by Prof. Evangelos Markatos.

Email address: [ekalyv [ at ] imperial.ac.uk].

	I am a Post-Doc Researcher working with Dr. Peter Pietzuch on the DISSP project. Before this, I finished my Ph.D. at the Computer Laboratory in the SRG netos group in Cambridge University under the supervision of Dr. Steven Hand and Dr. Ian Pratt. Prior to this I was a student in the Computer Science Department, University of Crete, Greece, where I obtained a M.Sc. and a B.Sc. degree and was supervised by Prof. Evangelos Markatos. Email address: [ekalyv [ at ] imperial.ac.uk].

Research Interests	My main research interests are in analysing complex distributed systems and building online models for dependability, performance tuning purposes and fault-tolerance. I enjoy working in the areas of distributed applications, operating systems, performance evaluation of complex distributed systems and at the intersection of mathematical reasoning and systems research.

Research	Dependable Internet-Scale Stream Processing (DISSP): Internet-Scale Streaming involves the real-time handling of data coming from world-wide distributed sources using a network of processing nodes across the Internet. In this environment, failures of nodes cause severe disruption to the execution of any query stream. In the DISSP project, we explore different ways to build a dependable system against unprecedented failures of processing nodes and stream sources. We emphasize on the adaptability of the framework to mask failures and deliver results that reflect the state of the infrastructure at the time of the query. Working with marco fiscato and peter pietzuch.
	Adaptive Resource Management for Virtualized Servers: Data center virtualization allows cost-effective server consolidation which can increase system throughput and reduce power consumption. Resource management of virtualized servers is an important and challenging task, especially when dealing with fluctuating workloads and complex multi-tier server applications. Recent results in control theory-based resource management have shown the potential benefits of adjusting allocations to match changing workloads. Key results of our approach include: (a) the integration of the Kalman filtering technique into linear feedback controllers that self-adapt to any workload condition without any a priori information; (b) controllers designed to online use the pair-wise resource coupling between components in multi-tier applications and improve the server performance against sharp and important workload increases. Working with themistoklis charalambous and steve hand.

Publications	Integrating Scale Out and Fault Tolerance in Stream Processing using Operator State Management Raul Castro Fernandez, Matteo Migliavacca, Evangelia Kalyvianaki, and Peter Pietzuch ACM Int. Conf. on Management of Data (SIGMOD) , 2013 (to appear). Overload Management in Data Stream Processing Systems with Latency Guarantees Evangelia Kalyvianaki, Themistoklis Charalambous, Marco Fiscato, and Peter Pietzuch 7th IEEE Int. Workshop on Feedback Computing (Feedback Computing), 2012. [pdf] Adaptive Provisioning of Stream Processing Systems in the Cloud Javier Cervio, Evangelia Kalyvianaki, Joaqun Salvacha, and Peter Pietzuch 7th IEEE Int. Workshop on Self Managing Database Systems (SMDB'12), 2012. Load-Balancing Stream Processing with the Cloud Wilhelm Kleiminger, Evangelia Kalyvianaki, and Peter Pietzuch 6th IEEE Int. Workshop on Self Managing Database Systems (SMDB'11), 2011. [pdf] SQPR: Stream Query Planning with Reuse Evangelia Kalyvianaki, Wolfram Wiesemann, Quang Hieu Vu, Daniel Kuhn, and Peter Pietzuch IEEE Int. Conf. on Data Engineering (ICDE'11), 2011. [paper] [tech report with NP proof] [poster(pdf)] A Min-Max Framework for CPU Resource Provisioning in Virtualized Servers using H-infinity Filters Themistoklis Charalambous and Evangelia Kalyvianaki IEEE Conference on Decision and Control (CDC'10), 2010. Resource Provisioning for Multi-Tier Virtualized Server Applications Evangelia Kalyvianaki, Themistoklis Charalambous, and Steven Hand Computer Measurement Group Journal (CMG Journal'10), Spring Issue 126, pages 6-17, 2010. [pdf] Self-Adaptive and Self-Configured CPU Resource Provisioning for Virtualized Servers Using Kalman Filters Evangelia Kalyvianaki, Themistoklis Charalambous, and Steven Hand 6th Int. Conference on Autonomic Computing and Communications (ICAC'09), 2009. [ACM DL] Applying Kalman Filters to Dynamic Resource Provisioning of Virtualized Server Applications Evangelia Kalyvianaki, Themistoklis Charalambous, and Steven Hand 3rd Int. Workshop on Feedback Control Implementation and Design in Computing Systems and Networks (FeBid'08), 2008. [pdf] On Dynamic Resource Provisioning for Consolidated Servers in Virtualized Data Centers Evangelia Kalyvianaki and Themistoklis Charalambous 8th Int. Workshop on Performability Modeling of Computer and Communication Systems, (PMCCS-8 2007). [pdf] Tuning Server Applications with OS Virtualization Support using Feedback Control Evangelia Kalyvianaki CMG Fellowship Winner 2006. Daedalus: Statistical Aggregation for Large-Scale Dynamic Networks Evangelia Kalyvianaki and Yu-En Lu Infocom Student Workshop 2006. [pdf] Building Adaptive Peer-To-Peer Systems Evangelia Kalyvianaki and Ian Pratt 4th Int. Conference on Peer-to-Peer Computing (P2P 2004), short paper session, August 25-27, Zurich, Switzerland. Wormhole IP over (Connectionless) ATM Manolis G. H. Katevenis, Iakovos Mavroidis, Georgios Sapountzis, Eva Kalyvianaki, Ioannis Mavroidis and Georgios Glykopoulos IEEE/ACM Transactions on Networking (ToN 2001), 9(5), October 2001, pp 650-661.