Article in Journal ART-2014-04

BibliographyMega, Cataldo; Waizenegger, Tim; Lebutsch, David; Schleipen, Stefan; Barney, J.M.: Dynamic cloud service topology adaption for minimizing resources while meeting performance goals.
In: IBM Journal of Research and Development. Vol. 58(2).
University of Stuttgart, Faculty of Computer Science, Electrical Engineering, and Information Technology.
pp. 1-10, english.
IBM, March 2014.
DOI: 10.1147/JRD.2014.2304771; ISSN: 0018-8646.
Article in Journal.
CR-SchemaH.2 (Database Management)
H.3 (Information Storage and Retrieval)
D.4.8 (Operating Systems Performance)
KeywordsCloud computing; Electronic countermeasures; Network topology; Optimization; Resource management; Service agreements; Time measurement
Abstract

Even in the cloud computing era, meeting service-level agreements (SLAs) of a computing service or application while significantly reducing the total cost of ownership (TCO) remains a challenge. Cloud and software defined environments (SDEs) are offering new opportunities for how resources can be utilized to an even higher degree than before—which leads to a reduced TCO for service providers and customers of a service. The traditional method of meeting an SLA is to assess peak workloads and size a system accordingly. This still leads to very low average compute resource utilization rates. This paper presents a novel dynamic and cost-efficient orchestration approach of multi-tenant capable, software defined system topologies based on a monitor-analyze-plan-execute (MAPE) concept. We present the mechanism involved in creating and applying these heuristics and show the results for a cloud-based enterprise content management (ECM) solution. Our approach allows the cloud provider to minimize its resource requirements while staying in compliance with SLAs.

Contacttim.waizenegger@ipvs.uni-stuttgart.de
Department(s)University of Stuttgart, Institute of Parallel and Distributed Systems, Applications of Parallel and Distributed Systems
Entry dateMay 5, 2014
   Publ. Institute   Publ. Computer Science