Bw man: Bandwidth manager for elastic services in the cloud
2014 (English)In: Proceedings - 2014 IEEE International Symposium on Parallel and Distributed Processing with Applications, ISPA 2014, Institute of Electrical and Electronics Engineers Inc. , 2014, p. 217-224, article id 6924450Conference paper, Published paper (Refereed)
Abstract [en]
The flexibility of Cloud computing allows elastic services to adapt to changes in workload patterns in order to achieve desired Service Level Objectives (SLOs) at a reduced cost. Typically, the service adapts to changes in workload by adding or removing service instances (VMs), which for stateful services will require moving data among instances. The SLOs of a distributed Cloud-based service are sensitive to the available network bandwidth, which is usually shared by multiple activities in a single service without being explicitly allocated and managed as a resource. We present the design and evaluation of BwMan, a network bandwidth manager for elastic services in the Cloud. BwMan predicts and performs the bandwidth allocation and tradeoffs between multiple service activities in order to meet service specific SLOs and policies. To make management decisions, BwMan uses statistical machine learning (SML) to build predictive models. This allows BwMan to arbitrate and allocate bandwidth dynamically among different activities to satisfy specified SLOs. We have implemented and evaluated BwMan for the OpenStack Swift store. Our evaluation shows the feasibility and effectiveness of our approach to bandwidth management in an elastic service. The experiments show that network bandwidth management by BwMan can reduce SLO violations in Swift by a factor of two or more.
Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers Inc. , 2014. p. 217-224, article id 6924450
Keywords [en]
Bandwidth Management, Cloud Computing, SLO, Artificial intelligence, Managers, Design and evaluations, Distributed clouds, Management decisions, Service instances, Service level objective, Statistical machine learning, Bandwidth
National Category
Engineering and Technology
Identifiers
URN: urn:nbn:se:ri:diva-46169DOI: 10.1109/ISPA.2014.37Scopus ID: 2-s2.0-84911421776ISBN: 9781479942930 (print)OAI: oai:DiVA.org:ri-46169DiVA, id: diva2:1457902
Conference
12th IEEE International Symposium on Parallel and Distributed Processing with Applications, ISPA 2014, 26 August 2014 through 28 August 2014
Note
Conference code: 108694
2020-08-132020-08-132023-05-08Bibliographically approved