FlexSplit: a workload-aware, adaptive load balancing strategy for media clusters

A number of technology and workload trends motivate us to consider a new request distribution and load balancing strategy for streaming media clusters. First, in emerging media workloads, a significant portion of the content is short and encoded at low bit rates. Additionally, media workloads display a strong temporal and spatial locality. This makes modern servers with gigabytes of main memory well suited to deliver a large fraction of accesses to popular files from memory. Second, a specific characteristic of streaming media workloads is that many clients do not finish playing an entire media file which results from the browsing nature of a large fraction of client accesses. In this paper, we propose and evaluate two new load-balancing strategies for media server clusters. The proposed strategies, FlexSplit and FlexSplitLard aim to efficiently utilize the combined cluster memory by exploiting specific media workload properties by "tuning" their behavior to media file popularity changes. The ability of the proposed policies to self-adapt to changing workloads across time while maintaining high performance makes these strategies an attractive choice for load balancing in media server clusters.

[1]  Philip S. Yu,et al.  The state of the art in locally distributed Web-server systems , 2002, CSUR.

[2]  Amin Vahdat,et al.  MediaGuard: a model-based framework for building streaming media services , 2005, IS&T/SPIE Electronic Imaging.

[3]  Wei Sun,et al.  Workload-aware load balancing for clustered Web servers , 2005, IEEE Transactions on Parallel and Distributed Systems.

[4]  Alec Wolman,et al.  Measurement and Analysis of a Streaming Media Workload , 2001, USITS.

[5]  Mary K. Vernon,et al.  Analysis of educational media server workloads , 2001, NOSSDAV '01.

[6]  Michael Stumm,et al.  Modular and Efficient Resource Management in the Exedra Media Server , 2001, USITS.

[7]  Amin Vahdat,et al.  Long-term Streaming Media Server Workload Analysis and Modeling , 2003 .

[8]  Katherine Guo,et al.  Silo, rainbow, and caching token: schemes for scalable, fault tolerant stream caching , 2002, IEEE J. Sel. Areas Commun..

[9]  Amin Vahdat,et al.  MediSyn: a synthetic streaming media service workload generator , 2003, NOSSDAV '03.

[10]  Ludmila Cherkasova,et al.  Characterizing locality, evolution, and life span of accesses in enterprise media server workloads , 2002, NOSSDAV '02.

[11]  Donald F. Towsley,et al.  Proxy prefix caching for multimedia streams , 1999, IEEE INFOCOM '99. Conference on Computer Communications. Proceedings. Eighteenth Annual Joint Conference of the IEEE Computer and Communications Societies. The Future is Now (Cat. No.99CH36320).

[12]  Erich M. Nahum,et al.  Locality-aware request distribution in cluster-based network servers , 1998, ASPLOS VIII.

[13]  Prashant J. Shenoy,et al.  A Demand Adaptive and Locality Aware (DALA) streaming media server cluster architecture , 2002, NOSSDAV '02.

[14]  Peter Parnes,et al.  Characterizing user access to videos on the World Wide Web , 1999, Electronic Imaging.

[15]  Michael B. Jones,et al.  The Tiger Video Fileserver , 1996 .

[16]  Olav Sandstå,et al.  Video server on an ATM connected cluster of workstations , 1997, Proceedings 17th International Conference of the Chilean Computer Science Society.

[17]  Ludmila Cherkasova,et al.  Measuring the capacity of a streaming media server in a Utility Data Center environment , 2002, MULTIMEDIA '02.