Main Profile

At A Glance

Statistical Virtualization: Scale as a Tool for Implementing Service Overlays

Google Tech TalksJanuary, 10 2008In this talk, we will explore the use of on-line non-parametric time seriesanalysis and prediction to build virtualized services in distributedcomputing environments. In particular, by analyzing and predicting thefuture behavior of one set of distributed services we explore how they can beamalgamated dynamically to implement a "virtual" service overlay withproperties that are not supported by any of the constituent servicecomponents (i.e. derive strictly from aggregation).We illustrate the approach by detailing distributed batch-scheduling mechanismsthat provide both reservation and co-allocation services in environments thatexplicitly do not support them. While our work focuses on national-scalescientific computing infrastructure, we believe its alternative approach tovirtualizing distributed systems abstractions is important in a largerscalable systems context for two reasons. First, because the methodology isinherently statistical, it improves with scale making scale a tool (ratherthan an impediment) in terms of implementation. Secondly, it shares manycommon features, both conceptually and implementationally with scalable searchservices making it possible, we believe, to explore the use ofcommercial search infrastructure in future work.Speaker: Rich WolskiMy origins, like those of most people born in North America during this century, are ambiguous and questionable. I am currently an Associate Professor in the Computer Science Department at the University of California, Santa Barbara which is, of course, located in Goleta California for all intents and purposes that do not involve the U.S. Mail. My past is checkered (it used to be plaid, but I've been politely informed that a past can only be so retro). Formerly, I enjoyed the hospitable climes offered by the Computer Science Department at the University of Tennessee. I've also done time as research faculty member in the U.C. San Diego CS&E Department where I researched CS and a little E (every now and then) in a decidedly pedagogical manner. My research interests include, but are not limited to, Computational Grid computing for performance, parallel and distributed systems, and the endless metaphysical search for the perfect coffee cup.
Length: 01:08:03

Contact

Questions about Statistical Virtualization: Scale as a Tool for Implementing Service Overlays

Want more info about Statistical Virtualization: Scale as a Tool for Implementing Service Overlays? Get free advice from education experts and Noodle community members.

  • Answer

Ask a New Question