Why elasticity matters
Date
2012-04-15
DOI
Authors
Schatzberg, Dan
Appavoo, Jonathan
Krieger, Orran
VanHensbergen, Eric
Version
OA Version
Citation
Schatzberg, Dan; Appavoo, Jonathan; Krieger, Orran; VanHensbergen, Eric. "Why Elasticity Matters", Technical Report BUCS-TR-2012-006, Computer Science Department, Boston University, April 15, 2012. [Available from: http://hdl.handle.net/2144/11394]
Abstract
In this paper we proposed a new research agenda focused on elasticity. We argued that elasticity is an important area of research and hypothesized that research in this area will lead to more efficient systems with less hoarding, new applications that exploit massive cloud resources elastically, and system software and libraries that will simplify the task of developing elastic applications.
We discussed some of our thoughts on a top-to-bottom cloud-scale system focused on elasticity. We argued that such a system will require: 1) a HW/IaaS layer that can quickly reallocated resources to different applications, 2) an event driven model where resource demand flows from the high level layers as transparently as possible to the lowest level of the system, and 3) a model of modularity that allows layers to be overridden as necessary and provides applications with a component model that enables the base elasticity to be exploited by new and advanced applications.