Scalable elastic systems architecture
Date
2012-04-15
DOI
Authors
Appavoo, Jonathan
Schatzberg, Dan
Version
OA Version
Citation
Appavoo, Jonathan; Schatzberg, Dan. "Scalable Elastic Systems Architecture", Technical Report BUCS-TR-2012-008, Computer Science Department, Boston University, April 15, 2012. [Available from: http://hdl.handle.net/2144/11396]
Abstract
Cloud computing has spurred the exploration and exploitation of elastic access to large scales of computing. To date the predominate building blocks by which elasticity has been exploited are applications and operating systems that are built around traditional computing infrastructure and programming models that are in-elastic or at best coarsely elastic. What would happen if application themselves could express and exploit elasticity in a fine grain fashion and this elasticity could be efficiently mapped to the scale and elasticity offered by modern cloud hardware systems? Would economic and market models that exploit elasticity pervade even the lowest levels? And would this enable greater efficiency both globally and individually? Would novel approaches to traditional problems such as quality of service arise? Would new applications be enabled both technically and economically? How to construct scalable and elastic software is an open challenge. Our work explores a systematic method for constructing and deploying such software. Building on several years of prior research, we will develop and evaluate a new cloud computing systems software architecture that addresses both scalability and elasticity. We explore a combination of a novel programming model and alternative operating systems structure. The goal of the architecture is to enable applications that inherently can scale up or down to react to changes in demand. We hypothesize that enabling such fine-grain elastic applications will open up new avenues for exploring both supply and demand elasticity across a broad range of research areas such as economic models, optimization, mechanism design, software engineering, networking and others.