Code distribution and data fusion in networks: A case study using overlays

Sunny So, K. L.Eddie Law

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Instead of executing a program at a centralized server system, is it possible to slice a program and run different slices dynamically across the Internet? This is exactly what this paper will examine and present. Concept of programmable nodes was introduced years ago. Commercially, active network concept was not feasible to operate in routers. But with the arrival of overlays, extension of programmable nodes to geographically distributed server farms may become a reality. Network service providers, application program providers, and end users constitute a novel business model. In the paper, a selected program is sliced, partitioned, and dispersed across different network servers for computations. An NP-complete problem (integer knapsack algorithm) is sliced into multiple small modules to execute in a networked prototype for validation. It has been demonstrated that small computational modules on servers allow speedy computations and fast convergence of data at receivers.

Original languageEnglish
Title of host publicationIEEE Canadian Conference on Electrical and Computer Engineering, Proceedings, CCECE 2008
Pages1683-1686
Number of pages4
DOIs
Publication statusPublished - 2008
Externally publishedYes
EventIEEE Canadian Conference on Electrical and Computer Engineering, CCECE 2008 - Niagara Falls, ON, Canada
Duration: 4 May 20087 May 2008

Publication series

NameCanadian Conference on Electrical and Computer Engineering
ISSN (Print)0840-7789

Conference

ConferenceIEEE Canadian Conference on Electrical and Computer Engineering, CCECE 2008
Country/TerritoryCanada
CityNiagara Falls, ON
Period4/05/087/05/08

Keywords

  • Data fusion
  • Distributed computing
  • Diversity coding
  • Overlay networks
  • Program slicing
  • Programmable node

Fingerprint

Dive into the research topics of 'Code distribution and data fusion in networks: A case study using overlays'. Together they form a unique fingerprint.

Cite this