9th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing

Research Article

Autonomic Swarms for Regenerative and Collaborative Networking

Download355 downloads
  • @INPROCEEDINGS{10.4108/icst.collaboratecom.2013.254108,
        author={Rui-ping Lua and Wee Keong Ng},
        title={Autonomic Swarms for Regenerative and Collaborative Networking},
        proceedings={9th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing},
        publisher={ICST},
        proceedings_a={COLLABORATECOM},
        year={2013},
        month={11},
        keywords={autonomic computing collaborative networking self-configuration self-optimization swarm networks overlay networks fast flux service networks},
        doi={10.4108/icst.collaboratecom.2013.254108}
    }
    
  • Rui-ping Lua
    Wee Keong Ng
    Year: 2013
    Autonomic Swarms for Regenerative and Collaborative Networking
    COLLABORATECOM
    IEEE
    DOI: 10.4108/icst.collaboratecom.2013.254108
Rui-ping Lua1,*, Wee Keong Ng1
  • 1: Nanyang Technological University
*Contact email: ruiping.lua@gmail.com

Abstract

The exponential growth of web services have necessitate the evolution of network infrastructures to meet this challenge. We envision the myriad of Internet connected devices coming together to provide a robust and reliable service network. We propose an Autonomous Swarm Network to provide autonomic capabilities to achieve our service quality goals while coping with complex and changing requirements of today’s web services particularly cost-effectiveness versus service assurance. To create a high-resilient network, we incorporated features of self-management, self-configuration, self-optimization and selfhealing strategies. Using a combination of fast-flux service networks, autonomic management and swarm algorithms, it becomes possible to build cost effective assurance for existing web services. We demonstrate the feasibility of our solution using the Nanyang Analytics Supercomputer with more than 20,000 agents against varying loads. We’ve also simulated algorithms and reconfiguration strategies. We subsequently developed a prototype swarm network of up to 500 machines.