A Decentralized Location and Routing
Infrastructure for Fault-tolerant Wide-area Applications

Ben Yanbin Zhao

Qualifying Examination Committee:
John D. Kubiatowicz (Chair)
Anthony D. Joseph
Satish Rao
John Chuang

Ph. D. Qualifying Examination
April 18, 2001

[Full Text of Proposal in PDF Format, 148KB]
[Talk Slides in PDF Format, 74KB]
[Talk Slides in Powerpoint SlideShow Format, 278KB]

Proposal Abstract

Today's Moore's Law increases in computational power and network bandwidth, combined with the increasing reach of networks into diverse environments and devices, offer new opportunities to and stretch the bounds of traditional network applications.  More specifically, new challenges of scalability, fault-tolerance and manageability stretch the limits of the communication components of applications. In this work, we propose the use of a global-scale routing and location infrastructure that leverages the abundant computational and network resources to facilitate a decentralized wide-area computing model.  We present an infrastructure prototype named Tapestry, and demonstrate its usefulness with novel large-scale network applications. Finally, we evaluate it by measuring its effectiveness against current challenges, and the resource costs these benefits incur in tradeoffs.