Scalable Peer-to-Peer Monitoring: Design and Measurement

MPhil Thesis Defence


Title: "Scalable Peer-to-Peer Monitoring: Design and Measurement"

By

Mr. Chun-Ho Yuen


Abstract

In order to assess the overall service quality and compare different 
design alternatives, the performance metrics of a client application (such 
as a streaming session) need to be continuously fed back to some 
monitoring servers. In order not to overwhelm the monitors by the volume 
of feedback traffic, much of the previous work has used distributed 
aggregation tree (DAT). However, this approach often leads to an overlay 
with high monitoring delay and network stress.

In this thesis, we present and measure a scalable monitoring network for 
distributed applications. There are multiple monitors keeping global 
performance statistics. The network is a two-tier overlay, where client 
applications report their performance to a proxy by means of an overlay, 
and the proxies form another overlay spanning tree among themselves to 
report to the monitors.  Such network provides better isolation against 
peer churns and hence better monitoring stability.  We study how to 
construct such an effective monitoring network minimizing the aggregation 
delay.  We first formulate the problem and show that it is NP-hard. Then 
we propose an adaptive and scalable protocol called SMon, which 
continuously reduces the network diameter in the presence of peer churns. 
Through simulations and actual measurements, we show that SMon achieves 
low monitoring delay and good network performance for distributed 
applications.


Date:			Thursday, 19 August 2010

Time:			3:00pm – 5:00pm

Venue:			Room 4472
 			Lifts 25/26

Committee Members:	Dr. Gary Chan (Supervisor)
 			Dr. Lin Gu (Chairperson)
 			Dr. Ke Yi


**** ALL are Welcome ****