[Colloquium] Ripeanu/Dissertation Defense/7-19-05

[Margaret Jaffey]

	Department of Computer Science/The University of Chicago

			*** Dissertation Defense ***

Candidate:  R. Matei Ripeanu

Date:  Tuesday, July 19, 2005

Time and Location: 10:00 a.m. in Ryerson 251

Title:  Using Peer-to-Peer Experience to Build Large-Scale Grid Services

Abstract:

	Understanding how to build self-organizing systems is paramount for
coping efficiently with the increased scale and complexity of the 
computing
systems we envision. Existing Grid infrastructure deployments provide 
the
real world requirements for such an exploration and can be used to 
evaluate
solutions in realistic settings.
	Multicast communication primitives have broad utility as building
blocks for distributed applications. However, creating and maintaining 
the
structures that support multicast primitives is challenging, 
particularly
when participating resources are volatile and their characteristics are
variable.
	We describe and evaluate UMM: a multi source multicast system based
on an unstructured overlay and a novel approach to extracting source
specific multicast trees. The strength of this solution comes from the
following two properties:  first, it decouples the overlay construction 
and
maintenance mechanism from the tree-extraction mechanism, allowing for
separate component optimization; second, it relies on soft-state and 
passive
data collection to adapt to the dynamics of the physical network, 
resulting
in low protocol complexity.
	Experimental and analytical evaluations demonstrate low
communication overhead, efficient network usage compared to alternative
solutions. In addition to these performance related characteristics UMM
displays four important properties. First, UMM is self organizing:
independent node decisions made based on partial, local information 
result
in desired global system behavior. UMM has the advantage of low 
complexity.
Second, UMM relies on soft-state and passive data collection to adapt to
participating end host and physical network dynamics. We assert that 
reduced
complexity is a highly desirable property of distributed, large scale
systems. Third, the state stored at each UMM node scales with the 
number of
sources and is independent of the number of passive participants. 
Finally,
UMM adapts to changes in underlying network topology characteristics,
performs well under significant levels of churn, and recovers quickly 
from
significant failures in the set of participating end-nodes with minimal
disruption to the multicast service offered.
	Additionally, to demonstrate UMM's practicality to implementing
higher level services, we evaluate it in an application context.  We 
design
and evaluate a replica location service that integrates three 
techniques:
UMM's flat overlay to obtain genuine decentralization and resilience 
when
facing network and node failures, probabilistic representations of 
replica
location information to achieve important space and bandwidth savings, 
and
soft-state protocols to decouple nodes state and achieve robustness.

Candidate's Advisor: Prof. Ian Foster

A draft copy of Mr. Ripeanu's dissertation is available in Ry 161A.


-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Margaret P. Jaffey				margaret at cs.uchicago.edu
Department of Computer Science
Student Support Rep (Ry 161A)		(773) 702-6011
The University of Chicago		http://www.cs.uchicago.edu
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-