Enabling High-Impact Applications in Large-Scale Systems in SMART Ways

                            Professor Xiaolin Li
                       Department of Computer Science
                          Oklahoma State University

                           March 5, 2010 (Friday)
                    10:00AM  Donuts; 10:15 -11:30 AM Talk
                    First floor auditorium, EC&S Building

Ever-increasing computing and storage resources provide great opportunities
to tackle grand challenge problems in science and engineering. However, it
is very challenging for scientists and engineers to efficiently and smartly
use these vast resources. Although eager to spearhead scientific discovery,
they are handicapped by similarly challenging problems in parallelization,
load balancing, runtime management, and complexities in large-scale systems
and applications.

This talk will present a SMART framework to address these issues from self
managing systems and algorithms perspectives. SMART is an integrated
framework of scalable adaptive runtime management algorithms, libraries,
and toolkit with friendly programming models. Following SMART ways,
scientists can write sequential programs to achieve automatic parallelism
and high performance and throughput without meticulous attention to handle
thousands of processors and petabytes of storage. The SMART toolkit
features a suite of components and abstractions to capture application
runtime states and manage their data movement. It is further enhanced with
a suite of application-aware self-managing algorithms to holistically
address various issues in computation, communication, data, I/O, and
asynchronism in systems with thousands of processors (such as clusters,
grids, and clouds). Some preliminary results and ongoing work for
scientific and bioinformatics applications will be presented too.

Other ongoing projects at S3Lab will be conceptually overviewed, such as
OKGems (enabling remotely programmable virtual laboratory of multimodal
cyber-physical systems on the future Internet), YouGo (physical search
engine via cyber-physical social networks), SmartGuard (firefighter
protective clothing and falling detection for elders), and CoopNet/CloudBay
(computational ebay).

Bio:
Xiaolin (Andy) Li is currently an Assistant Professor and the director of
Scalable Software Systems Laboratory in Computer Science Department at
Oklahoma State University. His research interests include parallel and
distributed systems, sensor networks, and cyber-physical systems. His
research has been sponsored by National Science Foundation (CAREER,
PetaApps, GENI, CRI, and MRI), other federal and state funding agencies,
industry partners, and OSU. He is in the executive committee of IEEE
Technical Committee of Scalable Computing (TCSC) and the coordinator of
Sensor Networks. He received the PhD in Computer Engineering from Rutgers
University. He is a member of IEEE and ACM.