Xin
Liu
(858) 405-2311
http://www-cse.ucsd.edu/~xinliu
OBJECTIVE
An industry research position in the fields of network, distributed
computing, and simulation.
RESEARCH
·
Networking
Performance Modeling and Simulation
·
Parallel
and Distributed Computing
·
Parallel
and Distributed Discrete Event Simulation
·
High
Speed Cluster Communication
EDUCATION
University of California San Diego San Diego , CA
Ph.D. candidate in Computer
Science Sept.
1998-Aug. 2004(expected)
·
Thesis:
Scalable Online Network Simulation for Modeling Grid Dynamics
·
Advisor:
Prof. Andrew A. Chien
Institute of Computing Technology
Master of Engineering - Computer Engineering
Sept.
1995 - July 1998
·
Thesis: Single Entry Point of Cluster Systems
·
Advisor: Prof. Jianping Fan
Tsinghua University Beijing , China
Bachelor of Science - Computer Science Sept. 1990 - July 1995
EXPERIENCE
·
One
of the major designers and lead developer of the MicroGrid
project. In the project, we are developing a set of scalable simulation tools and
models, which enable high fidelity simulations of Grid applications,
middleware, and resources and can be used as a scientific model for a Grid. We
can use these simulation tools to develop stable resource management, fault
resilience, and quality-of-service protocols over Internet.
·
The
lead designer and developer of Scalable
Online Network Simulator MaSSF, which is the key component of MicroGrid
system. Built above distributed discrete event
simulation, the scalability of network simulation can be improved by using
advanced load balance mechanisms and by exploiting the network topology. MaSSF
scales well up to 50,000 router network simulation using 256 cluster nodes.
Institute of Computing Technology, Research Assistant
·
The
lead designer and developer of the Single
Entry Point (SEP) for Dawning-2000 Supercluster
system, which is a cluster of 64 nodes of PowerPC/AIX box and the fastest
cluster in
Dawning Information Technology Company, Software Engineer (part-time)
·
Take
part in the design and implementation of a Highly
Available Dual-host Hot-standby System. The backup server can monitor the
status of the primary server and all applications running on it. When the
primary fails, the backup can automatically switch to be the primary server and
seamlessly accept following application requests. This was a really successful commercial
product at that time.
·
Performance
evaluation and analysis of a cluster system: High Performance Virtual Machine (HPVM).
The goal is to find out the bottleneck of the system, and find out some
heuristic programming rules to achieve best performance on this system.
·
Porting
of Fast Messages (a high speed cluster communication protocol)
from Myrinet to GigaNet.
And we evaluate the performance and scalability of GigaNet
system. It is implemented on a NT cluster.
Hewlett Packard Research Laboratory, Summer Intern
·
Design
a Cluster File System (CFS) on
storage area network at Multi-Computer Systems group of HP Lab. The motivation
is to utilize the increasing computing capacity on the disk controllers and
high speed switch network available for cluster system. This is done by
re-partitioning the functions of file system and moving some operations close
to the data to reduce the network traffic.
·
Teaching Assistant for three undergraduate courses
on Computer Architecture, Network Communication, and Discrete Mathematics
CONFERENCE PUBLICATION
·
Realistic
Large Scale Online Network Simulation,
Xin Liu and Andrew Chien,
submitted to in Proceedings of the ACM Conference on High Performance Computing
and Networking, SC2004,
·
Traffic-based Load Balance for Scalable Network
Emulation, Xin Liu and Andrew Chien, in Proceedings of
the ACM Conference on High Performance Computing and Networking, SC2003,
·
The MicroGrid: a Scientific Tool for Modeling Computational Grids ,
H. Song, Xin Liu, Dennis Jakobsen,
Ranjita Bhagwan, Xianan Zhang, Kenjiro Taura and Andrew Chien in Proceedings of SC2000, Dallas, Texas, November 2000
·
New Grid Scheduling and Rescheduling Methods in the GrADS Project,
(many authors including Alex Olugbile, Huaxia Xia, Xin Liu and Andrew Chien), in the Workshop
for Next Generation Software, Santa Fe, New Mexico, April 2004, in conjunction
with the IEEE International Parallel and Distributed Processing Symposium 2004.
JOURNAL
PUBLICATION
·
Validating
and Scaling the MicroGrid: A Scientific Instrument for
Grid Dynamics, Xin Liu, Huaxia Xia,
and Andrew Chien, submitted to Journal of Grid
Computing.
·
The MicroGrid: a Scientific Tool
for Modeling Computational Grids , Song, Xin Liu, Dennis Jakobsen, Ranjita Bhagwan, Xianan Zhang, Kenjiro Taura and Andrew Chien.
Scientific Programming, Vol. 8, Num 3, 2000, pp. 127-141.
TECHNICAL SKILLS
·
Specialist
at TCP/IP Socket programming and MPI parallel/distributed computing
·
Specialist
at IP protocols, including TCP/UDP, OSPF, and BGP4 protocols
·
Deep
understanding of Linux, from system administrate to system/kernel programming
·
Experience
on large software design, development, and management with team cooperation
·
Programming
languages: C/C++, Java, Perl, and MPI
HONOR
·
Best
paper finalist, SuperComputing 2000
MEMBERSHIP
·
IEEE
Computer Society
·
ACM
SIGCOMM, SIGMETRICS, SIGARCH
REFERENCES
Letters of Reference File is
available upon request from:
Andrew A. Chien, Professor Amin M. Vahdat, Associate
Professor
9500 Gilman Drive, Dept. 0114 9500
Gilman Drive, Dept. 0114
La Jolla, CA 92093-0114 La
Jolla, CA 92093-0114
Email: achien@ucsd.edu Email:
vahdat@cs.ucsd.edu
Phone: 858-822-2458 Phone:
858-534-4614
George Varghese, Professor
Email: varghese@cs.ucsd.edu
Phone: 858-822-0424