Our research efforts on the PERC project have had significant impact in the following areas:
1. Fault
tolerance: We have been
involved in the commercialization of fault tolerance technology through a
start-up company.
2.
Exploitation
of LRD in real-time resource management: Exploitation of LRD in congestion, resource, and traffic control
is still in its infancy. Tuan and Park
pioneered the work of exploiting LRD in congestion control, and used, based on
the conditional expectation, a simple estimation scheme to explore the
correlation structure. We have taken the heuristic rule-of-thumb to a rigorous
plateau, where the traffic prediction is rigorously made with the use of a
LMMSE predictor, and the calculation of packet dropping probability in AQM
and/or the window adjustment in TCP congestion control is pinpointed in the
context of steady-state dynamics. Moreover, this is achieved without requiring
router support or compromising simplicity and ease of implementation. The research results have been (or will be)
reported in IEEE INFOCOM 2002
(acceptance rate < 20%) and IEEE Int’l
Conf. on Network Protocols 2002 (acceptance rate < 15%).
3.
Centrality analysis within specific routing
protocols: We believe that the results of this research should lead to an
understanding of the impact and limitations of centrality in computer networks,
and the development of extended, secure, and robust routing protocols.
Specifically,
· We
believe that capturing the changing centrality description of the routing
topology will enable detection of some large scale network wide routing
attacks, such as may be wrought by compromised routers. We believe that this
detection can occur early, even before the changed forwarding tables are in
place and data packet forwarding occurs. A goal of centrality-based intrusion
monitoring is to abstract global network behavior locally at a router.
Subverting such monitoring, while causing a network wide attack, is harder
because of this abstraction. Given the nature of the information being
abstracted, centrality-based monitoring might not detect attacks where the
compromised routers are selectively misrouting packets. However, such attacks
would typically not have a disruptive effect on the network.
· We believe that this study will
suggest abstract specifications of router behavior that are both monitorable by
individual routers and which capture expected behavior of routers in a given
protocol; these specifications can detect rogue router behavior without
requiring prior knowledge of a router compromise.
4.
QoS-driven
multicast routing protocols: The research results on QoS-driven multicast routing have direct
applicability to the development and implementation of network
protocols/solutions in support of resource management in network-centric
warfare scenarios. Three journal papers
that document the research findings have been accepted for publication in two
of the most prestigious IEEE journals: IEEE
Trans. on Computers and IEEE/ACM
Trans. on Networking. Several
student-coauthored papers have been presented in highly-referenced IEEE conferences
(usually with acceptance rate usually less than 20%), e.g., IEEE INFOCOM and IEEE Int’l Conf. on Network Protocols.
5. Scalable, real-time video multicast
protocol: NS Solutions Corporation, one of the largest system integrators
in Japan, is interested in marketing the scalable, real-time video multicast
protocol that is developed under the support of AFOSR/MURI.
6. Transfer
of wireless technologies to U.S. forces: The results from the research carried out in the PERC have direct
applicability to the development and implementation of much needed solutions in
support of wireless internetworking solutions that are critical for a
full-spectrum dominance in the battlefield by the U.S. forces. Lessons learned in the Task Force XXI
experiments indicate that using COTS Internetoworking solutions will not
suffice to satisfy the communication needs of military networks. We are sharing
many of the results developed with support from the PERC with Raytheon, which
is applying the results to tactical wireless networks with multi-beam antennas
as part of the DARPA FCS program.
7.
Mitigating
misbehavior in routing protocols in MANETs: Our work on mitigating misbehavior in routing protocols in ad hoc
networks has generated a great deal of interest and follow-on work. Our experiments with Watchdog and Pathrater
were published in Mobicom 2000 (the premiere forum for research in mobile
networks). Since then, other
researchers such as Blazevic et al.
and Buchegger et al. have
experimented with other approaches involving heavier-weight security
mechanisms.
8. Packet
scheduling algorithms in MANETs: Our work on packet scheduling algorithms was presented as a
poster at MobiHoc 2002 in Switzerland.
It was selected as one of the best posters of the conference and thus we
were invited to publish the corresponding paper in ACM Mobile Computing and Communications Review (MC2R).
9. Standardization of Real-time CORBA: The
OMG Real-time CORBA specification defines standard interfaces and policies that
allow distributed real-time applications to configure and control the following
system resources:
·
Processor resources via thread pools, priority mechanisms,
intra-process mutexes, and a global scheduling service for real-time
applications with fixed priorities
·
Communication resources via protocol properties and explicit
bindings to server objects using priority bands and private connections and
·
Memory resources via buffering requests in queues and bounding
the size of thread pools.
Real-time
CORBA is designed for applications with hard real-time requirements, such as
avionics mission computing, as well as those stringent soft real-time
requirements, such as telecommunication call processing. Our work on real-time
middleware significantly influenced the OMG Real-time CORBA specification.
10. Standardization of extensible transports in
CORBA: Pluggable protocol capabilities are being standardized by the OMG in
the Extensible Transport Framework (ETF) specification effort. Although the OMG ETF has many commonalities
to TAO's pluggable protocols framework, ETF only standardizes the interface to
install transport protocols. Thus, GIOP
remains the sole ORB messaging protocol standard and is not itself pluggable in
the ETF specification. We expect that our work on PERC will enable future OMG
specifications to standardize pluggable high-level messaging protocols, as well
as the lower-level transport protocols.
11. Widespread adoption of Real-time CORBA in
the DoD and commercial industry: In addition to guiding the standardization
of Real-time CORBA and extensible transports in CORBA, the most tangible impact
of our work funded in part by PERC is the TAO ORB itself. TAO has been used in
hundreds of production software systems in research labs, military
applications, and commercial projects, including avionics mission computing
systems at Boeing, satellite communication systems at Lockheed and Raytheon,
telecommunication systems at Alcatel, BBN, Cisco, Lucent, Motorola, Nortel, and
Siemens, and medical systems at GE.
Since TAO is open-source software, the web site www.cs.wustl.edu/~schmidt/TAO.html
contains complete source code and documentation for TAO.
12. Bio-networking architecture: BBC News
featured an article that overviews the motivation and vision of the
Bio-Networking Architecture. In this article, adaptive object behaviors and
network applications in the Bio-Networking Architecture are described. Also,
key protocol designs in the Bio-Networking Architecture are currently being
discussed as a possible reference architecture at the Super Distributed Objects
group of the Object Management Group (OMG), the largest standard making bodies
for object oriented software technologies.
13. Bio-networking
architecture: One of the papers on the Bio-Networking Architecture and its
associated protocols received the best paper award at the 2002 IEEE SAINT (Symposium on Applications and the
Internet) conference. This is an IEEE conference specializing in Internet
applications with ten technical paper sessions, three panel sessions and 4
workshops.
14. Narses: Even though it is in its early
stages of development, the Narses
simulator has already been used by three research projects, including one
research group entirely unassociated with its authors. One example of its use may be seen in the
work on CUP, a controlled update protocol for propagating cached indices in
large peer-to-peer networks. We have
also been asked by the LOCKSS project to provide simulation support via Narses.
The goal of the LOCKSS project is to provide a robust peer-to-peer
network of digital libraries around the world.
Experiments comparing Narses to ns-2 are under submission to a
conference.
15. JavaSim software
release:
JavaSim provides
an extensible, reusable network simulation and emulation environments. Since its formal release, more than one
hundred groups have downloaded the software, including NIST, Oak Ridge National
Laboratory, Fujitsu Labs of America, CMU, Univ. of Toronto, Dartmouth College,
Georgia Tech, and Renesys, Inc. We are
currently collaborating with Oak Ridge National Laboratory on laying a Grid
networking layer (on top of the transport layer) in JavaSim to support simulation of Grid applications and with Fujitsu
Labs of America on extending JavaSim
for network emulation.
Education and Student Training: The research being carried in the
PERC project is part of the thesis research of twelve Ph.D. students:
1. Lichun
Bao has
successfully advanced to candidacy at University of California at Santa Cruz
and is expected to graduate during the Fall 2002 quarter. He will present papers at ACM Mobicom 2002 and IEEE ICNP 2002.
2. Mayur
Deshpande has
passed his candidacy exam at University of California at Irvine and expects to
graduate in late 2003 with an emphasis on protocols and algorithms for
reflective middleware that can run efficiently and scaleably on parallel
processors.
3. Yuan
Gao has
successfully advanced to candidacy at Ohio State University, is expected to
graduate in September 2002, and will join Lucent Technologies Bell
Laboratories, Murray Hills in September 2002 as a member of technical staff in
traffic measurement and control.
4. Guanghui
He is expected to
take his candidacy exam at University of Illinois at Urbana Champaign in
September 2002, and is working as a summer intern student (in Summer 2002) at
IBM T.J. Watson Research Center in the area of studying web traffic patterns
and user behaviors.
5. Ruppert Koch graduated from University
of California at Santa Barbara, and is currently a postdoctoral researcher at
the same University, while working for a start-up company.
6. Priya Narasimhan graduated from
University of California at Santa Barbara, was supported as a postdoctoral
researcher on this MURI project, and is now an assistant professor at Carnegie
Mellon University.
7. Carlos
O’Ryan has
successfully defended his thesis, is expected to graduate from University of
California at Irvine in Fall of 2002 and has joined Automated Trading Desk in
Charleston, SC, which specializes in the use of distributed real-time
middleware for online stock trading.
8. Irfan
Pyarali graduated
from University of California at Irvine in December 2001 and is currently the
founder of the OOMWorks startup company in Metuchen, NJ, which
specializes in distributed real-time middleware.
9. Soumya
Roy has
successfully advanced to candidacy at University of California at Santa Cruz
and is expected to graduate during the Fall 2002 quarter. He will present papers in IEEE SAWN 2002 and IEEE ICCCN 2002.
10. Hung-Ying
Tyan graduated
from Ohio State University in Fall 2001 and is currently a member of technical
staff at Fujitsu Labs of America, Sunnyvale, CA.
11. Yu
Wang is currently
a pre-Ph.D. candidate at University of California at Santa Cruz. He will present papers in IEEE SAWN 2002 and IEEE/ACM MASCOTS 2002.
12. Wenbing Zhao graduated from University
of California at Santa Barbara, and is currently a postdoctoral researcher at
the same University, while working for a start-up company.
Other Professional Activities: Recognized by the research
accomplishments (which are, in part, results of the grant), all the PIs on the
project have been actively involved in professional services:
1. Mary
Baker was elected
by the membership to be vice chair of the ACM
Special Interest Group on Mobility (SIGMOBILE). In addition, she is
co-chair of the technical program committee of the new annual ACM/USENIX
conference on higher-level issues in mobile systems. The first MobiSys
conference will take place in May 2003 and is already shaping up to be a
premier forum for publishing mobile systems work. http://www.usenix.org/events/mobisys03/.
She has also served on many program committees and editorial boards in the area
of mobility and is also one of the founding editorial board members of the new
IEEE Pervasive Computing magazine.
2. J .J.
Garcia-Luna-Aceves
was Program Co-Chair of the ACM MobiHoc 2002 Conference, held in Lausanne,
Switzerland on June 9--11, 2002.
3. Jennifer
Hou has served as
the chair of the technical program committee of IEEE Real-time Technology and Application Symposium in June 2000
and as the general chair of the same symposium in June 2001. In addition, she has served as the vice
chair of the technical program committee of
IEEE Int’l Conf. on Distributed Computing Systems in June 2002, the
tutorial chair of IEEE Int’l Conf. on
Computer, Communication, and Networks in November 2002, and the session
chair for IEEE
8th Int'l Conf. on Network Protocols, November 2000, SPIE conference
on scalability and traffic control in IP networks, August 2001, and IEEE
21st INFOCOM: The Conference on Computer Communication, June 2002. She has also served on the review panel for NSF CAREER program in 1998-2001, NSF Information Technology Research
program in 2000-2002, NSF Strategic Technology
for Internet Program in 2001, NSF
Computer and Computational Research program in 2001, and NSF Embedded and Hybrid Systems Program
in 2002.
4. Douglas
C. Schmidt has
served (or will serve) as the program co-chair for IEEE Real-time
Technology and Application Symposium June 2003, the program co-chair for the IFIP/ACM/USENIX Middleware 2003, 4th
International Conference on Distributed Systems Platforms June 2003, Area
vice-chair and session chair for Middleware at the 23rd IEEE International Conference on Distributed Computing Systems
(ICDCS) May 2003, Program co-chair for the 4th International Symposium on Distributed Objects and Applications
October 2002, Program Co-chair for the 3rd
International Symposium on Distributed Objects and Applications September
2001, Area vice-chair and session chair for Middleware at the IEEE International Conference on Distributed
Computing Systems (ICDCS) April 2001, and co-chair of the OMG Workshop on Real-time and Embedded CORBA,
in Reston, VA, July 2000.