PhD Student
Mailing address:
University of California, San Diego
Department of Computer Science and Engineering
9500 Gilman Dr., dept. code 0404
EBU 3B 3140
La Jolla, CA 92093-0404
e-mail: flavioATcs.ucsd.edu
Phone number: +1 858 822-5647
|
|
I have graduated!
- BSc in Electrical Engineering, Federal University of Rio de Janeiro (UFRJ), 1997.
- MSc in Electrical Engineering, COPPE/UFRJ,
1999.
- PhD in Computer Science, University of
California, San
Diego. In progress.
My research interests include:
- Dependable systems
- Distributed algorithms
- Internet applications
- Network protocols
- Operating systems
My first contact with academic research was back at
UFRJ,
when I was still an undergrad. In 1993, Prof. Otto Duarte offered me
this
opportunity to work along with grad students in the computer networking
group ( GTA ), which he lead at
that
time. The project I was involved more closely consisted of the
development
of transport protocols for multimedia applications. There were also
other
projects at that time related to the design of reliable multicast
transport
protocols and mechanisms for admission control in ATM networks.
The
tasks given to me in the context of the multimedia project consisted in
writing code for applications and for specific parts of the prototypes
under construction. Some years later, I was a master student in this
same
group and my advisor was still Prof. Otto Duarte. The topic of my
research
was the utilization of active networks to improve services for mobile
devices.
Briefly, the idea I developed in my master thesis was to use cache in
network
nodes to cope with message losses due to both the frequent handoffs of
mobile hosts and lossy wireless links for single-sender multicast
applications.
To evaluate the efficiency of the mechanisms I proposed, I did several
simulations using the network simulator ns-2. A product of my
master
thesis was an extension to ns that implements the mechanism
proposed
by Tennenhouse et al. in the ANTS
(Active Node Transfer System) project.
When I came to UCSD, I was very interested in mobile
computing.
During the summer of 2000, while advised by professors Keith Marzullo
and
Geoffrey M. Voelker, I started reading papers on mobile computing.
Although
I was mostly surveying what was being done in terms of mobile
computing,
I had a special interest for ad-hoc routing protocols. By the winter of
2001, Keith Marzullo, Geoffrey Voelker, and Stefan Savage started the
project
called RAMP (Reliable Adaptive
Multi-Path)
networks, which focus on the development of mechanisms to make
wide-area
and ad-hoc networks resilient to malicious attacks. I automatically
joined
the project since my advisors were involved in it and I was already
working
on a related topic.
I'm currently working on the problem of designing
distributed
systems for dependent failures. On the theoretical side, I've been
looking at the impact of assuming
dependent process failures in the design of fault-tolerant distributed
algorithms. In particular,
I've proved requirements on process replications for the Consensus
problem in different system models.
I'm currently working on the generalization of these results. On the
practical side, I've been working on a project called Phoenix, which is
a backup system designed to cope
with Internet catastrophes such as worms and viruses. The main idea is
that the threshold model
of failures is not adequate for such an application, because the
the maximum number
of hosts that can fail is potentially very high (over 300,000 hosts
were infected by the Code Red worm
according to a CAIDA analysis). Thus, a model of dependent failures is
clearly necessary in this case.
More details can be found on the technical reports below.
Present
Past
Workshops and Conferences
[HPDC'06] Zhang, X., Junqueira, F. P., Hiltunen, M., Marzullo, K., Schlichting, R., "Replicating nondeterministic services on grid environments". In the Proceedings of 15th International Symposium on High Performance Distributed Computing (HPDC-15), Paris, France, June 2006;
[DISC'05] Junqueira, F. P., and Marzullo, K., "Coterie availability in sites". In the Proceedings of 19th International Symposium on Distributed Computing (DISC'05), Krakow, Poland, September 2005;
[EuroPar'05] Junqueira, F. P., and Marzullo, K., " Replication predicates for dependent-failure algorithms ". In the Proceedings of 11th International Conference on Parallel and Distributed Computing (Euro-Par'05), Lisbon, Portugal, August 2005; (distinguished paper)
[HotDep'05] Junqueira, F. P., and Marzullo, K., " The virtue of dependent failures in multi-site systems". In the Proceedings of IEEE Workshop on Hot Topics in System Dependability (HotDep'05), Yokohama, Japan, June 2005;
[USENIX'05] Junqueira, F. P., Bhagwan, R., Hevia, A., Marzullo, K., and Voelker, G. M., "Surviving Internet Catastrophes". In the Proceedings of USENIX Annual Technical Conference, Anaheim, CA, USA, April 2005;
[HotOS'03] Junqueira, F. P., Bhagwan, R., Marzullo, K., Savage, S., and Voelker, G. M., " The Phoenix Recovery System: Recovering from the ashes of an Internet catastrophe ", in the Proccedings of HotOS IX,Lihue, HI, USA, May 2003;
[ICDCS' 03] Junqueira, F. P., Marzullo, K., " Synchronous Consensus for Dependent Process Failures ". In the Proceedings of the 23rd IEEE International Conference on Distributed Computing Systems (ICDCS'03), Providence, RI, USA, May 2003;
[FuDiCo' 02] Junqueira, F. P., Marzullo, K., " Designing Algorithms for Dependent Process Failures ". In the Proceedings of International Workshop on Future Directions in Distributed Computing, LNCS , Bertinoro, Italy, 2002;
[SBC'99] Junqueira, F. P. and Duarte, O. C. M. B., "Active Reliable Multicast with Adaptive Caching". In the proceeding of the XIX National Conference of the Brazilian Computer Society SBC'99, XXVI Integrated Symposium on Software and Hardware SEMISH'99, vol. 1, pp. 73-87, Rio de Janeiro, Brazil, July 1999;
[ICMCS' 99] de Lima, R. M., Junqueira, F. P., Gonçalves, P. A. da S. and Duarte, O. C. M. B., "SAMM: An Integrated Environment to Support Multimedia Synchronization of Pre-orchestrated Data". In the Proceedings of the IEEE International Conference on Multimedia Computing and Systems ICMCS'99, Vol. 2, pp. 929-933, Firenze, Italy, June 1999;
[SBRC' 99] Junqueira, F. P. and Duarte, O. C. M. B., "Active Networking Support for Mobile Multicast Communication". In the Proceedings of the XVII Brazilian Computer Networking Symposium SBRC'99, pp. 389-404, Salvador, Brazil, May 1999;
[SFBSID' 97] Junqueira F., Kosmalski L., Duarte O., "A QoS-based Routing Architecture using Active Networks", in proceedings of the 2nd SFBSID'97, Fortaleza, Brazil, November 1997;
Tech reports
[TR-WLE] Junqueira, F. P., Marzullo, K., " Weak Leader Election in the receive-omission failure model", UCSD Tech Report CS2005-0828, May 2005. Minor revision in August 2005;
[TR-Gen] Junqueira, F. P., and Marzullo, K., " On the Generalization of n>k*t ", Tech Report, April 2003;
[TR-Consensus] Junqueira, F. P., Marzullo, K., " Consensus for Dependent Process Failures ", UCSD Tech Report CS2003-0737, September 2002;
[TR-LB] Junqueira, F. P., Marzullo, K., " Lower Bound on the Number of Rounds for Synchronous Consensus with Dependent Process Failures ", UCSD Tech Report CS2003-0734, September 2002;
[TR-DF] Junqueira, F. P., Marzullo, K., Voelker, G. M., " Coping with Dependent Process Failures ", UCSD Tech Report CS2002-0723, December 2001;
[TR-Con] Junqueira, F. P., Teixeira, R. C., "Connectivity in the South American Internet ", UCSD Tech Report CS2003-0733,
March 2001;
Thesis and BSc Senior Project
[MSc-Thesis] Junqueira, F. P., "Reliable
Multicast Communication
for Mobile
Hosts in Active Networks", Master thesis presented to COPPE/UFRJ as
partial
fulfilment of the requirements for the degree of M.Sc. in electrical
engineering,
Rio de Janeiro, Brazil, August 1999;
[BSc-Project] Junqueira F., "XOCPN Automatic
Generation for a
Multimedia
Synchronization Service". Final project presented to DEL/UFRJ as
partial
fulfillment of the requirements for the degree of bachelor in
electrical
engineering, February 1997;
[DISC-Talk] Coterie availability in sites, DISC, Krakow, Poland, September 2005.
[Europar-Talk] Replication predicates for dependent-failure algorithms, Euro-Par, Lisbon, Portugal, August 2005.
[HotDep-Talk] The virtue of dependent failures in multi-site systems, HotDep, Yokohama, Japan, June 2005.
[USENIX-Talk] Surviving Internet Catastrophes, USENIX Annual Technical Conference, Anaheim, April 2005.
[HP-Talk] Surviving Internet Catastrophes, HP Labs, Palo Alto, December 2004.
[OSDI-WIP] Surviving Internet Catastrophes, OSDI Work-in-progress session, San Francisco, December 2004.
[MSR] Moving away from the Independent and Identically Distributed Failure Assumption, Microsoft Research, Cambridge, UK, September 2003.
[HotOS] The Phoenix Recovery System: Recovering from the Ashes of an Internet Catastrophe, HotOS'03, Lihue, Kauai, May 2003.
last modified:
03/23/2006
