|
PRIYA
NARASIMHAN
Assistant Professor
Electrical and Computer Engineering and Computer Science
 www
My research interests encompass all aspects of dependable distributed
systems. In particular, my research focuses on dependability for
distributed embedded and enterprise applications based on middleware,
such as EJB, Java RMI, CORBA, DCOM, Web Services and OSGi. My current
research projects include the development of distributed systems
and tools aimed at:
MEAD: Real-Time Fault-Tolerant Middleware. Today's middleware standards
make it possible to have either real-time or fault tolerance in isolation.
Supporting both real-time and fault tolerance is challenging because
the two QoS properties often impose conflicting requirements on the
system. This research focuses on strategies (i) to reconcile these
conflicts, (ii) to identify the real-time vs. fault tolerance trade-offs,
(iii) to continue to deliver adequate real-time and fault tolerance,
and (iv) to provide proactive, rather than the classical reactive,
fault-tolerance, even in the face of faults and changing resources.
Please see more details on the MEAD project at http://www.ece.cmu.edu/~mead
Starfish: Adaptive Survivable Infrastructures. Systems are increasingly
being connected to the Internet, which is an inherently unreliable,
insecure communication medium that is susceptible to network partitioning
and disconnected operation. This is aggravated further when an application
straddles multiple different enterprises, each with its own security/reliability
strategies and guarantees. This research focuses on developing adaptive
survivable infrastructures to support Internet-based applications
that must continue to operate in the face of faults, malicious attacks,
loss of resources, software upgrades, different user/usage profiles
and different applications. Please see more details on the Starfish
project at http://www.ece.cmu.edu/~starfish
As natural extensions of these research areas, I am also interested
in survivability for wireless systems, metrics to quantify the dependability
of systems, and mechanisms for proactive fault tolerance and recovery.
I have designed and implemented systems, such as Eternal and Immune,
to provide transparent reliability (protection against crash and
communication faults) and survivability (additional protection in
the face of malicious subversion of parts of the system), respectively,
to CORBA and Java applications. My experience with developing these
systems led to my participation in establishing the Fault Tolerant
CORBA industrial standard. As the CTO and VP of Engineering of a
startup company, I have had the additional opportunity to witness
the results of my research have a commercial impact.
|
