DARPA ISO Sponsored Research

2000 Project Summary
Computational Resiliency
Syracuse University

Project Website: 
Not yet available 
Quad Chart: 
Quad Chart - PPT Format
Any system that must operate in highly adverse environments, such as battlefield command and control, must be able to tolerate, recover from, and react to failures and attacks.

The standard fault-tolerant mechanisms to achieve this robust operation are based on some form of replications of critical information and resources. While this replication often provides graceful degradation of system performance, it is clearly not sufficient to aggressively recover assured operation, especially in the face of a coordinated attack.

The objective of this project is to develop technologies necessary for building distributed systems that are secure, remain available, and meet real-time performance constraints.

We are exploring techniques for constructing computationally resilient systems that deliver availability, security, and timeliness through a combination of techniques, including on-the-fly replication and reconfiguration, resource management, and camouflage.

Our approach is based on the metaphor of task threads as computational cockroaches infesting an apartment building (a distributed system). Roaches are notoriously resilient: no matter what form of attack, you never kill them all or completely prevent them from their goal of finding the resources they need to survive. To foil your eradication efforts, they use several techniques: they are highly mobile, moving from one place in the apartment (network) to another with speed and agility (thread migration). They continually reproduce to ensure that it is not possible to kill them all (replication). They flee from danger: if a light is turned on, they scurry away in all directions to hide behind cupboards in places of known safety (secure network zones). Our roaches will develop camouflage techniques to facilitate their spread.

We are pursuing three complementary efforts:

  1. Application-level and system-level software technologies intended to provide a concurrent-programming framework for resilient computing.

    These libraries provide basic replication, reconfiguration, and migration mechanisms; computational-camouflage support; resiliency specification frameworks; and process scheduleres capable of using the basic mechanisms to implement resiliency policy in response to intrusion events.

  2. Theoretical foundations to provide a framework to analyze and validate the protocols and policies underlying these technologies.

  3. Integration of existing DOD-related applications with our software to demonstrate its effectiveness in real-world situations.
Recent FY-2000 Accomplishments:
This project has just recently begun.
FY-2001 Plans:
Software Development: We will be implementing the essential thread-management primitives to migrate, merge and split threads; these primitive must properly track and maintain communication channels during migration and in the face of attack. We will also investigate simple camouflage and decoy deployments and initiate and initiate development of policy-specification toolkits and frameworks.

Formal Models: We will be identifying a core calculus and appropriate notions of equivalence/refinement to support reasoning about resource-constrained system behavior.

Integration: The primary integration task in FY-01 will be the port of a current DOD RADAR application to a distributed environment, in preparation for subsequent integration with CR software.

Technology Transition:
The computational artifacts of this effort will include system-level components for resourse management, thread management, scheduling, and reconfiguration. We expect that these artifacts will be built into commercial products, via Mobium Enterprises and the CASE Center located at Syracuse University.

The mission of the CASE Center is to develop a high-technology economy in New York State by fostering collaborations among the faculty and research projects at Syracuse University and New York State businesses. Mobium Enterprises is a small company incubated in the CASE Center that specializes in high-performance distributed computing for scientific and engineering applications and has already successfully deployed commercial products based on distributed-computing technologies.

We have already identified the following product areas of this project for commercialization:

  1. Secure real-time multispectral imaging and data dissemination
  2. Automated system reconfiguration in heterogeneous distributed computing systems
  3. Rule-based hierarchies for thread replication and memory management
Additionally, we will make the computational artifacts developed in this effort, as well as papers and reports describing them, accessible on the web.
Principal Investigators:
Steve J. Chapin
Dept of Electrical Engineering and Computer Science
2-177 CST
Syracuse University
Syracuse, NY 13244
(315) 443-4457 (voice)
(315) 443-1122 (FAX)
Susan Older
Dept of Electrical Engineering and Computer Science
2-177 CST
Syracuse University
Syracuse, NY 13244
(315) 443-4679 (voice)
(315) 443-1122 (FAX)

Admin Point of Contact:

Administrative Contact: Matthew Clark
Office of Sponsored Programs
Syracuse University
Syracuse, NY 13244
(315) 443-9356 (voice)
(315) 443-9361 (FAX)

Return to the ITS Project Summary Page  ]