Wireless ad hoc networking has become a critical technology for nodes to communicate with each other in the absence of infrastructures such as base stations. It will be used in a variety of applications including wireless sensor networks, disaster recovery, and military on-the-field communications. Significant advances have been made in the area of wireless ad hoc networking at all levels of the protocol stack, MAC, routing, congestion control, etc., with an eye toward configuring networks so as to optimize metrics such as capacity, connectivity, delay, etc. However, there remain numerous issues that have not been satisfactorily dealt with, ranging from identifying network configurations to optimize emerging security metrics, to how to effect various desirable network configurations: namely, how does an ad hoc network bootstrap itself to realize the promised performance and security? We address this question in our research. Specifically we focus on the following:

? Wireless network security: Characterize the tradeoff between performance and security in opportunistic wireless networks where eavesdroppers and/or jammers are present, and develop algorithms for providing secure communication in such environments.

? Wireless Network configuration: Configure networks at the time of deployment, paying particular attention to facilitating the selection of "friendly jammers" and opportunistic relays as needed to enhance security.

The outcome of this research will constitute a significant advance in the development of theoretical foundations, practical algorithms, and network architecture for configuring and securing wireless networks.

Project Report

Securing information networks is of significant importance in modern society. This need has been exacerbated as users switch to mobile communication devices for access, as a communication signal sent wirelessly through the air can be eavesdropped by any adversary in proximity of the transmitter. This project has looked broadly at the start up and ongoing operation of secure wireless networks containing various numbers of users, ranging from a point-to-point connection between two users, to wireless networks whose number of users grows without bound. In the context of wireless communication between two users, we have initiated the study of the foundations of covert communications, where the goal is not only to keep the contents of the message secret from an eavesdropper, but also to hide the very existence of the communication. This extra layer of privacy has taken on added importance due to revelations in recent years of the surveillance of communication links to collect valuable "meta-data" (whom is talking to whom), rather than the actual message contents. Whereas the field of covert wireless communications has deep roots in a military context, the fundamental limits that determine the maximum rate at which information can be transmitted covertly had not been uncovered prior to this project. In particular, this project has established a method such that, in some period of time, a user can covertly transmit an amount of information proportional to the square root of the length of that time period, without detection of that transmission by an adversary. The project also established that a user cannot transmit more than that amount of information using any scheme. This was the first result establishing the fundamental limits of covert communications. As users of any mobile device know, maximizing battery lifetime is a critical goal for a wireless communications system. Hence, minimizing the energy required for communication, which generally dominates the energy budget of mobile devices, is of high priority. This project has considered how to minimize the energy required for transmission in moderate-sized secure wireless networks. Consider a network where the message is sent through a route consisting of multiple wireless "hops" between system devices on its way from the transmitter to the receiver. In a traditional approach, the route for the message would be chosen to minimize the energy consumption without the consideration of security, and then energy would be consumed in securing each of the individual hops. Here, we have considered how the selection of the route to convey the message from transmitter to receiver can utilize information on the security vulnerabilities of each link, be it to an eavesdropper trying to listen to the signal or an adversary trying to jam the signal, to select a route that will save energy in a secure transmission from the transmitter to the receiver. Significant energy savings are achieved through the proposed designs versus previous methods that decouple route selection and the securing of the individual hops. As evidenced by the now-traditional Internet or the more recently emerging Internet-of-Things, information networks can grow rapidly to extremely large scales. When such explosive growth occurs, behaviors can emerge that had been heretofore unobserved and expected. Hence, the scaling of various networks properties as the size of the network becomes very large has emerged in recent years as an active area of study. This project has considered how to secure a message from eavesdroppers in this context. In particular, consider a very large wireless network with many pairs of transmitters and receivers all trying to talk to each other. Suppose that there are many eavesdroppers also located in the network. A pertinent question is how many eavesdroppers can be present before secure operation of the network is compromised. In this project, we have developed a conclusive result: through a construction developed here, secure transmission at the traditional rate can take place in the presence of any number of eavesdroppers, and the network nodes need not even be aware of the location of those eavesdroppers. Finally, this project has considered how to initialize an ad hoc wireless network, where a collection of nodes might be dropped into a given scenario, turned on, and expected to form a communication network. The first step in such a network operation is the classic problem of "neighbor discovery", where the nodes need to figure out the identity of the nodes whom are within their radio range and can be reached directly without hopping through another node. This project has developed efficient methods for neighbor discovery that require neither information on the number of neighbors nor synchronization among the nodes. We have also shown how feedback from nodes that they have discovered neighbors can significantly improve the performance of such algorithms.

Agency
National Science Foundation (NSF)
Institute
Division of Computer and Network Systems (CNS)
Type
Standard Grant (Standard)
Application #
1018464
Program Officer
Darleen L. Fisher
Project Start
Project End
Budget Start
2010-09-01
Budget End
2014-08-31
Support Year
Fiscal Year
2010
Total Cost
$425,000
Indirect Cost
Name
University of Massachusetts Amherst
Department
Type
DUNS #
City
Hadley
State
MA
Country
United States
Zip Code
01035