Program Notes
Sensor Networks You Can Trust
As the military makes growing use of networks of sensors for key aspects of operations, the Air Force is funding research to find new ways of determining whether those sensors—a physical device, a human collecting intelligence or a computer—can be trusted to provide accurate data.
The Air Force Research Lab (AFRL) at Wright-Patterson Air Force Base, Ohio, recently awarded a $100,000 contract to Modus Operandi, a software technology company, to develop a cutting-edge framework for evaluating the integrity of the end components in a communications network. The contract, titled “Green Wave: Assuring Trust between the Edges,” is sponsored by the Office of the Secretary of Defense under the Small Business Innovation Research program.
Under Green Wave, Modus Operandi will develop an open architectural framework, along with a supporting protocol stack, that allows the trust associated with network elements to be evaluated and acted upon within a networked environment. This will include a sophisticated semantic technology service that allows subjective information to be incorporated into trust calculations. Modus Operandi will also provide an analysis of the security of the overall architecture.
According to Mark Heileman, vice president, advanced programs, for Modus Operandi, the research will build upon the overarching vision of layered sensing laid out earlier this year by the AFRL Sensors Directorate, which calls for a global web of sensors to provide decision-makers with timely, actionable and trusted information for better situational awareness.
Trust, Heileman noted, has many definitions, even in this specific context. “But when our customer defines trust, they specifically mean whether a node has been compromised or not,” he said. “That compromise could be malicious, for example by a cyber-attack, or non-malicious— the node has failed or is providing bad data for some reason. We’re looking to apply the trust metric to the question of whether the nodes have been compromised. You can think of the local areas that interoperate with multiple sensors, and one is providing a different reading than the others. You want to determine whether data from that sensor is trustworthy.”
Efforts to manage trust, he continued, have typically been approached from two ends of a spectrum. “The reputation-based approach used by eBay tends to apply to peer-to-peer networks, where there are recommendations about trustworthiness. Usually there is some kind of recommendation system, with transitivity—if A trusts B, and C trusts A, then C can trust B. It’s called reputation-based trust. The other end of the spectrum is policybased trust, which typically involves credentialing from a certificate authority. We’re looking at a hybrid approach that involves both.
“Another issue in managing trust is that it has traditionally been treated in a communications network as a binary metric—the node is either trusted or not trusted, and if it’s not trusted, you ignore it. We’re taking the viewpoint that a trust metric could be a gradation, between completely trusted or untrusted nodes. Part of our research into a metric will be looking at how we account for that gradation. It could be that the context or the environment has changed, and as a result you either want to increase or decrease the trust level,” he said.
Military interest in evaluating the trustworthiness of a sensor or other network node has been spurred by several factors, Heileman suggested. One is the growing prevalence of cyber attacks and intrusions on military networks, which could compromise results from sensors. Another is evidence that counterfeit computer chips are making their way into the military supply chain, raising the risks both of operational failures and of secret, malicious “backdoors” that could allow unauthorized leaks of information.
Modus Operandi also hopes to apply its expertise in semantic technology to address another key issue in sensor networking, which is that the vast number of sensor systems on the Global Information Grid use different technology languages, making it difficult for sensor nodes to communicate with one another. “One of the goals of the research is to see how we can effectively communicate trust between end nodes.
We’re looking at semantic technology as being an interpreter between end nodes that speak different languages,” Heileman said. ♦
* Compiled by KMI Media Group staff
