Wade Trappe

Research

 

During the past few years, I have led my group in an effort to revisit the concepts of security, and examine whether there are additional notions for security in the wireless domain. In this regard, my group has leveraged our strengths in radio resource management, physical layer technologies, and systems-level experimentation. By combining these aspects with more traditional cryptographic security methodologies, we have been able to produce results that will have a practical impact on current and future system deployment.

 

Supporting this objective, my current portfolio of research efforts at WINLAB include both traditional cryptographic security techniques, as well as a variety of methods that emphasize security and privacy issues that are uniquely associated with wireless systems. A survey of several current research thrusts in wireless security is provided below:

 

Staggered TESLA and DoS-Resistance in Secure Multicasting: Many techniques for multicast authentication employ the principle of delayed key disclosure. These methods introduce delay in authentication, employ receiver-side buffers, and are susceptible to denial of service (DoS) attacks. Delayed key disclosure schemes have a binary concept of authentication and do not incorporate any notion of partial trust. We have developed a scheme, Staggered TESLA, which provides multi-grade authentica tion, reduces the delay needed to filter forged multicast packets, and consequently mitigates the effects of DoS attacks. Staggered TESLA achieves this notion of multi-level trust through the use of multiple, staggered authentication keys in creating message authentication codes (MACs) for a multicast packet. Complementary techniques for reducing delay through the assurance of the trustworthiness of entities in a neighborhood of the source, and through the introduction of additional key distributors within a network have been studied. Staggered TESLA has been experimentally validated in an adversarial setting using the ORBIT Wireless Testbed (www.orbit-lab.org).

 

Related Publications: Staggered TESLA was originally described in a Globecom paper, and a significantly more expanded investigation was provided in an IEEE Trans. Information Forensics and Security journal article. The prototype demonstration of Staggered TESLA was presented recently at WINLAB during a regular industrial advisory board meeting.

 

Privacy Augmented Relaying of Information from Sensors (PARIS): A first line of defense for protecting sensor communications involves employing cryptography. However, these methods cannot address the complete spectrum of privacy issues that will arise in sensor networks. Specifically, security solutions are inadequate for protecting the privacy of contextual information surrounding a sensor application. In this project, we are developing solutions that provide contextual privacy for wireless sensor networks. PARIS addresses the following critical contextual privacy issues: source-location privacy (where was the source of a sensor reading located?), temporal privacy (when did the transmission originate?), and traffic privacy (can meaning be inferred from the size of the message alone?). A suite of privacy-enhanced routing algorithms has been developed for deployment on sensor networks. One of our proposed routing schemes, phantom routing, has been shown to provide significantly enhanced source-location privacy while incurring controlled amounts of additional overhead (i.e. latency and energy usage).

 

Related Publications: We originally mapped out the problem of contextual privacy in an ACM SASN paper, where we  provided an initial investigation of several routing schemes in. A more thorough investigation has appeared in our ICDCS paper. Subsequent work on temporal privacy has appeared in ICDCS 2007.

 

Defenses for Attacks of Radio interference in WIreless Networks (DARWIN): Here, I am targeting a severe threat that looms on the horizon for a large class of commercial wireless networks. Many wireless networks, such as 802.11, are susceptible to simple forms of radio interference attacks (i.e. jamming) that can prevent other wireless devices from even being to transmit or receive. DARWIN consists of methods for diagnosing the presence of radio interference, as well as defense mechanisms for coping with radio interference. One promising defense strategy that we have developed is channel surfing, whereby wireless devices individually detect the presence of interference and autonomously adjust their operating frequencies in order to evade the interference and reestablish connectivity on a different channel. The DARWIN channel surfing protocols have been prototyped on a testbed of 30 wireless nodes. This technology promises to be very important in the future as it provides a means for networks to evolve and repair themselves in the presence of intentional or accidental radio interference. A second approach, known as spatial retreats, has been proposed and involves wireless nodes altering their spatial positions in order to best evade RF interference.

 

Related Publications: An overview of the jamming problem is described in our IEEE Network survey paper and our original ACM Wise paper, while the issue of detecting jamming has been more thoroughly investigated in our ACM Mobihoc, and spatial retreats is presented in our IEEE Trans. Parallel and Distributed Systems paper. A channel surfing prototype was built, which involved 30 node sensor network, and has appeared in ACM IPSN conference. More detailed explanation of our prototype will appear in ACM Trans. on Sensor Networks.

 

Security Via Lower Layer Enforcements (SEVILLE): Another direction where I am using wireless-specific properties involves using the wireless medium itself to enhance traditional approaches to authentication and confidentiality. I have been developing a way to extract forge-resistant signatures that can identify wireless transmitters without requiring conventional cryptographic authentication mechanisms. This is particularly important because, without some form of identification, wireless networks are susceptible to spoofing, whereby one device can imitate another device. There are many scenarios, such as a hotspot at a local coffee shop, where the constantly evolving customer base makes it impractical to deploy conventional authentication methods. Unfortunately, for these networks, it is a very simple matter to launch spoofing attacks and introduce evil twin access points. SEVILLE  makes it possible to detect spoofing and thus can thwart spoofing by triggering appropriate countermeasures. Initial experiments for SEVILLE have been conducted on the ORBIT Wireless Testbed (www.orbit-lab.org), where time averaged readings have been used to establish coarse, but effective, signatures for identifying wireless devices and detecting spoofing. Further directions for investigation involve utilizing more finely measured physical layer measurements, such as exists in the digitized waveforms available to software defined radio platforms, to conduct authentication and facilitate key establishment in support of higher-layer confidentiality services.

 

Related Publications: The SEVILLE project is currently one of the major focuses of my research group, with funding coming from NSF, DARPA and industry. We provide an overview of the problem area in our position paper, which appeared at ACM Wise 2006. Subsequently, we have published information-theoretic analysis of secret communication in CISS, Allerton, and ISIT. On the systems side, we have published a paper on relationship-based strategies for detecting device spoofing, which appeared in IEEE Trans. on Information Forensics and Security. Specific physical-layer based authentication methods were explored in an ICC paper, and a journal version will shortly appear in IEEE Trans. on Wireless Communications. Ongoing work is focusing on implementing SEVILLE using software defined radio platforms, and preliminary results have appeared in our ACM Wireless Security Worskhop paper, and our IEEE SDR Workshop paper (where we showed that physical layer authentication can be achieved using WiFi waveforms and SDR platforms!)

 

Securing Wireless Localization: As more location-dependent services get deployed, the very mechanisms that provide location information will become the target of misuse and attacks. Therefore, as we move forward with deploying wireless systems that support location services, it is prudent to integrate security into the protection of localization techniques. My group has examined the problem of securing the localization infrastructure. We have identified a variety of non-cryptographic attacks that can be launched against localization algorithms. In order to defend against such attacks, we have proposed two different approaches. In the first approach, rather than introduce countermeasures for every possible attack, we have sought to provide localization-specific, attack-tolerant mechanisms that shield the localization infrastructure from threats that bypass traditional security defenses. The idea is to live with bad nodes rather than eliminate all possible bad nodes. The second strategy that we are investigating involves developing a suite of attack-diagnosis mechanisms that identify the presence of localization attacks, and can be thought of as statistical intrusion detection mechanisms for wireless localization systems.

 

Related Publications: In our IPSN paper various localization-specific attacks have been identified and several

robust statistical methods were investigated for coping with these attacks. A comparative study investigating the natural robustness of different localization algorithms was presented in our DCOSS paper. The effect of landmark deployment on localization has been investigated in a 2006 SECON paper.

 

Spatio-Temporal Access Control (STAC): Historically, wireless networks have freed users from the confines of static, wired networks, and traditionally a service has access control mechanisms that are not based upon the geographic properties associated with the wireless user. The fact that wireless networks are becoming increasingly ubiquitous, however, suggests that it is not necessary to restrict access to services based solely on conventional identity-based authenticators. Rather, the wireless infrastructure can facilitate location-aware computing paradigms, where services are only accessible if the user is in the right place at the right time. For example, location-aware security services, such as ensuring that a file can only be accessed within a specific secure room, or that a laptop no longer functions when it is taken outside of a building, are not only desirable but will soon become feasible. I am investigating techniques to provide access control based on the spatio-temporal context surrounding mobile users. To facilitate this new form of security, we are investigating methods for trustworthy localization of wireless nodes, new forms of challenge-response protocols that involve wireless entity proving that they are in a specific location, the ontological representation of access control policies using automata, and network architectures that support spatio-temporal access control.

 

Related Publications: The research into building STAC systems is at an early stage, and currently one

paper has appeared in the 2006 ACM Workshop on Security of Ad hoc and Sensor Networks.