This paper presents a password authenticated group key exchange protocol without the password sharing assumption. To obtain the passwords, wireless devices are used to extract short secrets at the physical layer. Then, users in our protocol can establish a group key at higher layers with light computation consumptions. Thus, our protocol is a cross-layer design. Additionally, our protocol is a compiler, i.e., it can transform any provably secure two-party password authenticated key exchange protocol into a password authenticated group key exchange protocol with only one more round of communications. Besides, the proposed protocol is proved secure in the standard model.
The argumentative structure of texts is increasingly exploited for analysis tasks. Most approaches, however, model only local structure of single arguments. This article investigates how to capture the global structure of the discourse-level argumentation of a text. We propose to model global structure as a flow of task-related rhetorical moves and to then compare the flow to common flow patterns. Thereby, we map texts into the feature space of global structures. In sentiment analysis and essay scoring experiments, our model proves effective and more domain-robust than strong baselines. We discuss the universality of the model for discourse-level argumentation analysis.
Analyzing the security of Wearable Internet-of-Things (WIoT) devices is considered a complex task due to their heterogeneous nature. In addition, there is currently no mechanism that performs security testing for WIoT devices in different contexts. In this paper, we propose an innovative security testbed framework targeted at wearable devices that is designed to conduct a set of security tests and analyze the behavior of WIoT devices in different user contexts, by realistically simulating environmental conditions in which they operate. The architectural design and a proof-of-concept for testbed operation, demonstrating the detection of context-based attacks executed by smartwatch devices, are presented.
The Evolved Packet System-based Authentication and Key Agreement (EPS-AKA) protocol of the LTE network does not support IoT objects and has various security imitations including: transmission of the object's (user/device) identity and key set identifier in plaintext over the network, synchronization issue, large overhead, limited identity privacy, and security attacks vulnerabilities. In this paper, we propose a new secure and efficient AKA protocol for the LTE network that supports secure and efficient communications among various IoT devices as well as among the users. Analysis shows that our protocol is secure, efficient, privacy-preserved, and reduces bandwidth consumption during authentication.
We consider a crowdsourcing system where mobile users form a local community to help each other transfer data. In particular, we design a system and algorithms to solve the problems of optimizing the perspectives of mobile clients who request nearby mobile users to transfer data, and mobile hotspots who open their cellular network connections and serve requests from mobile clients. Our evaluation results on an Android testbed and a packet-level simulator indicate that: (i) mobile clients can tune preferred trade-off between cost and delay, (ii) mobile hotspots comply with all delay bounds, and (iii) our system is efficient and practical.
In Online Social Networks, user reputation scores are computed according to two orthogonal perspectives: helpfulness-based reputation (HBR) and centrality-based reputation (CBR). In HBR approaches, the most reputable users post the most helpful reviews; in CBR approaches, the most reputable users occupy the most central positions in the network encoding trust links. We used datasets extracted from CIAO, Epinions and Wikipedia and five centrality measures to calculate CBR scores. We showed that CBR scores can predict HBR ones: because user reviews are sparse, we could leverage trust relationships to spot those users producing the most helpful reviews for the whole community.
Wireless-enabled devices owned by a user has seen a huge growth in number over the last few years with one third of adults in the United States currently having three wireless devices - smartphone, laptop and tablet. This paper provides a study of the network usage behavior of today's multi-device users, using a data collected from a large university campus of over 30,000 users. The study reveals several interesting findings about multi-device users like, how current DHCP configurations are oblivious to multiple devices which result in inefficient utilization of available IP address space.
Conventional security techniques are not ideal solutions for security and privacy of data in resource constrained IoT systems. Typical battery operated IoT devices cannot afford using cryptographic algorithms due to their high power consumption. We have proposed an intrusion detection and prevention mechanism by implementing an intelligent security architecture using Random Neural Networks. The application's source code is also instrumented to detect illegal memory accesses, using a novel tag checking technique. The proposed solution is implemented in a smart building system, which successfully detects the presence of any suspicious node and anomalous activity with 97.23% accuracy and minimal performance overhead.