Biblio

Current cryptographic solutions used in information technologies today like Transport Layer Security utilize algorithms with underlying computationally difficult problems to solve. With the ongoing research and development of quantum computers, these same computationally difficult problems become solvable within reasonable (polynomial) time. The emergence of large-scale quantum computers would put the integrity and confidentiality of today’s data in jeopardy. It then becomes urgent to develop, implement, and test a new suite of cybersecurity measures against attacks from a quantum computer. This paper explores, understands, and evaluates this new category of cryptosystems as well as the many tradeoffs among them. All the algorithms submitted to the National Institute of Standards and Technology (NIST) for standardization can be categorized into three major categories, each relating to the new underlying hard problem: namely error code correcting, algebraic lattices (including ring learning with errors), and supersingular isogenies. These new mathematical hard problems have shown to be resistant to the same type of quantum attack. Utilizing hardware clock cycle registers, the work sets up the benchmarks of the four Round 3 NIST algorithms in two environments: cloud computing and embedded system. As expected, there are many tradeoffs and advantages in each algorithm for applications. Saber and Kyber are exceedingly fast but have larger ciphertext size for transmission over a wire. McEliece key size and key generation are the largest drawbacks but having the smallest ciphertext size and only slightly decreased performance allow a use case where key reuse is prioritized. NTRU finds a middle ground in these tradeoffs, being better than McEliece performance wise and better than Kyber and Saber in ciphertext size allows for a use case of highly varied environments, which need to value speed and ciphertext size equally. Going forward, the benchmarking system developed could be applied to digital signature, another vital aspect to a cryptosystem.

This work investigates entropy-based prevention of Distributed Denial-of-Service (DDoS) attacks for Software Defined Networks (SDN). The experiments are conducted on a virtual SDN testbed setup within Mininet, a Linux-based network emulator. An arms race iterates on the SDN testbed between offense, launching botnet-based DDoS attacks with progressive sophistications, and defense who is deploying SDN controls with emerging technologies from other faucets of cyber engineering. The investigation focuses on the transmission control protocol’s synchronize flood attack that exploits vulnerabilities in the three-way TCP handshake protocol, to lock up a host from serving new users.The defensive strategy starts with a common packet filtering-based design from the literature to mitigate attacks. Utilizing machine learning algorithms, SDNs actively monitor all possible traffic as a collective dataset to detect DDoS attacks in real time. A constant upgrade to a stronger defense is necessary, as cyber/network security is an ongoing front where attackers always have the element of surprise. The defense further invests on entropy methods to improve early detection of DDoS attacks within the testbed environment. Entropy allows SDNs to learn the expected normal traffic patterns for a network as a whole using real time mathematical calculations, so that the SDN controllers can sense the distributed attack vectors building up before they overwhelm the network.This work reveals the vulnerabilities of SDNs to stealthy DDoS attacks and demonstrates the effectiveness of deploying entropy in SDN controllers for detection and mitigation purposes. Future work includes provisions to use these entropy detection methods, as part of a larger system, to redirect traffic and protect networks dynamically in real time. Other types of DoS, such as ransomware, will also be considered.

Anomalous users detection in social network is an imperative task for security problems. Motivated by the great power of Graph Neural Networks(GNNs), many current researches adopt GNN-based detectors to reveal the anomalous users. However, the increasing scale of social activities, explosive growth of users and manifold technical disguise render the user detection a difficult task. In this paper, we propose an innovate Relevance-aware Anomalous Users Detection model (RAU-GNN) to obtain a fine-grained detection result. RAU-GNN first extracts multiple relations of all types of users in social network, including both benign and anomalous users, and accordingly constructs the multiple user relation graph. Secondly, we employ relevance-aware GNN framework to learn the hidden features of users, and discriminate the anomalous users after discriminating. Concretely, by integrating Graph Convolution Network(GCN) and Graph Attention Network(GAT), we design a GCN-based relation fusion layer to aggregate initial information from different relations, and a GAT-based embedding layer to obtain the high-level embeddings. Lastly, we feed the learned representations to the following GNN layer in order to consolidate the node embedding by aggregating the final users' embeddings. We conduct extensive experiment on real-world datasets. The experimental results show that our approach can achieve high accuracy for anomalous users detection.

With wide applications like surveillance and imaging, securing underwater acoustic Mobile Ad-hoc NETworks (MANET) becomes a double-edged sword for oceanographic operations. Underwater acoustic MANET inherits vulnerabilities from 802.11-based MANET which renders traditional cryptographic approaches defenseless. A Trust Management Framework (TMF), allowing maintained confidence among participating nodes with metrics built from their communication activities, promises secure, efficient and reliable access to terrestrial MANETs. TMF cannot be directly applied to the underwater environment due to marine characteristics that make it difficult to differentiate natural turbulence from intentional misbehavior. This work proposes a trust model to defend underwater acoustic MANETs against attacks using a machine learning method with carefully chosen communication metrics, and a cloud model to address the uncertainty of trust in harsh underwater environments. By integrating the trust framework of communication with the cloud model to combat two kinds of uncertainties: fuzziness and randomness, trust management is greatly improved for underwater acoustic MANETs.

With the development of IoT, smart health has significantly improved the quality of people's life. A large amount of smart health monitoring system has been proposed, which provides an opportunity for timely and efficient diagnosis. Nevertheless, most of them ignored the impact of environment on patients' health. Due to the openness of the communication channel, data security and privacy preservation are crucial problems to be solved. In this work, an environment aware privacy-preserving authentication protocol based on the fuzzy extractor and elliptic curve cryptography (ecc) is designed for health monitoring system with mutual authentication and anonymity. Edge computing unit can authenticate all environmental sensors at one time. Fuzzy synthetic evaluation model is utilized to evaluate the environment equality with the patients' temporal health index (THI) as an assessment factor, which can help to predict the appropriate environment. The session key is established for secure communication based on the predicted result. Through security analysis, the proposed protocol can prevent common attacks. Moreover, performance analysis shows that the proposed protocol is applicable for resource-limited smart devices in edge computing health monitoring system.

System penetration testing is an important measure of discovering information system security issues. This paper summarizes and analyzes the high-risk problems found in the penetration testing of the artificial storm prediction system for power grid storm disasters from four aspects: application security, middleware security, host security and network security. In particular, in order to overcome the blindness of PGRDAIPS current SQL injection penetration test, this paper proposes a SQL blind bug based on improved second-order fragmentation reorganization. By modeling the SQL injection attack behavior and comparing the SQL injection vulnerability test in PGRDAIPS, this method can effectively reduce the blindness of SQL injection penetration test and improve its accuracy. With the prevalence of ubiquitous power internet of things, the electric power information system security defense work has to be taken seriously. This paper can not only guide the design, development and maintenance of disaster prediction information systems, but also provide security for the Energy Internet disaster safety and power meteorological service technology support.