Applied Research

Browsing history sniffing attacks enable a party who controls a webpage to learn the browsing history of individuals who visit that webpage. Protecting browsing history is crucial to user privacy. If a user’s browsing history can be exposed by simply visiting a page controlled by the attacker, this can have serious consequences for the user, including discrimination in access to employment, housing and health opportunities, political persecution, and even the risk of arrest in the case of users in more oppressive regimes. In this project, we seek to advance the state of the art in both attacks and defenses for browser history sniffing. In particular, we will explore novel methods to increase the effectiveness of such attacks; at the same time, we will attempt to establish the foundations for defenses that allow optimization of the browser engine while protecting user privacy.

Cloud infrastructures are undergoing substantial changes as memory disaggregation architectures and technologies emerge. Memory disaggregation collects and combines unused memory resources from multiple cloud servers, and makes the resources available as a unified memory pool. It is mainly driven by the increasing demand for large memory capacity to support the deployment and execution of memory-intensive applications, such as graph analytics, AI/machine learning, and in-memory databases. Memory disaggregation is a flexible, efficient, and cost-effective solution for allocating and managing memory resources for these applications. It is made possible by the recent advancements in networking technologies (RDMA) and hardware interfaces (CXL), as well as improved memory virtualization technology.

For large-scale memory-intensive applications, data structures, such as B-trees, hash tables, and graphs are the main way to organize data in memory and a major performance factor. Our research will investigate how these and other key data structures should be adapted and/or redesigned for optimizing their performance in disaggregated memory systems. To better handle large amounts of data in distributed data structures and support fast data access and for efficient use of disaggregated memory, the underlying memory virtualization software must also be renovated and provide a more efficient, scalable, and flexible way to allocate and manage memory resources. Our research will investigate the required changes in this functionality as well.

Verification is important to ensure integrity in multiple use-cases, such as verification of the sender of a message or verification of the results computed over a database at the cloud. Let us discuss the verification of the sender of a message. A digital signature (an electronic and encrypted stamp on digital messages) verifies the authenticity of digital messages. In other words, a digital signature allows the receiver to verify the sender who signs the message (i.e., authentication), as well as verify the message originated from the signer/sender was not altered (i.e., message integrity). Any change made to the signed data invalidates the signature. A digital signature is created using asymmetric encryption techniques, having private and public keys. RSA and ElGamal signatures are examples of digital signatures. Such signature techniques are based on one of three hard mathematical problems: the integer factorization problem, the discrete logarithm problem, and/or the elliptic-curve discrete logarithm problem. The major drawback of the most commonly used digital signatures is that they are not post-quantum secure. Recently, several efforts have been placed to develop post-quantum secure encryption techniques. The National Institute of Standards and Technology (NIST), recently, announced four winners of a years-long competition to develop new quantum-secure encryption standards. However, one of them (based on Supersingular Isogeny Diffie-Hellman SIDH) was broken. In contrast, most current symmetric cryptographic algorithms (such as AES1) and hash functions (such as SHA-256 or SHA-512) are considered to be secure against attacks by quantum computers. We plan to explore methods to ensure authenticity and integrity (even in the presence of quantum attacks, i.e., attacks carried out by quantum computers) using publicly published pseudo-random arrays.

Scheduling Theory is one of the most studied fields in Operations Research and Computer Science. In a typical scheduling problem we are given a set of jobs that are to be processed on a set of machines so as to optimize certain objectives. In many scheduling problems resources (machines) are limited and thus not all jobs can be executed (by their due date). Sometimes a solution that optimizes the given objective may be ``unfair''. In loose terms this means that the solution may give preference to several jobs over others. We plan to explore algorithms that ensure a certain type of fairness, called proportionate fairness, in various scheduling scenarios.

Shales are tight rocks with ultra-low porosity and contain hydrocarbons (oil and gas). Both the United States (USA) and Israel have abundant shale formations containing oil and gas. The USA is currently the largest oil and gas producer in the world with the majority of its shale hydrocarbon production from the Permian basin, Texas, whereas the largest kerogen shale deposits in Israel occur in the Shfela area. For the recovery of oil and gas, shale formations are hydraulically fractured by injecting water-based fluids under high pressure. To address sustainability and global warming, it is now proposed to store greenhouse gases (CO2) and hydrogen (H2) in depleted hydrocarbon formations such as shales. Shale is composed of clays and silts, and they are bound by carbonates to form ultra-tight matrices that contain interconnected nano- to micro- to macro-voids with total porosity of 1-6%. These voids store oil and gas. Shale formations were isolated for millions of years and now the introduction of fracturing fluids, as well as storage of CO2 and H2, may cause the ultra-tight shale matrix to soften, and resulting in  the voids to seal thereby preventing the extraction of oil and gas as well as possible leakage of stored CO2 and H2 gases. The interaction artificially introduced fluids and shale formations, leading to softened shale is now attributed to the interlayer expansion of clays and dissolution of carbonates. In this project we are investigating the effect of shale softening on enhanced oil and gas recovery, as well as on storage of CO2 and H2. The shale and fluid interactions will be quantified through the lens of mineralogical, pore network, and mechanical property changes. 

The ultimate target of this proposal is to develop MXene-based electrodes for efficient electrochemical degradation of emerging organic pollutants. Electrochemical degradation has emerged as the most promising technique to remove organic pollutants from surface or ground water in terms of efficiency, electrode recyclability, and cost effectiveness. The choice of electrode materials plays a major role in the efficient electrochemical degradation. Among numerous kinds of materials, MXenes are considered as the most promising candidates due to their combined properties of superior electrical conductivity and good electrocatalytic activity, as well as the ease for fabrication. However, their use in electrochemical degradation is still a virgin territory to be explored. Here, different kinds of MXene-based electrodes will be generated for the electrochemical degradation of emerging organic pollutants, through solving the major issue of environmental stability of MXenes. The novelty of the project lies in the exploration of MXenes for electrochemical degradation and a deep understanding of their performance and mechanism. We expect that highly efficient electrochemical degradation of different kinds of organic pollutants can be achieved using the MXene-based electrodes. In addition, a variety of environmentally stable MXene solutions will be obtained for further applications in the field of not just environment, but also energy storage, catalysis, and electronics. Therefore, funding for will be needed for student scholarships, materials, consumables, and research center service fees. This seed grand is critical for the proposed work to generate enough preliminary results for future external funding applications.

PFAS are a group of synthetic organofluorine surfactants that are increasingly detected in many environmental matrices such as groundwater. This project will develop and evaluate electrically assisted adsorption of a broad range of per- and polyfluoroalkyl substances (PFAS) using externally charged electrically conducting membranes (ECMs). Adsorption and electro adsorption kinetics and capacity were evaluated for varying carbon materials using QCM under variations of DC charges/currents and accordingly used on composite electrically conductive membranes. Furthermore, the electro-oxidation of PFAS that adsorbed on the electrode membrane surface was also studied. The impact of water chemistry on electro-adsorption and electro-oxidaiton of PFAS were examined. The research unravels novel concentration and destruction approaches for trace-level PFAS from contaminated water (e.g., groundwater).