Scalable, Energy-Aware Fault Tolerance Approaches for Large Scale Systems
The U.S. Department of Energy has identified resilience and energy consumption as key challenges for future extreme-scale systems. It is commonly thought that current checkpoint/restart methods wouldn’t scale. This project aims at proposing new fault tolerance scheme to minimize energy as well as increase scalability for future large scale system.
People: Xiaolong Cui, Bryan Mills, Longhao Li, Taieb Znati, and Rami Melhem
Community Resilience to Hazards
The National Academies’ report laid out the challenge of defining and building resilience to hazards as a long-term effort that would engage the ‘whole nation’ including scientists, governmental agencies at all levels of jurisdiction, private and nonprofit organizations, and communities. To meet this challenge, it is essential to define, design, and demonstrate an interdisciplinary, dynamic processes that will transform societal understanding of risk and enable self-organized, collective action to support the resilient management of hazards. Addressing this goal, we utilize the threat of Near-Field Tsunami(NFT) in a location prone to this risk, Padang, West Sumatra, Indonesia as a case study, as it includes all components of hazard evolution and hazard response, our study includes developing the concepts, methods, instruments, and models needed to detect NFTs, transmit the data through a network of sensors, analyze its likely impact on the community, and represent this risk visually to emergency managers.
People: Fuli Ai, Xerandy and Taieb Znati
Cognitive Radio Networks Research
Dynamic Spectrum Access
Static spectrum allocation results in large portions of unused spectrum, both spatially and temporarily, leading to an apparent spectrum scarcity and deployment difficulty. Dynamic Spectrum Access (DSA), enabled by cognitive radios, has emerged as a promising solution to dramatically improve spectrum utilization and reduce overhead deployment, with significant impact on a wide range of wireless services applications, including online learning and multimedia conferencing, emergency care and public safety, smart electric grid, and mobile healthcare. A cognitive radio has the ability to leverage situational knowledge and intelligent processing to autonomously and dynamically adjust its operating and environmental parameters to maximize spectrum utilization. While conceptually simple, the realization of DSA is challenging. This research program addresses several fundamental problems related to DSA, focusing on the development of new theories, frameworks and protocols, to enhance throughput, delay, and fairness in future wireless and cognitive radio networks.
People: Daniel Petrov, Debarun Das and Taieb Znati
Dynamic Spectrum Enforcement
As spectrum is shared more intensively, it will become necessary to automate enforcement procedures to the extent possible so that the outcomes of the enforcement process can occur in near real time and at scale. Much of the research and practice to date has focused on interference protection of the incumbent and on preventative (ex ante) approaches. Static ex ante approaches (such as exclusion or protection zones) are not easily adaptable to policy changes. Database-driven ex ante enforcement (e.g. TV White Spaces database and the Spectrum Access System, or SAS) can be more responsive to changes but require constant connectivity with radios and may not be geographically fine-grained. This research project addresses the development of efficient policy-based reasoning engines to reduce the social cost of ex ante enforcement, advancement of techniques to automate the enforcement of events after they occur (ex post enforcement) and examination of different institutional strategies for ex post enforcement (third party, self enforcement, cooperative mutual enforcement) with the goal of understanding under what circumstances each approach would apply to spectrum sharing.
People: Debarun Das and Taieb Znati
Intelligent Network Security
Distributed Denial of Services (DDoS) attacks continue to be one of the 
most challenging threats in the Internet. The intensity, frequency and sophistication of these attacks are increasing at an alarming rate. Numerous schemes have been proposed to mitigate the impact of DDoS attacks. These schemes use a number of approaches to protect against DDoS attacks, including utilizing stochastic analysis to monitor and identify network traffic behaviors, exploiting the entropy information of network traffic to characterize the network status, and leveraging machine learning techniques to recognize network traffic patterns. Despite significant advances in the state-of-the-art, DDoS attacks remain a challenge problem. In this project, we explore the potential of a distributed, adaptive and intelligent DDoS detection and mitigation system.
People: Xiaoyu Liang, and Taieb Znati
Ubiquitous and Pervasive Computing and Tetherless Care
Despite the advances in routing and media access control technologies, 
little progress has been made towards large-scale deployment of services and applications in pervasive and ubiquitous environments. The lack of a fixed infrastructure, coupled with the time-varying characteristics of the underlying network topology make service delivery challenging. The goal of this research is to address the fundamental design issues of a service infrastructure for ubiquitous environments and provide a comprehensive solution which is robust, scalable, secure and takes into consideration node mobility and resource constraints.
Tetherless care is a novel healthcare delivery paradigm that enables interaction between caregivers and patients beyond the confines of traditional points of care. The advances in health information technology in support of the patient-centric paradigm coupled with the advances in body area networks and home health monitoring have created opportunities to enable ambulatory care and to strengthen the ongoing dialog between patient and caregiver with tangible, unbiased medical data.
People: Longhao Li and Taieb Znati