Showing papers in "IEEE Transactions on Emerging Topics in Computing in 2018"

Green Resource Allocation Based on Deep Reinforcement Learning in Content-Centric IoT

Abstract: In the era of information, the green services of content-centric IoT are expected to offer users the better satisfaction of Quality of Experience (QoE) than that in a conventional IoT. Nevertheless, the network traffic and new demands from IoT users increase along with the promising of the content-centric computing system. Therefore, the satisfaction of QoE will become the major challenge in the content-centric computing system for IoT users. In this article, to enhance the satisfaction of QoE, we propose QoE models to evaluate the qualities of the IoT concerning both network and users. The value of QoE does not only refer to the network cost, but also the Mean Opinion Score (MOS) of users. Therefore, our models could capture the influence factors from network cost and services for IoT users based on IoT conditions. Specially, we mainly focus on the issues of cache allocation and transmission rate. Under this content-centric IoT, aiming to allocate the cache capacity among content-centric computing nodes and handle the transmission rates under a constrained total network cost and MOS for the whole IoT, we devote our efforts to the following two aspects. First, we formulate the QoE as a green resource allocation problem under the different transmission rate to acquire the best QoE. Then, in the basis of the node centrality, we will propose a suboptimal dynamic approach, which is suitable for IoT with content delivery frequently. Furthermore, we present a green resource allocation algorithm based on Deep Reinforcement Learning (DRL) to improve accuracy of QoE adaptively. Simulation results reveal that our proposals could achieve high QoE performance for content-centric IoT.

Auction-Based Resource Allocation for Sharing Cloudlets in Mobile Cloud Computing

Abstract: Driven by pervasive mobile devices and ubiquitous wireless communication networks, mobile cloud computing emerges as an appealing paradigm to accommodate demands for running power-hungry or computation-intensive applications over resource-constrained mobile devices. Cloudlets that move available resources closer to the network edge offer a promising architecture to support real-time applications, such as online gaming and speech recognition. To stimulate service provisioning by cloudlets, it is essential to design an incentive mechanism that charges mobile devices and rewards cloudlets. Although auction has been considered as a promising form for incentive, it is challenging to design an auction mechanism that holds certain desirable properties for the cloudlet scenario. In this paper, we propose an incentive-compatible auction mechanism (ICAM) for the resource trading between the mobile devices as service users (buyers) and cloudlets as service providers (sellers). ICAM can effectively allocate cloudlets to satisfy the service demands of mobile devices and determine the pricing. Both the theoretical analysis and the numerical results show that the ICAM guarantees desired properties with respect to individual rationality, budget balance and truthfulness (incentive compatibility) for both the buyers and the sellers, and computational efficiency.

Personalized Search Over Encrypted Data With Efficient and Secure Updates in Mobile Clouds

Abstract: Mobile cloud computing has been involved as a key enabling technology to overcome the physical limitations of mobile devices toward scalable and flexible mobile services. In the mobile cloud environment, searchable encryption, which enables direct search over encrypted data, is a key technique to maintain both the privacy and usability of outsourced data in cloud. On addressing the issue, many research efforts resolve to using the searchable symmetric encryption (SSE) and searchable public-key encryption (SPE). In this paper, we improve the existing works by developing a more practical searchable encryption technique, which can support dynamic updating operations in the mobile cloud applications. Specifically, we make our efforts on taking the advantages of both the SSE and SPE techniques, and propose PSU, a Personalized Search scheme over encrypted data with efficient and secure Updates in mobile cloud. By giving thorough security analysis, we demonstrate that the PSU can achieve a high security level. Using extensive experiments in a real-world mobile environment, we show that the PUS is more efficient compared with the existing proposals.

Protecting Location Privacy for Task Allocation in Ad Hoc Mobile Cloud Computing

Abstract: Mobile cloud computing (MCC) is an emerging cloud-computing paradigm that integrates cloud computing and mobile computing to enable many useful mobile applications. However, the large-scale deployment of MCC is hindered by the concerns on possible privacy leakage. In this paper, we investigate the privacy issues in the ad hoc MCC, and propose a framework that can protect the location privacy when allocating tasks to mobile devices. Our mechanism is based on differential privacy and geocast, and allows mobile devices to contribute their resources to the ad hoc mobile cloud without leaking their location information. We develop analytical models and task allocation strategies that balance privacy, utility, and system overhead in an ad hoc mobile cloud. We also conduct extensive experiments based on real-world data sets, and the results show that our framework can protect location privacy for mobile devices while providing effective services with low system overhead.

Towards Real-Time Object Detection on Embedded Systems

Abstract: Convolutional neural network (CNN) based methods have achieved great success in image classification and object detection tasks. However, unlike the image classification task, object detection is much more computation-intensive and energy-consuming since a large number of possible object proposals need to be evaluated. Consequently, it is difficult for object detection methods to be integrated into embedded systems with limited computing resources and energy supply. In this paper, we propose a pipelined object detection implementation on the embedded platform. We present a comprehensive analysis of state-of-the-art object detection algorithms and select Fast R-CNN as a possible solution. Additional modifications on the Fast R-CNN method are made to fit the specific platform and achieve trade-off between speed and accuracy on embedded systems. Finally, a multi-stage pipelined implementation on the embedded CPU $+$ GPU platform with duplicated module-parallelism is proposed to make full use of the limited computation resources. The proposed system is highly energy-efficient and close to real-time performance. In the first Low-Power Image Recognition Challenge (LPIRC), our system achieved the best result with mAP/Energy of 1.818e-2/(W $\cdot$ h) on the embedded Jetson TK1 CPU $+$ GPU platform.

Decentralized and Optimal Resource Cooperation in Geo-Distributed Mobile Cloud Computing

Abstract: Mobile cloud computing is a key enabling technology in the era of the Internet of Things. Geo-distributed mobile cloud computing (GMCC) is a new scenario that adds geography consideration in mobile cloud computing. In GMCC, users are able to access cloud resources that are geographically close to their mobile devices. This is expected to reduce the communication delay and the service providers’ cost compared with the traditional centralized approach. In this paper, we focus on resource sharing through the cooperation among the service providers in geo-distributed mobile cloud computing. Then, we propose two different strategies for efficient resource cooperation in geographically distributed data centers. Furthermore, we present a coalition game theoretical approach to deal with the competition and the cooperation among the service providers. Utility functions have been specifically considered to incorporate the cost related to virtual machine migration and resource utilization. Illustrative results indicate that our proposed schemes are able to efficiently utilize limited resources with quality-of-service consideration.

Approximate Computing Using Multiple-Access Single-Charge Associative Memory

Abstract: Memory-based computing using associative memory is a promising way to reduce the energy consumption of important classes of streaming applications by avoiding redundant computations. A set of frequent patterns that represent basic functions are pre-stored in Ternary Content Addressable Memory (TCAM) and reused. The primary limitation to using associative memory in modern parallel processors is the large search energy required by TCAMs. In TCAMs, all rows that match, except hit rows, precharge and discharge for every search operation, resulting in high energy consumption. In this paper, we propose a new Multiple-Access Single-Charge (MASC) TCAM architecture which is capable of searching TCAM contents multiple times with only a single precharge cycle. In contrast to previous designs, the MASC TCAM keeps the match-line voltage of all miss-rows high and uses their charge for the next search operation, while only the hit rows discharge. We use periodic refresh to control the accuracy of the search. We also implement a new type of approximate associative memory by setting longer refresh times for MASC TCAMs, which yields search results within 1–2 bit Hamming distances of the exact value. To further decrease the energy consumption of MASC TCAM and reduce the area, we implement MASC with crossbar TCAMs. Our evaluation on AMD Southern Island GPU shows that using MASC (crossbar MASC) associative memory can improve the average floating point units energy efficiency by 33.4, 38.1, and 36.7 percent (37.7, 42.6, and 43.1 percent) for exact matching, selective 1-HD and 2-HD approximations respectively, providing an acceptable quality of service (PSNR > 30 dB and average relative error <10 percent). This shows that MASC (crossbar MASC) can achieve 1.77X (1.93X) higher energy savings as compared to the state of the art implementation of GPGPU that uses voltage overscaling on TCAM.

A Parity-Preserving Reversible QCA Gate with Self-Checking Cascadable Resiliency

Abstract: A novel Parity-Preserving Reversible Gate (PPRG) is developed using Quantum-dot Cellular Automata (QCA) technology. PPRG enables rich fault-tolerance features, as well as reversibility attributes sought for energy-neutral computation. Performance of the PPRG design is validated through implementing thirteen standard combinational Boolean functions of three variables, which demonstrate from 10.7 to 41.9 percent improvement over the previous gate counts obtained with other reversible and/or preserving gate designs. Switching and leakage energy dissipation as low as 0.141 eV and 0.294 eV, for 1.5 $E_k$ energy level are achieved using PPRG, respectively. The utility of PPRG is leveraged to design a one-bit full adder with 171 cells occupying only 0.19 $\mu \text{m}^2$ area. Finally, fault detection and isolation properties are formalized into a concise procedure. PPRG-based circuits capable of self-configuring active recovery for selected three-variable standard functions are realized using a memoryless method irrespective of garbage outputs.

SEARE: A System for Exercise Activity Recognition and Quality Evaluation Based on Green Sensing

Abstract: Green-computing technology and energy-saving design have become the focus of research in various fields in recent years. As a ubiquitously deployed infrastructure, WiFi can be considered as a platform for green sensing, and a plethora of efforts have been made in WiFi-based passive detection recently. However, little work has been done on the exercise activity recognition. In this paper, we propose SEARE, a novel energy-efficient solution using WiFi for exercise activity recognition. It is prototyped by fine-grained CSI extracted from existing commercial WiFi devices. Different from traditional features like mean or max value exploited in previous activity recognition works, involving either time or frequency information, we select CSI-waveform shape as activity feature, which contains the information from both of these two domains. A series of de-noise methods are designed, including low-pass, PCA, and median filtering, where PCA can remove the in-band noise that traditional low-pass filters fail to do. Finally the evaluation of activities quality can be made. Extensive experimental result validates the great performance of SEARE in both LOS and NLOS scenarios, with average recognition accuracies of 97.8 and 91.2 percent respectively.

Exploiting Voltage Regulators to Enhance Various Power Attack Countermeasures

Abstract: The security implications of on-chip voltage regulation on the effectiveness of various voltage/frequency scaling-based countermeasures such as random dynamic voltage and frequency scaling (RDVFS), random dynamic voltage scaling (RDVS), and aggressive voltage and frequency scaling (AVFS) are investigated. The side-channel leakage mechanisms of different on-chip voltage regulator topologies are mathematically analyzed and verified with circuit level simulations. Correlation coefficient between the input data and monitored power consumption of a cryptographic circuit is used as the security metric to compare the impact of different on-chip voltage regulators when implemented with the aforementioned countermeasures. As compared to a cryptographic circuit without countermeasure, the RDVFS technique implemented with an on-chip switched-capacitor voltage converter reduces the correlation coefficient over 80 percent and over 92 percent against differential and leakage power analysis attacks, respectively, through masking the leakage of the clock frequency and supply voltage information in the monitored power profile.

Hardware-Software Codesign of Wireless Transceivers on Zynq Heterogeneous Systems

Abstract: Recently, wireless technology has seen many new devices, protocols, and applications. As standards adapt to keep pace with hardware availability and user needs, the trend points towards systems that achieve high data rates with low energy consumption. Moreover, there is an emerging vision of a transceiver architecture that can adapt to multiple protocols, existing and evolving. This architecture maps computation to underlying heterogeneous computing elements, composed of processors and field programmable gate array (FPGA) fabric. Here, we introduce a method for modeling a generic orthogonal frequency division multiplexing (OFDM) wireless transceiver on the Zynq system-on-chip by decomposing the standard specifications into a set of functional blocks used in multiple protocols. Implementing the 802.11a physical (PHY) layer as an example, our approach creates Simulink model variants for both transmitter and receiver, each with a different boundary between hardware and software components. We use these models to generate hardware description language (HDL) code and bitstream for the programmable logic and C code with an executable for the advanced RISC machine (ARM) processor. We validate, profile, and analyze the models using metrics including frame time, resource utilization, and energy consumption. Our results demonstrate how to select a co-design configuration considering execution time and energy, and show how our platform can be reused for multiple-input multiple-output (MIMO) and protocol coexistence.

Demographic Information Prediction: A Portrait of Smartphone Application Users

Abstract: Demographic information is usually treated as private data ( e.g., gender and age), but has been shown great values in personalized services, advertisement, behavior study and other aspects. In this paper, we propose a novel approach to make efficient demographic prediction based on smartphone application usage. Specifically, we firstly consider to characterize the data set by building a matrix to correlate users with types of categories from the log file of smartphone applications. Then, by considering the category-unbalance problem, we make use of the correlation between users’ demographic information and their requested Internet resources to make the prediction, and propose an optimal method to further smooth the obtained results with category neighbors and user neighbors. The evaluation is supplemented by the dataset from real world workload. The results show advantages of the proposed prediction approach compared with baseline prediction. In particular, the proposed approach can achieve 81.21 percent of Accuracy in gender prediction. While in dealing with a more challenging multi-class problem, the proposed approach can still achieve good performance ( e.g. , 73.84 percent of Accuracy in the prediction of age group and 66.42 percent of Accuracy in the prediction of phone level).

Energy-Aware Bio-Signal Compressed Sensing Reconstruction on the WBSN-Gateway

Abstract: Technology scaling enables today the design of ultra-low power wearable bio-sensors for continuous vital signs monitoring or wellness applications. Such bio-sensing nodes are typically integrated in Wireless Body Sensor Network (WBSN) to acquire and process biomedical signals, e.g., Electrocardiogram (ECG), and transmit them to the WBSN gateway, e.g., smartphone, for online reconstruction or features extraction. Both bio-sensing node and gateway are battery powered devices, although they show very different autonomy requirements (weeks versus days). The rakeness -based Compressed Sensing (CS) proved to outperform standard CS, achieving a higher compression for the same quality level, therefore reducing the transmission costs in the node. However, most of the research focus has been on the efficiency of the node, neglecting the energy cost of the CS decoder. In this work, we evaluate the energy cost and real-time reconstruction feasibility on the gateway, considering different signal reconstruction algorithms running on a heterogeneous mobile SoC based on the ARM big.LITTLETM architecture. The experimental results show that it is not always possible to obtain the theoretical QoS under real-time constraints. Moreover, the standard CS does not satisfy real-time constraints, while the rakeness enables different QoS-energy trade-offs. Finally, we show that in the optimal setup (OMP, $n=128$ ) heterogeneous architectures make the CS decoding task suitable for wearable devices oriented to long-term ECG monitoring.

Optimizing Network Traffic for Spiking Neural Network Simulations on Densely Interconnected Many-Core Neuromorphic Platforms

Abstract: In this paper we present a new Partitioning and Placement methodology able to maps Spiking Neural Network on parallel neuromorphic platforms. This methodology improves scalability/reliability of Spiking Neural Network (SNN) simulations on many-core and densely interconnected platforms. SNNs mimic brain activity by emulating spikes sent between neuron populations. Many-core platforms are emerging computing targets that aim to achieve real-time SNN simulations. Neurons are mapped to parallel cores, and spikes are sent in the form of packets over the on-chip and off-chip network. However, the activity of neuron populations is heterogeneous and complex. Thus, achieving an efficient exploitation of platform resources is a challenge that often affects simulation scalability/reliability. To address this challenge, the proposed methodology uses customised SNN to profile the board bottlenecks and implements a SNN partitioning and placement (SNN-PP) algorithm for improving on-chip and off-chip communication efficiency. The cortical microcircuit SNN was simulated and performances of the developed SNN-PP algorithm were compared with performances of standard methods. These comparisons showed significant traffic reduction produced by the new method, that for some configurations reached up to 96X. Results demonstrate that it is possible to consistently reduce packet traffic and improve simulation scalability/reliability with an effective neuron placement.

On Achieving Asynchronous Energy-Efficient Neighbor Discovery for Mobile Sensor Networks

Abstract: Recent advances in mobile sensor networks (MSNs) lead to a wide demand of wireless communication based applications. However, due to the battery technology constraint, many MSNs-based applications are confined by the limited power resource capacity. Thus, discovering neighbors with minimal power consumption and latency becomes an indispensable characteristic to guarantee the feasibility of above applications. Most of previously proposed time-slotted-based neighbor discovery protocols excessively idealize the power consumption model, which ignores the power consumption and time duration of the transient state. In this paper, we propose a more practical model named enhanced power consumption model that considers the power consumption and time duration of the transient state. We then propose the asynchronous energy-efficient neighbor discovery protocols called Quick-Connect ( $Q-Connect$ ) including $Q-Connect_A$ , $Q-Connect_U$ and $Q-Connect_{UI}$ protocols, each of which can provide a strict upper bound on the discovery latency. We consider both the slot-aligned and slot-unaligned cases. For slot-aligned case, we propose the $Q-Connect_A$ protocol, which can greatly reduce the worst-case discovery latency. For slot-unaligned case, we first propose the $Q-Connect_U$ protocol, based on which we further propose an improved protocol called $Q-Connect_{UI}$ . Finally, we conduct state-based simulations to illustrate the effectiveness of the proposed $Q-Connect$ protocols.

Exploiting Task Elasticity and Price Heterogeneity for Maximizing Cloud Computing Profits

Abstract: This paper exploits cloud task elasticity and price heterogeneity to propose an online resource management framework that maximizes cloud profits while minimizing energy expenses. This is done by reducing the duration during which servers need to be left on and maximizing the monetary revenues when the charging cost for some of the elastic tasks depends on how fast these tasks complete, while meeting all the resource requirements. Comparative studies conducted using Google data traces show the effectiveness of our proposed framework in terms of improving resource utilization, reducing energy expenses, and increasing cloud profits.

Memristive Crossbar Memory Lifetime Evaluation and Reconfiguration Strategies

Abstract: Among the emerging technologies and devices for highly scalable and low power memory architectures, memristors are considered as one of the most favorable alternatives for next generation memory technologies. They are attracting great attention recently, due to their many appealing characteristics such as non-volatility and compatibility with CMOS fabrication process. But beside all memristor advantages, their drawbacks including manufacturing process variability and limited read/write endurance, could risk their future utilization. This paper will evaluate the impact of reliability concerns in lifetime of memristive crossbars and will present the design basis of two proposed reconfiguration approaches in memristive crossbar-based memories, in order to extend the system lifetime by utilizing available resources in an intense way and without need of failure recovery. It is observed that the adaptive reconfiguring approach can improve the crossbar reliability and extend its lifetime up to 65 percent in comparison with non-adaptive reconfiguration strategy.

On Minimizing Energy Consumption in FiWi Enhanced LTE-A HetNets

Abstract: As an integration of optical fiber networks and LTE-A heterogeneous networks (HetNets), fiber-wireless (FiWi) enhanced LTE-A HetNets will play a key role in supporting large-scale mobile data transmission. To improve the energy efficiency of the networks, we should put as much equipment into sleep state, and thus few user equipment (UE) can access Internet. In contrast, to maximize the UE connection all equipment should be in active state, which results in very high energy consumption. Obviously, there is a tradeoff between energy consumption minimization and UE connection maximization. However, existing works mostly focused on designing energy saving schemes for access networks but little can be found taking UE connection constraints into account. Toward this end, we provide this paper to explore the problem of minimizing energy consumption with UE connection constraints, where a constrained optimization problem is formulated and three solutions are presented, i.e., a brute force algorithm, a heuristic greedy algorithm, and a snowball rolling algorithm. Numerical results reveal that the heuristic greedy and snowball rolling algorithms have much higher computational efficiency than the brute force method, and can also achieve near-optimal solutions for the problem of minimizing energy consumption with UE connection constraints in FiWi enhanced LTE-A HetNets.

Analytical Model for Resistivity and Mean Free Path in On-Chip Interconnects with Rough Surfaces

Abstract: Planar copper interconnects suffer from surface roughness that results in performance degradation. This paper presents a novel analytical model for calculation effective resistivity and mean free path in on-chip copper interconnects. The closed form expressions are obtained from a generalized surface and grain boundary scattering approach that is combined with Mandelbrot-Weierstrass ( MW ) fractal function. It is observed that resistivity increases while mean free path reduces significantly for rough on-chip interconnects when compared with that of smooth lines. Current and future technology nodes i.e., 45 nm, 22 nm, 13 nm and 7 nm are considered for our analysis. The analytical models are validated against industry standard field solvers Ansys Q3D Extractor and previous data available in literature that exhibit excellent accuracy. Finally, we also present computational overhead in terms of simulation time, matrix size, number of tetrahedrons and memory for different values of roughness and technology nodes.

Massively Parallel Analog Computing: Ariadne’s Thread Was Made of Memristors

Abstract: This paper explores memristive grids where emergent computation arises through collective device interactions. Computing efficiency of the grids is studied in several scenarios and new composite memristive structures are utilized in shortest path and maze-solving computations. The dependence of the computing medium behavior on the symmetry of both the underlying geometry and the employed devices, is validated through SPICE-level circuit simulations, which highlight important computing inefficiencies. Particular circuit-models of memristive connections enable precise mapping of the target application on the computing medium. Extraordinary functionalities emerge when novel memristive computing components, comprising different electrical characteristics from their structural elements, are introduced in the grid. Applying assisted-computation, by incorporating the concept of Ariadne’s thread, leaded to better computing results, which could find application in routing and path computing problems.

Comprehensive Reliability-Aware Statistical Timing Analysis Using a Unified Gate-Delay Model for Microprocessors

Abstract: A framework is proposed to perform timing analysis of state-of-art microprocessors considering the impact of process-voltage-temperature (PVT) variations and the aging effect, including bias temperature instability (BTI), hot carrier injection (HCI), and time-dependent dielectric breakdown (TDDB). In this work, not only statistical timing analysis (StTA) due to each wearout mechanism is studied individually, but also the performance degradation while all these wearout mechanisms happen simultaneously is analyzed. Moreover, this work takes into account realistic use scenarios which include active, standby, and sleep modes. A unified gate-delay model, which combines both PVT variations and the aging effect, is constructed via a technique called multivariate adaptive regression splines (MARSP). Then a timing engine, which consists of two parts: a block-based analyzer and a path-based analyzer, is built to perform PVT-reliability-aware timing analysis. The accuracy and effectiveness of our framework has been verified on large industrial designs, like the LEON3 microprocessor, through a comparison with SPICE.

Classification and Mapping of Adaptive Security for Mobile Computing

Abstract: Context: Mobile computing has emerged as a disruptive technology that has empowered its users with portable, connected and context-aware computation. However, issues such as resource poverty, energy efficiency and specifically data security and privacy represent critical challenges for mobile computing. Objective: The objective of this work is to systematically identify, taxonomically classify and map the state-of-research on adaptive security (a.k.a. self-protection) for mobile computing. Methodology: We followed evidence based software engineering method to conduct a systematic mapping study of 43 qualitatively selected studies - published from 2003 to 2017 - on adaptive security for mobile computing. Results and Conclusions: Classification and mapping of the research highlights three prominent themes that support adaptive security for (i) Mobile Device Data and Resources, (ii) Mobile to Mobile Communication, and (iii) Mobile to Server Communication. Mapping analysis suggests that security of mobile device data and resources is the most researched theme. The mapping study highlights that active and futuristic research are primarily focused on security as a service, whereas; the frequent research challenges relate to self-protecting mobile devices, user-driven privacy decisions and context-aware security. The results of the mapping study facilitate knowledge transfer that can benefit researchers and practitioners to understand the role of adaptive and context-aware security in mobile computing environments.

Toward Reproducibility in Online Social Network Research

Abstract: The challenge of conducting reproducible computational research is acknowledged across myriad disciplines from biology to computer science. In the latter, research leveraging online social networks (OSNs) must deal with a set of complex issues, such as ensuring data can be collected in an appropriate and reproducible manner. Making research reproducible is difficult, and researchers may need suitable incentives, and tools and systems, to do so. In this paper, we explore the state-of-the-art in OSN research reproducibility, and present an architecture to aid reproducibility. We characterize the reproducible OSN research using three main themes: 1) reporting of methods; 2) availability of code; and 3) sharing of research data. We survey 505 papers and assess the extent to which they achieve these reproducibility objectives. While systems-oriented papers are more likely to explain data-handling aspects of their methodology, social science papers are better at describing their participant-handling procedures. We then examine incentives to make research reproducible, by conducting a citation analysis of these papers. We find that sharing data are associated with increased citation count, while sharing method and code does not appear to be. Finally, we introduce our architecture which supports the conduct of reproducible OSN research, which we evaluate by replicating an existing research study.

PADS: Design and Implementation of a Cloud-Based, Immersive Learning Environment for Distributed Systems Algorithms

Abstract: As distributed systems become more complex, understanding the underlying algorithms that make these systems work becomes even harder. Traditional learning modalities based on didactic teaching and theoretical proofs alone are no longer sufficient for a holistic understanding of these algorithms. Instead, an environment that promotes an immersive, hands-on learning of distributed systems algorithms is needed to complement existing teaching modalities. Such an environment must be flexible to support the learning of a variety of algorithms. The environment should also support extensibility and reuse since many of these algorithms share several common traits with each other while differing only in some aspects. Finally, it must also allow students to experiment with large-scale deployments in a variety of operating environments. To address these concerns, we use the principles of software product lines and model-driven engineering, and adopt the cloud platform to design an immersive learning environment called the Playground of Algorithms for Distributed Systems (PADS). A prototype implementation of PADS is described to showcase use cases involving BitTorrent Peer-to-Peer file sharing, ZooKeeper-based coordination, and Paxos-based consensus, which show the benefits of rapid deployment of the distributed systems algorithms. Results from a preliminary user study are also presented.

Memory Physical Aware Multi-Level Fault Diagnosis Flow

Abstract: Advanced methods of fault detection and diagnosis become increasingly important for the improvement of reliability, safety and efficiency in nanoscale designs. Because the existing approaches do not give a deeper insight and usually do not allow a comprehensive fault diagnosis, multi-level model based methods of fault detection were developed by using hierarchy of detection and diagnosis methods. This contribution proposes a memory physical (scrambling) aware multi-level fault diagnosis flow which is generic and applicable both for planar- and FinFET-based memories. In addition, special test algorithms for classification of static and dynamic faults are discussed while for classification of FinFET-specific faults a new test algorithm March FFDD is proposed. The flow is validated on 16nm FPGA board as well as it has been applied to numerous chips enabling successful physical failure analysis (PFA). At the end of the paper some real-life case scenarios of the flow application are presented.

Bayesian Cooperative Coalition Game as a Service for RFID-Based Secure QoS Management in Mobile Cloud

Abstract: In this paper, the problem of maintaining the quality of service with respect to high communication cost, available bandwidth, and security is investigated in a mobile cloud environment. A Bayesian cooperative coalition game is formulated in which learning automata stationed at Radio frequency identification (RFID) readers are assumed as the players. These players form a coalition among themselves using newly defined payoff value function (PVF) based upon the communication cost, network bandwidth, and storage requirements. Each player in the game makes his best efforts to increase his individual PVF for which it selects one of the strategies from the strategy space using the conditional probability. For each action taken by the players, they may get a reward or a penalty from the environment according to which they update their action probability vector. Tags are arranged in acyclic directed graph, and a secure access control algorithm is proposed, which results a complexity of $O(\log (n))$ in comparison with earlier solutions having the complexity of $O(n)$ . The proposed scheme is evaluated in comparison with other schemes by extensive simulations using various metrics. The results obtained show that the proposed scheme is effective in maintaining the RFID-based secure data dissemination in the mobile cloud environment. In particular, there is a reduction of 20% in computational time, and overhead generated with an increase of more than 30% in packet delivery ratio.

Mitigating Process Variability for Non-Volatile Cache Resilience and Yield

Abstract: While inclusion of emerging technology-based Non-Volatile Memory (NVM) devices in on-chip memory subsystems offers excellent potential for energy savings and scalability, their sensing vulnerability creates Process Variation (PV) challenges. This paper presents a circuit-architecture cross-layer solution to realize a radically-different approach to leveraging as-built variations via specific Sense Amplifier (SA) design and use. This novel approach, referred to as a Self-Organized Sub-bank (SOS) design, assigns the preferred SA to each Sub-Bank (SB) based on a PV assessment, resulting in energy consumption reduction and increased read access reliability. To improve the PV immunity of SAs, two reliable and power efficient SAs, called the Merged SA (MSA) and the Adaptive SA (ASA) are introduced herein for use in the SOS scheme. Furthermore, we propose a dynamic PV and energy-aware cache block migration policy that utilizes mixed SRAM and STT-MRAM banks in Last Level Cache (LLC) to maximize the SOS bandwidth. Our experimental results indicate that SOS can alleviate the sensing vulnerability by 89 percent on average, which significantly reduces the risk of application contamination by fault propagation. Furthermore, in the light of the proposed block migration policy, write performance is improved by 12.4 percent on average compared to the STT-MRAM-only design.

Disaster Rescue Task Scheduling: An Evolutionary Multiobjective Optimization Approach

Abstract: Efficient rescue task scheduling plays a key role in disaster rescue operations. In real-world applications, such an emergency scheduling problem often involves multiple objectives, complex constraints, inherent uncertainty, and limited response time requirement. In this paper, we propose a fuzzy multiobjective optimization problem of rescue task scheduling, the aim of which is to simultaneously maximize the task scheduling efficiency and minimize the operation risk for the rescue team. We then develop an efficient multiobjective biogeography-based optimization (EMOBBO) algorithm for solving the problem. To cope with the uncertainty, we employ three correlated fuzzy ranking criteria, and use the concept of fuzzy dominance for comparing the dominance relation of solutions. In EMOBBO, we define new migration and mutation operators for effectively evolving the permutation-based solutions, use a problem-specific solution rearrangement mechanism for filtering out inefficient solutions, and employ a local neighborhood structure to suppress premature convergence. Computational experiments show that the proposed EMOBBO algorithm outperforms some state-of-the-art evolutionary multiobjective optimization algorithms, and our algorithm has been successfully applied to several real-world disaster rescue operations in recent years.

Network-on-Chip Multicast Architectures Using Hybrid Wire and Surface-Wave Interconnects

Abstract: The network-on-chip (NoC) has been introduced as an efficient communication backbone to tackle the increasing challenges of on-chip communication. Nevertheless, merely metal-based NoC implementation offers only limited performance and power scalability in terms of multicast and broadcast traffics. To meet scalability demands, this paper addresses the system-level challenges for intra-chip multicast communication in a proposed hybrid interconnects architecture. This hybrid NoC combines and utilizes both regular metal on-chip interconnects and new type of wireless-NoC (WiNoC) which is Zenneck surface wave interconnects (SWI). Moreover, this paper embeds novel multicast routing and arbitration schemes to address system-level multicast-challenges in the proposed architecture. Specifically, a design exploration of contention handling in SWI layer is considered in both centralized and decentralized manners. Consequently, the hybrid wire-SWI architecture avoids overloading the network, alleviates the formation of traffic hotspots and avoid deadlocks that are typically associated with state-of-the-art multicast handling. The evaluation is based on a cycle-accurate simulation and hardware description. It demonstrates the effectiveness of the proposed architecture in terms of power consumption (up to around $10$ x) and performance (around $22$ x) compared to regular NoCs. These results are achieved with negligible hardware overheads. This study explore promising potential of the proposed architecture for current and future NoC-based many-core processors.

A Novel Low Complexity Logic Encryption Technique for Design-for-Trust

Abstract: The outsourcing of several untrusted intellectual property designs makes the development of integrated circuits vulnerable to piracy, overbuilding, reverse engineering, and hardware Trojan (HT). To thwart the hardware-based attacks, several design-for-trust (DFT) techniques are reported in the literature. In this paper, we analyze that the existing methods are inefficient to prevent hardware Trojan attacks and also exhibit significant design overhead. Hence, we propose a novel DFT technique that effectively increases the immunity of design against the Trojan attack while requiring minimal overhead. In the proposed technique, various new light-weight key-gate topologies are proposed that effectively increase the triggering probability of Trojan at the rare-triggered nets by encrypting the design. We also propose a new encryption algorithm that identifies an optimal node using a new metric called vulnerability factor in-order to replace it with the proposed key-gate. Our encryption algorithm significantly reduces the number of vulnerable-nets with minimum replacements. The simulation results show that the proposed key-gates reduce on an average 34.2 percent and 35.1 percent per-gate area and energy respectively over the stack-based key-gates. Finally, our DFT technique gives on an average 94 percent area overhead reduction as compared to the best known DFT technique on the ISCAS-85 benchmarks.