Recent Submissions

  • Blue Superluminescent Diodes with GHz Bandwidth Exciting Perovskite Nanocrystals for High CRI White Lighting and High-Speed VLC

    Alatawi, Abdullah A.; Holguin Lerma, Jorge Alberto; Kang, Chun Hong; Shen, Chao; Dursun, Ibrahim; Sinatra, Lutfan; Albadri, Abdulrahman M.; Alyamani, Ahmed Y.; Ng, Tien Khee; Bakr, Osman; Ooi, Boon S. (Conference on Lasers and Electro-Optics, OSA, 2019-05-08) [Conference Paper]
    A 442-nm GaN-based superluminescent diode (SLD) is demonstrated with a GHz modulation bandwidth and a linewidth of 7 nm. When use for exciting CsPbBr3-perovskite nanocrystal-phosphor, warm-white light with a high CRI of 91 was achieved.
  • A Low Power Hardware Implementation of Izhikevich Neuron using Stochastic Computing

    Ismail, Aya A.; Shaheen, Zeinab A.; Rashad, Osama; Salama, Khaled N.; Mostafa, Hassan (2018 30th International Conference on Microelectronics (ICM), IEEE, 2019-05-02) [Conference Paper]
    This paper introduces the hardware implementation of one of the most popular spiking neuron models which is Izhikevich model. The main target of this implementation is to reduce area and power consumed by the Spiking Neural Network (SNN) neurons as the SNN consists of a large number of neurons to mimic the human brain. Therefore, stochastic computing techniques are used to perform the squaring term that consumes much of the power in the Izhikevich neuron model equations. A hardware implementation of the model is proposed to show the area and power consumption to help the SNN designers to choose between stochastic-based multipliers and the approximate multipliers considering their power, area, and accuracy constraints.
  • A Simple, Easy to Fabricate Miniaturized Microfluidic Gradient Generator for Drug Testing Devices

    Sivashankar, Shilpa; Alamoudi, Kholod; Agambayev, Sumeyra; Mashraei, Yousof; Mkaouar, Hend; Khashab, Niveen M.; Salama, Khaled N. (2018 30th International Conference on Microelectronics (ICM), IEEE, 2019-05-02) [Conference Paper]
    To date, although microfabrication technologies for fabricating a microfluidic device are advanced, they are still time-consuming and laborious. Hence, we demonstrate the fabrication of microfluidic devices with a fast and easy maskless Ultraviolet (UV) projection method based on a stereolithography process in less than 5 mins. The flow model analysis by COMSOL gives the design concept of the gradient demonstrated. The fabricated chip is a miniaturized 25×25 mm2 gradient chip that produces gradient by maintaining equal width and length of each channel throughout the device. The design of the gradient is dependent on diffusion of molecules and hence is well suited for low flow rate applications such as drug delivery or cell related studies. The biocompatibility of the resins in their native form and with surface modification was evaluated by injecting cell culture medium to culture Human cervical cell line (HeLa) cells. Drug (Doxorubicin) screening was demonstrated by the viability of HeLa cells using Cell Counting Kit-8 (CCK-8) calorimetric assay. The miniaturized size of the chip aids these gradient generators to find applications in drug testing Lab-on-chip/Micro Total analysis systems (μTAS) and organ-on-chip devices.
  • Moving Bits with a Fleet of Shared Virtual Routers

    Kathiravelu, Pradeeban; Chiesa, Marco; Marcos, Pedro; Canini, Marco; Veiga, Luis (2018 IFIP Networking Conference (IFIP Networking) and Workshops, IEEE, 2019-04-26) [Conference Paper]
    The steady decline of IP transit prices in the past two decades has helped fuel the growth of traffic demands in the Internet ecosystem. Despite the declining unit pricing, bandwidth costs remain significant due to ever-increasing scale and reach of the Internet, combined with the price disparity between the Internet's core hubs versus remote regions. In the meantime, cloud providers have been auctioning underutilized computing resources in their marketplace as spot instances for a much lower price, compared to their on-demand instances. This state of affairs has led the networking community to devote extensive efforts to cloud-assisted networks - the idea of offloading network functionality to cloud platforms, ultimately leading to more flexible and highly composable network service chains.We initiate a critical discussion on the economic and technological aspects of leveraging cloud-assisted networks for Internet-scale interconnections and data transfers. Namely, we investigate the prospect of constructing a large-scale virtualized network provider that does not own any fixed or dedicated resources and runs atop several spot instances. We construct a cloud-assisted overlay as a virtual network provider, by leveraging third-party cloud spot instances. We identify three use case scenarios where such approach will not only be economically and technologically viable but also provide performance benefits compared to current commercial offerings of connectivity and transit providers.
  • SeqST-ResNet: A Sequential Spatial Temporal ResNet for Task Prediction in Spatial Crowdsourcing

    Zhai, Dongjun; Liu, An; Chen, Shicheng; Li, Zhixu; Zhang, Xiangliang (Database Systems for Advanced Applications, Springer International Publishing, 2019-04-23) [Conference Paper]
    Task appearance prediction has great potential to improve task assignment in spatial crowdsourcing platforms. The main challenge of this prediction problem is to model the spatial dependency among neighboring regions and the temporal dependency at different time scales (e.g., hourly, daily, and weekly). A recent model ST-ResNet predicts traffic flow by capturing the spatial and temporal dependencies in historical data. However, the data fragments are concatenated as one tensor fed to the deep neural networks, rather than learning the temporal dependencies in a sequential manner. We propose a novel deep learning model, called SeqST-ResNet, which well captures the temporal dependencies of historical task appearance in sequences at several time scales. We validate the effectiveness of our model via experiments on a real-world dataset. The experimental results show that our SeqST-ResNet model significantly outperforms ST-ResNet when predicting tasks at hourly intervals and also during weekday and weekends, more importantly, in regions with intensive task requests.
  • A Multicopter Design Software Tool for Automated Generation of Simulation and Visualization Models

    Shaqura, Mohammad; Shamma, Jeff S. (Informatics in Control, Automation and Robotics, Springer International Publishing, 2019-04-18) [Conference Paper]
    Multicoptor unmanned aerial vehicles (UAVs) are popular robotics platforms in various research and applications fields. Research in robotics, control, estimation and computer vision relies heavily on open-source software and hardware to build custom UAV. This is motivated by lower cost of material and rapid development desire. The presented tool automates the task of obtaining realistic models for simulation and visualization of multicoptors using state-of-the-art Computer Aided Design engineering tools (CAD). Users interact with the software through a desktop application that offers interface to CAD tools, hardware database and simulation files generation. Custom models can be generated for three popular multirotor simulators. Modeling parameters accuracy has been validated using data of IRIS+ quadcopter model.
  • Joint Scheduling and Beamforming via Cloud-Radio Access Networks Coordination

    Douik, Ahmed; Dahrouj, Hayssam; Al-Naffouri, Tareq Y.; Alouini, Mohamed-Slim (2018 IEEE 88th Vehicular Technology Conference (VTC-Fall), IEEE, 2019-04-15) [Conference Paper]
    Cloud radio access network (CRAN) emerges as a promising architecture for large-scale interference management. This paper addresses the benefit of one particular type of coordinated resource allocation in CRANs through the combined effect of joint scheduling and beamforming. Consider the downlink of a CRAN where the cloud is connected to several remote radio heads (RRHs), each equipped with multiple antennas. The transmit frame of every RRH is formed by several radio resource blocks (RRBs), each capable of serving multiple single-antenna users via spatial multiplexing using beamforming. The paper focuses on the problem of maximizing the network-wide weighted sum-rate by jointly determining the set of scheduled users at each RRB, and their corresponding beamforming vectors. The main contribution of the paper is to solve such a mixed discrete-continuous optimization problem using a graph-theoretical based approach. The paper introduces the joint scheduling and beamforming graph, wherein each independent set accounts for a feasible schedule and feasible beamforming vectors. Afterward, the joint scheduling and beamforming problem is shown to be equivalent to a maximum independent set problem in the proposed graph. Simulation results suggest that the proposed joint solution provides appreciable performance improvements as compared to the classical iterative approach.
  • Reproducibility in Benchmarking Parallel Fast Fourier Transform based Applications

    Aseeri, Samar; Muite, Benson K.; Takahashi, Daisuke (Companion of the 2019 ACM/SPEC International Conference on Performance Engineering - ICPE '19, ACM Press, 2019-04-05) [Conference Paper]
    An overview of concerns observed in allowing for reproducibility in parallel applications that heavily depend on the three dimensional distributed memory fast Fourier transform are summarized. Suggestions for reproducibility categories for benchmark results are given.
  • A Novel Framework for Node/Edge Attributed Graph Embedding

    Sun, Guolei; Zhang, Xiangliang (Advances in Knowledge Discovery and Data Mining, Springer International Publishing, 2019-04-04) [Conference Paper]
    Graph embedding has attracted increasing attention due to its critical application in social network analysis. Most existing algorithms for graph embedding utilize only the topology information, while recently several methods are proposed to consider node content information. However, the copious information on edges has not been explored. In this paper, we study the problem of representation learning in node/edge attributed graph, which differs from normal attributed graph in that edges can also be contented with attributes. We propose GERI, which learns graph embedding with rich information in node/edge attributed graph through constructing a heterogeneous graph. GERI includes three steps: construct a heterogeneous graph, take a novel and biased random walk to explore the constructed heterogeneous graph and finally use modified heterogeneous skip-gram to learn embedding. Furthermore, we upgrade GERI to semi-supervised GERI (named SGERI) by incorporating label information on nodes. The effectiveness of our methods is demonstrated by extensive comparison experiments with strong baselines on various datasets.
  • HCCI Octane Number Scale in a Pressure-Temperature Diagram

    Masurier, Jean-Baptiste; Elkhazraji, Ali; Mohammed, Abdulrahman; Johansson, Bengt (SAE Technical Paper Series, SAE International, 2019-04-02) [Conference Paper]
    A new approach for investigating combustion behavior of practical fuels under homogeneous charge compression ignition (HCCI) conditions was developed with the help of a cooperative fuel research (CFR) engine. The method uses a set of two pressure-temperature diagrams and two charts, each with an octane number scale based on primary reference fuels (PRF), created from experimental results by sweeping the intake temperature. The two pressure-temperature diagrams report conditions leading to the start of the low temperature combustion and the start of the main combustion, respectively. Additional two charts - required compression ratio and fraction of low temperature heat release charts - describe global combustion behavior and the importance of the low temperature combustion. Each diagram and chart, together with their respective octane number scale, allow to examine the combustion behavior of practical fuels by comparing their combustion behavior with those of the PRFs. Finally, octane numbers representing the various combustion behaviors of a practical fuel can be rated. Application of the method with two low-octane number surrogate fuels led to the following main results. The required compression ratio chart provides a quick description of the combustion behavior. The pressure-temperature diagrams indicate the ease with which a fuel ignites under low temperature combustion and main combustion regimes. An extra pressure-temperature diagram reports start and end of the negative temperature coefficient regime and highlights that this regime is independent of the fuel. Accordingly, each combustion regime is clearly defined in the pressure-temperature diagram. The fraction of low temperature heat release finally indicates how low temperature combustion vanishes. Finally, octane numbers for each practical fuel were rated from each diagram and chart. Rated octane numbers suggest that a single PRF cannot reflect the entire combustion behavior of a practical fuel; but multiple PRFs are required for HCCI combustion.
  • Mechanism Triggering Pre-Ignition in a Turbo-Charged Engine

    Singh, Eshan; Dibble, Robert W. (SAE Technical Paper Series, SAE International, 2019-04-02) [Conference Paper]
    Pre-ignition in modern engines is largely attributed to oil-fuel mixture droplets igniting before the spark timing. Researchers have also found pre-ignition events to be triggered by high hydrocarbon emissions from the previous cycle as well as late spark timing in the previous cycle. Additionally, an ideally scavenged engine was not found to be limited by pre-ignition. These observations point to a significant role of residuals in triggering pre-ignition events. Current work studies pre-ignition in a probabilistic approach. The effect of residuals and in-cylinder thermodynamic state is studied by varying the exhaust back pressure and intake air temperature respectively. Experiments were performed with a fixed mass flow rate of air + fuel and intake air temperature while the exhaust back pressure was varied. Intake air pressure varied in response to fixed intake temperature. Pre-ignition and super-knock count increased with increasing exhaust back pressure. In the next set of experiments, mass flow rate of air + fuel and intake air pressure were fixed, while the exhaust back pressure was varied. Intake air temperature was varied to fix the intake air pressure constant. Pre-ignition counts generally increased with increasing intake temperature, although the exhaust back pressure decreased. Number of super-knock cycles correlated directly with intake air temperature. Conclusively, the current study shows that probability of a pre-ignition event relies on (a) the likelihood of precursor generation (from fuel impinging the liner), (b) the likelihood of precursors being held back in cylinder (related to exhaust back pressure) and (c) the reactivity of bulk mixture (related to in-cylinder temperature).
  • Effect of Different Fluids on Injection Strategies to Suppress Pre-Ignition

    Singh, Eshan; Hlaing, Ponnya; Shi, Hao; Dibble, Robert W. (SAE Technical Paper Series, SAE International, 2019-04-02) [Conference Paper]
    Pre-ignition is an abnormal engine combustion phenomenon where the inducted fuel-air charge ignites before the spark ignition. This premature combustion phenomenon often leads to heavy knocking events. The mixture preparation plays a critical role in pre-ignition tendency for a given load. Literature shows efforts made towards improving pre-ignition-limited-IMEP by splitting the injection pulse into multiple pulses. In this study, two direct injectors are used in a single cylinder research engine. A centrally mounted direct injector was used to inject Coryton Gasoline (RON 95) fuel early in the intake stroke. A second fluid was injected late in the compression stroke to suppress pre-ignition. The fluids used in the second direct injector was varied to see the effects of the molecule and its physical and chemical property on pre-ignition suppression tendency. Methanol, ethanol, water, and gasoline were tested as second fluid. Engine tests were conducted at 2000 rpm and at an intake pressure of 2.1 bar (abs). Although alcohols show high pre-ignition tendency as fuels, they were most effective at pre-ignition suppression when injected later in the compression stroke. The pre-ignition suppression led to a decrease in IMEP and an increase in cycle-to-cycle variation. Water injection was highly effective at maintaining peak IMEP values. Water injection was further explored for pre-ignition suppression. The water injection helped reduce pre-ignition count when injected at two different injection times each in intake, compression and late exhaust stroke.
  • Auto-windowed Super-virtual Interferometry via Machine Learning: A Strategy of First-arrival Traveltime Automatic Picking for Noisy Seismic Data

    Lu, Kai; Feng, Shihang (SEG 2018 Workshop: SEG Maximizing Asset Value Through Artificial Intelligence and Machine Learning, Beijing, China, 17-19 September 2018, Society of Exploration Geophysicists and the Chinese Geophysical Society, 2019-04-02) [Conference Paper]
    Supervirtual interferometry (SVI) was developed to significantly enhance the signal-to-noise ratio of noisy first arrivals. However, a time window must be specified that contains these first arrivals, and the window should be no wider than several times the dominant period of the source wavelet. The accurate specification of this window is very challenging for noisy data and involves manual picking. To overcome this problem, we propose to automatically pick these windows via machine learning methods. Convolutional neural network (CNN) and density-based spatial clustering of applications with noise (DBSCAN) are used to distinguish first-arrival signals completely buried in noise. Numerical tests validate that this method can accurately specify the correct window as well as that of a human interpreter. The benefit is an automatic means for picking first-arrival traveltimes in noisy traces from a large 3D data set.
  • Automatic Semblance Picking by a Bottom-up Clustering Method

    Chen, Yuqing (SEG 2018 Workshop: SEG Maximizing Asset Value Through Artificial Intelligence and Machine Learning, Beijing, China, 17-19 September 2018, Society of Exploration Geophysicists and the Chinese Geophysical Society, 2019-04-02) [Conference Paper]
    Semblance picking is an important but tedious labor-intensive processing procedure in the petroleum industry. For a large 3D dataset, this task becomes extremely time-consuming. In this paper, we present an automatic semblance picking technique based on the K-means clustering algorithm. K-means clustering method can automatically partition different clusters of energy in the semblance spectrum into different groups. The centroid of each group is the automatically picked semblance point. A synthetic and field data example is shown in this paper to illustrate the effectiveness of this method.
  • Combustion Stratification and Dynamic Flame Tracing Analysis of Partially Premixed Combustion in a Compression Ignition Engine Fueled with Low-Octane Fuel

    An, Yanzhao; Shi, Hao; Vallinayagam, R; Sim, Jaeheon; Chang, Junseok; Johansson, Bengt (SAE Technical Paper Series, SAE International, 2019-04-02) [Conference Paper]
    Partially premixed combustion (PPC) is a low-temperature combustion concept, which is between conventional diesel compression ignition (CI) and homogeneous charge compression ignition (HCCI). In PPC mode, the start of injection timing (SOI) is earlier than that of CI and later than that of HCCI and stratified in-cylinder fuel/air mixture can be formed to control the auto-ignition by the fuel injection timing. Gasoline fuel is beneficial for PPC mode because of its superior resistance to auto-ignition, which can enhance fuel-air charge mixing process with longer ignition delay time. The scope of this study is to investigate in-cylinder auto-ignition, combustion evolution, combustion stratification, and engine-out emissions at PPC operating mode under lean and low load engine conditions with different injection timings. Primary reference fuel PRF77, was selected as the low-octane test fuel. Fuel-tracer planar laser-induced fluorescence (PLIF) imaging and high-speed color imaging based on natural flame luminosity were performed to visualize fuel injection, spray-wall interaction, and subsequent combustion evolution. Based on the intensity of high sped combustion images, combustion stratification and dynamic flame tracing were evaluated to gain insights into the combustion evolution. Combustion stratification analysis indicated that more inhomogeneous low-temperature combustion was achieved at earlier fuel injection timings along with decreased natural flame luminosity and increased soot emission. Fuel-trapping in piston crevice zone was visualized by fuel-tracer PLIF. Fuel-trapping in squish zone and crevice zone was measured and linked to the formation of unburned hydrocarbon when stronger spray-wall interaction occurs under PPC operating mode. Injector dribbling during the late stage of combustion was found to be as an important source of soot formation through high-speed color imaging and dynamic flame tracing analysis.
  • The Physical and Chemical Effects of Fuel on Gasoline Compression Ignition

    Vallinayagam, R.; Hlaing, Ponnya; AlRamadan, Abdullah S.; An, Yanzhao; Sim, Jaeheon; Chang, Junseok; Johansson, Bengt (SAE Technical Paper Series, SAE International, 2019-04-02) [Conference Paper]
    In the engine community, gasoline compression ignition (GCI) engines are at the forefront of research and efforts are being taken to commercialize an optimized GCI engine in the near future. GCI engines are operated typically at Partially Premixed Combustion (PPC) mode as it offers better control of combustion with improved combustion stability. While the transition in combustion homogeneity from convectional Compression Ignition (CI) to Homogenized Charge Compression Ignition (HCCI) combustion via PPC has been comprehensively investigated, the physical and chemical effects of fuel on GCI are rarely reported at different combustion modes. Therefore, in this study, the effect of physical and chemical properties of fuels on GCI is investigated. In-order to investigate the reported problem, low octane gasoline fuels with same RON = 70 but different physical properties and sensitivity (S) are chosen. Fuels with comparable sensitivity and RON are chosen to study the impact of physical properties on GCI. On the other hand, by keeping the same RON and physical properties, the effect of sensitivity on GCI is investigated. In this regard, three test fuels such as RON 70 gasoline (S=0.7), PRF 70 (S=0) and RON 70 gasoline (S=7) are chosen in the present study. Herein, RON 70 gasoline (S=0.7) and PRF 70 have similar RON and sensitivity but different physical properties; however, RON 70 gasoline (S=0.7) and RON 70 gasoline (S=7) have the same RON and physical properties but different sensitivity. These test fuels were tested in a heavy-duty CI engine at a compression ratio of 17.8 under different combustion modes. The experimental investigation reveals that RON 70 gasoline (S=0.7) and PRF 70 (S=0) behaves the same in terms of combustion behavior (combustion phasing, ignition delay, in cylinder pressure and rate of heat release) regardless of the difference in physical properties. While nitrogen oxide (NOX) and soot emissions are comparable between RON 70 gasoline (S=0.7) and PRF 70 at all combustion modes, the hydrocarbon (HC) and carbon monoxide (CO) emissions are slightly higher for PRF 70 when compared to RON 70 gasoline (S=0.7) at HCCI mode but not at PPC and CI modes due to the impact of physical properties. On the other hand, due to higher sensitivity, the reactivity for RON 70 gasoline (S=7) is improved to advance the combustion phasing at HCCI combustion mode when compared to RON 70 gasoline (S=0.7). At HCCI mode, the HC emissions are lower for high sensitive gasoline when compared to low sensitive gasoline whereas they are comparable at PPC and CI combustion modes. The NOX and soot emissions are comparable at HCCI modes whereas high sensitivity gasoline shows slightly decreased NOX and increased soot emissions, respectively, at PPC and CI combustion modes when compared to low sensitive gasoline.
  • Should We Walk or Take a Car for Minimum Greenhouse Gas Emissions?

    Babayev, Rafig; Johansson, Bengt (SAE Technical Paper Series, SAE International, 2019-04-02) [Conference Paper]
    This paper compares the greenhouse gas (GHG) emissions attributed to driving a popular production vehicle powered by an internal combustion engine (ICE), as well as a hybrid electric vehicle (HEV), with GHG emissions associated with walking, running and bicycling. The purpose of this study is to offer a different perspective on the problem of global warming due to anthropogenic causes, specifically on transportation and eating patterns. In order to accurately estimate emissions, a full life cycle of food has been considered coupled with energy expenditures of the aforementioned activities obtained from several different sources and averaged for more reliable results. The GHG emissions were calculated for Sweden, the UK, and the US. Depending on the availability of certain data, the methodology for different countries was altered slightly. The question whether walking, running or taking a bicycle is better for the environment than driving a car cannot be answered uniquely. This study demonstrates that the answer depends on several factors, such as diet composition, the number of people commuting, vehicle powertrain, as well as the country analyzed. The conclusion is that if one has an eco-friendly diet and travels alone the preferred modes of transport would be bicycling, walking and running, the cleanest of which by far is bicycling. However, if the diet has a higher CO2 footprint, as in the case of diets containing a large amount of meat and/or imported products, then the preference shifts towards cars, among which the most environmentally friendly are hybrid electric vehicles. The same conclusion applies to the cases where the number of people commuting together exceeds two-three persons.
  • Efficient Assimilation of Crosswell Electromagnetic Data Using Ensemble-Based History-Matching Framework

    Zhang, Yanhui; Hoteit, Ibrahim (SPE Reservoir Simulation Conference, Society of Petroleum Engineers, 2019-03-27) [Conference Paper]
    An ensemble-based history-matching framework is proposed to enhance the characterization of petroleum reservoirs through the assimilation of crosswell electromagnetic (EM) data. As one of advanced technologies in reservoir surveillance, crosswell EM tomography can provide a cross-sectional conductivity map and hence saturation profile at an interwell scale by exploiting the sharp contrast in conductivity between hydrocarbons and saline water. Incorporating this new information into reservoir simulation in combination with other available observations is therefore expected to enhance the forecasting capability of reservoir models and to lead to better quantification of uncertainty. The proposed approach applies ensemble-based data-assimilation methods to build a robust and flexible framework under which various sources of available measurements can be readily integrated. Because the assimilation of crosswell EM data can be implemented in different ways (e.g., components of EM fields or inverted conductivity), a comparative study is conducted. The first approach integrates crosswell EM data in its original form which entails establishing a forward model simulating observed EM responses. In this work, the forward model is based on Archie's law that provides a link between fluid properties and formation conductivity, and Maxwell’s equations that describe how EM fields behave given the spatial distribution of conductivity. Alternatively, formation conductivity can be used for history matching, which is obtained from the original EM data through inversion using an adjoint gradient-based optimization method. Because the inverted conductivity is usually of high dimension and very noisy, an image-oriented distance parameterization utilizing fluid front information is applied aiming to assimilate the conductivity field efficiently and robustly. Numerical experiments for different test cases with increasing complexity are carried out to examine the performance of the proposed integration schemes and potential of crosswell EM data for improving the estimation of relevant model parameters. The results demonstrate the efficiency of the developed history-matching workflow and added value of crosswell EM data in enhancing the characterization of reservoir models and reliability of model forecasts.
  • Matrix Algebra Framework for Portable, Scalable and Efficient Query Engines for RDF Graphs

    Jamour, Fuad Tarek; Abdelaziz, Ibrahim; Chen, Yuanzhao; Kalnis, Panos (Proceedings of the Fourteenth EuroSys Conference 2019 CD-ROM on ZZZ - EuroSys '19, ACM Press, 2019-03-22) [Conference Paper]
    Existing query engines for RDF graphs follow one of two design paradigms: relational or graph-based. We explore sparse matrix algebra as a third paradigm and propose MAGiQ: a framework for implementing SPARQL query engines that are portable on various hardware architectures, scalable over thousands of compute nodes, and efficient for very large RDF datasets. MAGiQ represents the RDF graph as a sparse matrix and defines a domain-specific language of algebraic operations. SPARQL queries are translated into matrix algebra programs that are oblivious to the underlying computing infrastructure. Existing matrix algebra libraries, optimized for each particular architecture, are called to execute the program and handle the performance issues. We present three case studies of matrix algebra back-end libraries: SuiteSparse, Matlab, and CombBLAS; we demonstrate how MAGiQ can effortlessly be ported on a variety of architectures such as Intel CPUs, NVIDIA GPUs, and Cray XC40 supercomputers. Our experiments on large-scale real and synthetic datasets show that MAGiQ performs comparably to or better than existing specialized SPARQL query engines for data-intensive queries, scales to very large computing infrastructures, and handles datasets with up to 512 billion triples.
  • Flexible Design of Millimeter-Wave Cache Enabled Fog Networks

    Emara, Mostafa; ElSawy, Hesham; Sorour, Sameh; Al-Ghadhban, Samir; Alouini, Mohamed-Slim; Al-Naffouri, Tareq Y. (2018 IEEE Globecom Workshops (GC Wkshps), IEEE, 2019-03-18) [Conference Paper]
    Ultra-densification, millimeter wave (mmW) communications, and proactive network-edge caching, utilized within mmW fog networks (mmFNs), are foreseen to provide tangible gains for broadband access, network capacity, and latency. However, caching implementation in mmFN imposes high capital expenditure (CAPEX) due to the ultra-high density of base stations (BSs). For a given caching CAPEX, it may be more efficient to install higher capacity caches in a fraction of the BSs than installing smaller capacity caches in every BSs. In the former case, wireless self-backhauling of mmW systems can be exploited to share the cache contents stored in a given cache enabled BSs (CE-BSs) with other BSs in the network. In this regards, this paper develops a mathematical model, based on stochastic geometry, to study the tradeoff between the cache size and intensity of CE-BSs on the probability that requested popular contents are retrieved from the network edge, denoted as the hit probability. Assuming a power-law inverse relationship between the cache size and intensity of CE-BSs, an optimization problem is formulated and solved for the intensity of CE-BSs and probabilistic file placement in caches such that the hit probability is maximized. The results show that neither installing small caches in every BS nor having sufficiently high capacity caches (i.e., that confine all popular files) installed in small number of BSs exploit the full potential of mmFN. Instead, there exists an optimal balance between the cache size and intensity of CE-BSs, which depends on the network parameters such as the applied caching strategy, required rate, total intensity of BSs, popular content distribution, and cache size/intensity relationship.

View more