Search

In this work, we empirically derive the scheduler's behavior under concurrent workloads for NVIDIA's Pascal, Volta, and Turing microarchitectures. In contrast to past studies that suggest the scheduler uses a round-robin policy to assign thread blocks to streaming multiprocessors (SMs), we instead find that the scheduler chooses the next SM based on the SM's local resource availability. We show how this scheduling policy can lead to significant, and seemingly counter-intuitive, performance degradation; for example, a decrease of one thread per block resulted in a 3.58X increase in execution time for one kernel in our experiments. We hope that our work will be useful for improving the accuracy of GPU simulators and aid in the development of novel scheduling algorithms.

With maximum atom-utilization efficiency, single atom catalysts (SACs) are surging as a new research frontier in catalysis science. However, fabricating SACs and maintaining their thermodynamic stability remain challenging and thus uneconomical, largely due to the lack of fundamental understanding of the formation and stabilization mechanisms of single atoms (SAs). Through systematic density functional theory (DFT) calculations and machine learning algorithms, we present a rational design guidance for the feasibility of transition metal SA formation on a defective carbon surface and the oxygen reduction reaction (ORR) activity of the resulting SAs. We show that the dispersion of a metal nanoparticle (NP) into a stable array of SAs on a defective carbon surface is governed by the decomposition energy barrier of the NP and the diffusion barrier of SAs on a carbon surface. An intrinsic descriptor that correlates the catalytic activity of a SAC with the topological, bonding, and electronic structures of the SAC and its bonding carbon defect site is revealed. The few-shot machine learning algorithm further enables a 130 000-fold reduction of the time needed to calculate the ORR activity of SACs from DFT, and thus allows us to predict the ORR activity of SACs of all transition metals within an error of 8.33%. The results from this study offer a mechanistic and quantitative guidance for rational selection of transition metal and optimal synthesis conditions to fabricate SACs with desirable electrocatalyst activity in emerging energy applications.

Multiprincipal element alloys (MPEAs) have gained surging interest due to their exceptional properties unprecedented in traditional alloys. However, identifying an MPEA with desired properties from a huge compositional space via a cost-effective design remains a grand challenge. To address this challenge, the authors present a highly efficient design strategy of MPEAs through a coherent integration of molecular dynamics (MD) simulation, machine learning (ML) algorithms, and genetic algorithm (GA). The ML model can be effectively trained from 54 MD simulations to predict the stiffness and critical resolved shear stress (CRSS) of CoNiCrFeMn alloys with a relative error of 2.77% and 2.17%, respectively, with a 12 600-fold reduction of computation time. Furthermore, by combining the highly efficient ML model and a multi-objective GA, one can predict 100 optimal compositions of CoNiCrFeMn alloys with simultaneous high stiffness and CRSS, as verified by 100 000 ML-accelerated predictions. The highly efficient and precise design strategy can be readily adapted to identify MPEAs of other principal elements and thus substantially accelerate the discovery of other high-performance MPEA materials.

Popular big data frameworks, ranging from Hadoop MapReduce to Spark, rely on garbage-collected languages, such as Java and Scala. Big data applications are especially sensitive to the effectiveness of garbage collection (i.e., GC), because they usually process a large volume of data objects that lead to heavy GC overhead. Lacking in-depth understanding of GC performance has impeded performance improvement in big data applications. In this paper, we conduct the first comprehensive evaluation on three popular garbage collectors, i.e., Parallel, CMS, and G1, using four representative Spark applications. By thoroughly investigating the correlation between these big data applications' memory usage patterns and the collectors' GC patterns, we obtain many findings about GC inefficiencies. We further propose empirical guidelines for application developers, and insightful optimization strategies for designing big-data-friendly garbage collectors.

Executing deep-learning inference on cloud servers enables the usage of high complexity models for mobile devices with limited resources. However, pre-execution time-the time it takes to prepare and transfer data to the cloud-is variable and can take orders of magnitude longer to complete than inference execution itself. This pre-execution time can be reduced by dynamically deciding the order of two essential steps, preprocessing and data transfer, to better take advantage of on-device resources and network conditions. In this work, we present PieSlicer, a system for making dynamic preprocessing decisions to improve cloud inference performance using linear regression models. PieSlicer then leverages these models to select the appropriate preprocessing location. We show that for image classification applications PieSlicer reduces median and 99th percentile pre-execution time by up to 50.2ms and 217.2ms respectively when compared to static preprocessing methods.

Summary The oxygen evolution reaction (OER) is a critical reaction for energy-related applications, yet suffers from its slow kinetics and large overpotential. It is desirable to develop effective OER electrocatalysts, such as single-atom catalysts (SACs). Here, we demonstrate machine learning (ML)-accelerated prediction of OER overpotential of all transition metals. Based on density functional theory (DFT) calculations of 15 species of SACs, we design a topological information-based ML model to map the OER overpotentials with atomic properties of the corresponding SACs. The trained ML model not only yields remarkable prediction precision (relative error of 6.49%) but also enables a 130,000-fold reduction of prediction time in comparison with pure DFT calculation. Furthermore, an intrinsic descriptor that correlates the overpotential of an SAC with its atomic properties is revealed. The approach and results from this study can be readily applicable to screen other SACs and significantly accelerate the design of high-performance catalysts for many other reactions., Graphical abstract, Highlights • We present a topology-based machine learning (ML) approach to predict OER activity • The prediction by the ML model is of high precision (relative error of 6.49%). • The ML model is 130,000 times faster than pure density function theory calculation, Artificial Intelligence; Catalysis; Electrochemistry; Energy Materials

Stream clustering is an important data mining technique to capture the evolving patterns in real-time data streams. Today’s data streams, e.g., IoT events and Web clicks, are usually high-speed and contain dynamically-changing patterns. Existing stream clustering algorithms usually follow an online-offline paradigm with a one-record-at-a-time update model, which was designed for running in a single machine. These stream clustering algorithms, with this sequential update model, cannot be efficiently parallelized and fail to deliver the required high throughput for stream clustering.In this paper, we present DistStream, a distributed framework that can effectively scale out online-offline stream clustering algorithms. To parallelize these algorithms for high throughput, we develop a mini-batch update model with efficient parallelization approaches. To maintain high clustering quality, DistStream’s mini-batch update model preserves the update order in all the computation steps during parallel execution, which can reflect the recent changes for dynamically-changing streaming data. We implement DistStream atop Spark Streaming, as well as four representative stream clustering algorithms based on DistStream. Our evaluation on three real-world datasets shows that DistStream-based stream clustering algorithms can achieve sublinear throughput gain and comparable (99%) clustering quality with their single-machine counterparts.

Volumetric video (VV) streaming has drawn an increasing amount of interests recently with the rapid advancements in consumer VR/AR devices and the relevant multimedia and graphics research. While the resource and performance challenges in volumetric video streaming have been actively investigated by the multimedia community, the potential security and privacy concerns with this new type of multimedia have not been studied. We for the first time identify an effective threat model that extracts 3D face models from volumetric videos and compromises face ID-based authentications To defend against such attack, we develop a novel volumetric video security mechanism, namely VVSec, which makes benign use of adversarial perturbations to obfuscate the security and privacy-sensitive 3D face models. Such obfuscation ensures that the 3D models cannot be exploited to bypass deep learning-based face authentications. Meanwhile, the injected perturbations are not perceivable by the end-users, maintaining the original quality of experience in volumetric video streaming. We evaluate VVSec using two datasets, including a set of frames extracted from an empirical volumetric video and a public RGB-D face image dataset. Our evaluation results demonstrate the effectiveness of both the proposed attack and defense mechanisms in volumetric video streaming.

Smartphones have recently become a popular platform for deploying the computation-intensive virtual reality (VR) applications, such as immersive video streaming (a.k.a., 360-degree video streaming). One specific challenge involving the smartphone-based head mounted display (HMD) is to reduce the potentially huge power consumption caused by the immersive video. To address this challenge, we first conduct an empirical power measurement study on a typical smartphone immersive streaming system, which identifies the major power consumption sources. Then, we develop QuRate, a quality-aware and user-centric frame rate adaptation mechanism to tackle the power consumption issue in immersive video streaming. QuRate optimizes the immersive video power consumption by modeling the correlation between the perceivable video quality and the user behavior. Specifically, QuRate builds on top of the user's reduced level of concentration on the video frames during view switching and dynamically adjusts the frame rate without impacting the perceivable video quality. We evaluate QuRate with a comprehensive set of experiments involving 5 smartphones, 21 users, and 6 immersive videos using empirical user head movement traces. Our experimental results demonstrate that QuRate is capable of extending the smartphone battery life by up to 1.24X while maintaining the perceivable video quality during immersive video streaming. Also, we conduct an Institutional Review Board (IRB)-approved subjective user study to further validate the minimum video quality impact caused by QuRate.

Target To construct a method for generating object detection dataset based on the virtual environment. The generated dataset can be used for object detection tasks based on deep learning algorithms. Methods The procedural generation method was used to create the city's virtual environment, and also computer graphics were used for rendering and automatic labeling. Results We constructed a virtual reality environment and collected 1500 images through the virtual environment, including 1307 images containing valid vehicle and pedestrian information, and trained a deep learning model based on this dataset. Conclusions A virtual reality environment is successfully created, and the generated dataset can be used to train deep learning object detection algorithms, and the trained models can also effectively perform object detection in real world.

Infrastructure-as-a-Service (IaaS) cloud platforms rent computing resources with different cost and availability tradeoffs. For example, users may acquire virtual machines (VMs) in the spot market that are cheap, but can be unilaterally terminated by the cloud operator. Because of this revocation risk, spot servers have been conventionally used for delay and risk tolerant batch jobs. In this paper, we develop risk mitigation policies which allow even interactive applications to run on spot servers. Our System, SpotCheck is a derivative cloud platform, and provides the illusion of an IaaS platform that offers always-available VMs on demand for a cost near that of spot servers, and supports unmodified applications. SpotCheck’s design combines virtualization-based mechanisms for fault-tolerance, and bidding and server selection policies for managing the risk and cost. We implement SpotCheck on EC2 and show that it i) provides nested VMs with 99.9989 percent availability, ii) achieves upto 2-5 $\times$ cost savings compared to using on-demand VMs, and iii) eliminates any risk of losing VM state.

Online applications that serve global workload have become a norm and those applications are experiencing not only temporal but also spatial workload variations. In addition, more applications are hosting their backend tiers separately for benefits such as ease of management. To provision for such applications, traditional elasticity approaches that only consider temporal workload dynamics and assume well-provisioned backends are insufficient. Instead, in this article, we propose a new type of provisioning mechanisms—geo-elasticity, by utilizing distributed clouds with different locations. Centered on this idea, we build a system called DBScale that tracks geographic variations in the workload to dynamically provision database replicas at different cloud locations across the globe. Our geo-elastic provisioning approach comprises a regression-based model that infers database query workload from spatially distributed front-end workload, a two-node open queueing network model that estimates the capacity of databases serving both CPU and I/O-intensive query workloads and greedy algorithms for selecting best cloud locations based on latency and cost. We implement a prototype of our DBScale system on Amazon EC2’s distributed cloud. Our experiments with our prototype show up to a 66% improvement in response time when compared to local elasticity approaches.

Distributed clouds offer a choice of data center locations for providers to host their applications. In this paper, we consider distributed clouds that host virtual desktops which are then accessed by users through remote desktop protocols. Virtual desktops have different levels of latency-sensitivity, primarily determined by the actual applications running and affected by the end users’ locations. In the scenario of mobile users, even switching between 3G and WiFi networks affects the latency-sensitivity. We design VMShadow, a system to automatically optimize the location and performance of latency-sensitive VMs in the cloud. VMShadow performs black-box fingerprinting of a VM’s network traffic to infer the latency-sensitivity and employs both ILP and greedy heuristic based algorithms to move highly latency-sensitive VMs to cloud sites that are closer to their end users. VMShadow employs a WAN-based live migration and a new network connection migration protocol to ensure that the VM migration and subsequent changes to the VM’s network address are transparent to end-users. We implement a prototype of VMShadow in a nested hypervisor and demonstrate its effectiveness for optimizing the performance of VM-based desktops in the cloud. Our experiments on a private as well as the public EC2 cloud show that VMShadow is able to discriminate between latency-sensitive and insensitive desktop VMs and judiciously moves only those that will benefit the most from the migration. For desktop VMs with video activity, VMShadow improves VNC’s refresh rate by 90% by migrating virtual desktop to the closer location. Transcontinental remote desktop migrations only take about 4 min and our connection migration proxy imposes 13 μs overhead per packet.

The evaluation of teaching quality is the key point in the guarantee system of teaching quality in schools. Overall monitoring, evaluation and feedback of teachers’ teaching activities should be done from different perspectives, different levels and different channels. But the current research in this field, the objective factors that influences the teaching quality are overlooked. So this paper proposes “structure entropy weight method” combining subjective evaluating and objective evaluating based on entropy theory. It is a weight coefficient structure analysis method combining qualitative analysis and quantitative analysis with the basic idea of typical ranking by combining Delphi survey which is gathering experts’ opinions with fuzzy analysis, then calculating the entropy value of the typical ranking according to the given formula, analyzing the blindness, and dealing the data with potential variance. The structure entropy weight method provides a new method for evaluation indicators system wi...

This poster focuses on addressing the power consumption issues in 360-degree immersive video streaming on smart phones, an emerging virtual reality (VR) application in the consumer video market. We first conducted a power measurement study that indicates VR view generation as the major power consumption source. Then, we developed an edge-based immersive streaming system called EdgeVR that offloads the power-consuming view generation operation from the smart phone to the edge. Through our preliminary evaluations using EdgeVR, we identified the challenge of Motion-to-Photon latency associated with offloading. To reduce such latency, we propose a viewport prediction-based pre-rendering mechanism at the edge to ensure the quality of experience in the VR application.

Deep learning models are increasingly used for end-user applications, supporting both novel features such as facial recognition, and traditional features, e.g. web search. To accommodate high inference throughput, it is common to host a single pre-trained Convolutional Neural Network (CNN) in dedicated cloud-based servers with hardware accelerators such as Graphics Processing Units (GPUs). However, GPUs can be orders of magnitude more expensive than traditional Central Processing Unit (CPU) servers. These resources could also be under-utilized facing dynamic workloads, which may result in inflated serving costs. One potential way to alleviate this problem is by allowing hosted models to share the underlying resources, which we refer to as multi-tenant inference serving. One of the key challenges is maximizing the resource efficiency for multi-tenant serving given hardware with diverse characteristics, models with unique response time Service Level Agreement (SLA), and dynamic inference workloads. In this paper, we present Perseus, a measurement framework that provides the basis for understanding the performance and cost trade-offs of multi-tenant model serving. We implemented Perseus in Python atop a popular cloud inference server called Nvidia TensorRT Inference Server. Leveraging Perseus, we evaluated the inference throughput and cost for serving various models and demonstrated that multi-tenant model serving led to up to 12% cost reduction., Comment: 8 pages, 5 figures, and 6 tables. In proceedings of International Conference on Cloud Engineering (IC2E) 2020

Device-to-Device (D2D) communication allows user devices in proximity to directly communicate with each other through reusing resources in cellular communication system. D2D improves the system capacity while it also raises some challenges on interference. This paper adopts reconfigurable antenna in D2D communication generating directional signal transmission between D2D pairs which mitigates the interference to other users. The simulations show that such scheme improve the system capacity compared to the traditional one.

Credibility is the key performance index and the lifeline for system simulation. Verification, validation and accreditation (VV&A) of all system simulation life can insure the credibility, and credibility evaluation is a very necessary approach to obtain the credibility. Combat simulation is a special simulation for its many characteristics of multi-level, more elements, overall emergence, uncertainty, morbid and multi-granularity model, and so on. Credibility is vital to combat simulation. The evaluation method is one of the aspects of credibility evaluation. In this paper, the methods of credibility evaluation are researched. The common evaluation methods are analyzed and compared. Aiming at the characteristics of combat simulation system, the subjective comprehensive evaluation method based on subject matter expert (SME) is put forward, which is applied to a combat simulation. The practice shows that the subjective comprehensive evaluation method is very effective. The results are sufficiently credible to support the application.

Residential networks are home to increasingly diverse devices, including embedded devices that are part of the Internet of Things phenomenon, leading to new management and security challenges. However, current residential solutions that rely on customer premises equipment (CPE), which often remains deployed in homes for years without updates or maintenance, are not evolving to keep up with these emerging demands. Recently, researchers have proposed to outsource the tasks of managing and securing residential networks to cloud-based security services by leveraging software-de ned networking (SDN). However, the use of cloud-based infrastructure may have performance implications. In this paper, we measure the performance impact and perception of a residential SDN using a cloud-based controller through two measurement studies. First, we recruit 270 residential users located across the United States to measure residential latency to cloud providers. Our measurements suggest the cloud controller architecture provides 90% of end-users with acceptable performance with judiciously selected public cloud locations. When evaluating web page loading times of popular domains, which are particularly latency-sensitive, we found an increase of a few seconds at the median. However, optimizations could reduce this overhead for top websites in practice.

Matrix factorization techniques have been widely used as a method for collaborative filtering for recommender systems. In recent times, different variants of deep learning algorithms have been explored in this setting to improve the task of making a personalized recommendation with user-item interaction data. The idea that the mapping between the latent user or item factors and the original features is highly nonlinear suggest that classical matrix factorization techniques are no longer sufficient. In this paper, we propose a multilayer nonlinear semi-nonnegative matrix factorization method, with the motivation that user-item interactions can be modeled more accurately using a linear combination of non-linear item features. Firstly, we learn latent factors for representations of users and items from the designed multilayer nonlinear Semi-NMF approach using explicit ratings. Secondly, the architecture built is compared with deep-learning algorithms like Restricted Boltzmann Machine and state-of-the-art Deep Matrix factorization techniques. By using both supervised rate prediction task and unsupervised clustering in latent item space, we demonstrate that our proposed approach achieves better generalization ability in prediction as well as comparable representation ability as deep matrix factorization in the clustering task., version 3

Modern mobile applications are benefiting significantly from the advancement in deep learning, e.g., implementing real-time image recognition and conversational system. Given a trained deep learning model, applications usually need to perform a series of matrix operations based on the input data, in order to infer possible output values. Because of computational complexity and size constraints, these trained models are often hosted in the cloud. To utilize these cloud-based models, mobile apps will have to send input data over the network. While cloud-based deep learning can provide reasonable response time for mobile apps, it restricts the use case scenarios, e.g. mobile apps need to have network access. With mobile specific deep learning optimizations, it is now possible to employ on-device inference. However, because mobile hardware, such as GPU and memory size, can be very limited when compared to its desktop counterpart, it is important to understand the feasibility of this new on-device deep learning inference architecture. In this paper, we empirically evaluate the inference performance of three Convolutional Neural Networks (CNNs) using a benchmark Android application we developed. Our measurement and analysis suggest that on-device inference can cost up to two orders of magnitude greater response time and energy when compared to cloud-based inference, and that loading model and computing probability are two performance bottlenecks for on-device deep inferences., Comment: Accepted at The IEEE International Conference on Cloud Engineering (IC2E) conference 2018

Aiming at the problems in the modeling of assembly sequence planning, this paper proposes a assembly model based on the semantic and two-dimensional matrix which is easy to calculate and expression. The paper complement the parts attribute semantic and it is used to make the assembly model .So the parts about the attributes of parts assembly can be used in the ASP. Assembly relations including Assembly Precedence Relations (APRs), Contact semantic, Connection and Transmission semantic are transformed into a matrix model. Finally shaft structure is used as a example to represent a competitive advantage of this model.

Discovering and modeling lead-lag relations is a critical task in a variety of domains, including energy management, financial markets and environment monitoring. This task becomes more challenging when processing massive and highly dynamic data sources, often produced by sensors and live feeds that collect data about evolving entities in the real world. To cope with this data volume and velocity, distributed real-time computation systems have been proposed in the last years, although the problem of the lead-lag relation mining and modeling has not been deeply explored in this context. In this paper, we propose DL2-Miner, a novel distributed data mining framework for lead-lag relations based on this computational paradigm. DL2-Miner addresses the fundamental data mining task of uncovering interactions in evolving entities, and encompasses a lead-lag relation detection module with communication and computation optimization and a probabilistic model for lead-lag relation occurrence inference. It is implemented on top of the open source distributed real-time computation system Apache Storm, and preliminary experiments show promising results of our approach.

The general computer-aided design (CAD) software cannot meet the mould design requirement of the autoclave process for composites, because many parameters such as temperature and pressure should be considered in the mould design process, in addition to the material and geometry of the part. A framed-mould computer-aided design system (FMCAD) used in the autoclave moulding process is proposed in this paper. A function model of the software is presented, in which influence factors such as part structure, mould structure, and process parameters are considered; a design model of the software is established using object oriented (O-O) technology to integrate the stiffness calculation, temperature field calculation, and deformation field calculation of mould in the design, and in the design model, a hybrid model of mould based on calculation feature and form feature is presented to support those calculations. A prototype system is developed, in which a mould design process wizard is built to integrate the input information, calculation, analysis, data storage, display, and design results of mould design. Finally, three design examples are used to verify the prototype.

The dramatic rise of time-series data in a variety of contexts, such as social networks, mobile sensing, data centre monitoring, etc., has fuelled interest in obtaining real-time insights from such data using distributed stream processing systems. One such extremely valuable insight is the discovery of correlations in real-time from large-scale time-series data. A key challenge in discovering correlations is that the number of time-series pairs that have to be analyzed grows quadratically in the number of time-series, giving rise to a quadratic increase in both computation cost and communication cost between the cluster nodes in a distributed environment. To tackle the challenge, we propose a framework called AEGIS. AEGIS exploits well-established statistical properties to dramatically prune the number of time-series pairs that have to be evaluated for detecting interesting correlations. Our extensive experimental evaluations on real and synthetic datasets establish the efficacy of AEGIS over baselines.

The dramatic rise of time-series data produced in a variety of contexts, such as stock markets, mobile sensing, sensor networks, data centre monitoring, etc., has fuelled the development of large-scale distributed real-time computation systems (e.g., Apache Storm, Samza, Spark Streaming, S4, etc.). However, it is still unclear how certain time series mining tasks could be performed using such new emerging systems. In this paper, we focus on the task of efficiently discovering statistically significant correlations among a large number of time series via a distributed realtime computation engine. We propose a framework referred to as SigCO. In SigCO, we put forward a novel partition-aware data shuffling, which is able to adaptively shuffle time series data only to the relevant nodes of the distributed real-time computation engine. On the other hand, in SigCO we design a δ-hypercube structure based correlation computation approach which is capable of pruning unnecessary correlation computations. Finally, our extensive experimental evaluations on real and synthetic datasets establish that SigCO outperforms the baseline approaches in terms of diverse performance metrics.

Cloud spot markets enable users to bid for compute resources, such that the cloud platform may revoke them if the market price rises too high. Due to their increased risk, revocable resources in the spot market are often significantly cheaper (by as much as 10×) than the equivalent non-revocable on-demand resources. One way to mitigate spot market risk is to use various fault-tolerance mechanisms, such as checkpointing or replication, to limit the work lost on revocation. However, the additional performance overhead and cost for a particular fault-tolerance mechanism is a complex function of both an application's resource usage and the magnitude and volatility of spot market prices. We present the design of a batch computing service for the spot market, called SpotOn, that automatically selects a spot market and fault-tolerance mechanism to mitigate the impact of spot revocations without requiring application modification. SpotOn's goal is to execute jobs with the performance of on-demand resources, but at a cost near that of the spot market. We implement and evaluate SpotOn in simulation and using a prototype on Amazon's EC2 that packages jobs in Linux Containers. Our simulation results using a job trace from a Google cluster indicate that SpotOn lowers costs by 91.9% compared to using on-demand resources with little impact on performance.

Motivated by the emergence of distributed clouds, we argue for the need for geo-elastic provisioning of application replicas to effectively handle temporal and spatial workload fluctuations seen by such applications. We present DB Scale, a system that tracks geographic variations in the workload to dynamically provision database replicas at different cloud locations across the globe. Our geo-elastic provisioning approach comprises a regression-based model to infer the database query workload from observations of the spatially distributed front-end workload and a two-node open queueing network model to provision databases with both CPU and I/O-intensive query workloads. We implement a prototype of our DB Scale system on Amazon EC2's distributed cloud. Our experiments with our prototype show up to a 66% improvement in response time when compared to local elasticity approaches.

Infrastructure-as-a-Service (IaaS) cloud platforms rent resources, in the form of virtual machines (VMs), under a variety of contract terms that offer different levels of risk and cost. For example, users may acquire VMs in the spot market that are often cheap but entail significant risk, since their price varies over time based on market supply and demand and they may terminate at any time if the price rises too high. Currently, users must manage all the risks associated with using spot servers. As a result, conventional wisdom holds that spot servers are only appropriate for delay-tolerant batch applications. In this paper, we propose a derivative cloud platform, called SpotCheck, that transparently manages the risks associated with using spot servers for users. SpotCheck provides the illusion of an IaaS platform that offers always-available VMs on demand for a cost near that of spot servers, and supports all types of applications, including interactive ones. SpotCheck's design combines the use of nested VMs with live bounded-time migration and novel server pool management policies to maximize availability, while balancing risk and cost. We implement SpotCheck on Amazon's EC2 and show that it i) provides nested VMs to users that are 99.9989% available, ii) achieves nearly 5x cost savings compared to using equivalent types of on-demand VMs, and iii) eliminates any risk of losing VM state.

Distributed clouds offer a choice of data center locations to application providers to host their applications. In this paper we consider distributed clouds that host virtual desktops(VDs) which are then accessed by their users through remote desktop protocols. VDs have different sensitivities to latency, primarily determined by the types of applications running (games or video players are more sensitive to latency) and the end users' locations. We design VMShadow, a system to automatically optimize the location and performance of latency-sensitive VDs in the cloud. VMShadow performs black-box fingerprinting of a VM's network traffic to infer its latency-sensitivity and employs a greedy heuristic based algorithm to move highly latency-sensitive VMs to cloud sites that are closer to their end users. VMShadow employs WAN-based live migration and a new network connection migration protocol to ensure that the VM migration and subsequent changes to the VM's network address are transparent to end-users. We implement a prototype of VMShadow in a nested hypervisor and demonstrate its effectiveness for optimizing the performance of VM-based desktops in the cloud. Our experiments on a private and the public EC2 cloud show that VMShadow is able to discriminate between latency-sensitive and insensitive desktop applications and judiciously move only those VMs that will benefit the most. For desktop VMs with video activity, VMShadow improves VNC's refresh rate by 90%. Further our connection migration proxy, which utilizes dynamic rewriting of packet headers, imposes a rewriting overhead of only 13μs per packet. Trans-continental VM migrations take about 4 minutes.

As various kinds of sensors penetrate our daily life (e.g., sensor networks for environmental monitoring, GPS for localization and navigation), the efficient management of massive amount of sensor data becomes increasingly important at present. Many sensor data management systems are implemented based on key-value stores in the cloud; the traditional solutions based on relational database lack scalability to accommodate the large-scale sensor data efficiently. Meanwhile, model-view sensor data management, which stores the sensor data in the form of modelled segments, largely reduces the amount of raw data. However, currently there is no index and query optimizations on these modelled segments in the cloud, which results in full table scan for query processing in the worst case. In this paper, we propose an innovative model index for sensor data segments in key-value stores (KVM-index). KVM-index consists of two interval indices on the time and sensor value dimensions respectively, each of which has an in-memory search tree and a secondary list materialized in the key-value store. This in-memory and key-value composite structure enables to update new incoming sensor data segments with constant network I/O. Second, for time (or value)-range and point queries a MapReduce-based approach is designed to process the discrete predicate-related ranges of the table of KVM-index, thereby eliminating computation and communication overheads incurred by accessing irrelevant parts of the index table in conventional MapReduce programs. Finally, we propose a cost based adaptive strategy for the KVM-index-MapReduce framework to process composite queries on both time and value dimensions. As proved by extensive experiments in a private cloud, our approach outperforms in query response time both MapReduce-based processing of the raw sensor data and multiple alternative approaches of querying model-view sensor data.

We present VMShadow, a system that automatically optimizes the location and performance of applications based on their dynamic workloads. We prototype VMShadow and demonstrate its efficacy using VM-based desktops in the cloud as an example application. Our experiments on a private cloud as well as the EC2 cloud, using a nested hypervisor, show that VMShadow is able to discriminate between location-sensitive and location-insensitive desktop VMs and judiciously moves only those that will benefit the most from the migration. For example, VMShadow performs transcontinental VM migrations in ~ 4 mins and can improve VNC's video refresh rate by up to 90%.

The simulation model of the super-mild hybrid electrical vehicle is established through the simulation software ADVISOR. The fuzzy logic torque distribution controller of motor and engine is designed. The drive cycle selects Urban Dynamometer Driving Schedule (UDDS). The simulation results show that the fuzzy torque controller of motor and engine can properly distribute the torque. The fuel economy and emission performance are improved.

This paper builds plans for emergency management of forestry enterprises, and provide a new effective use emergency document for forestry enterprises to respond to emergenciesway based on the knowledge element linking technology, because the emergency management of forestry enterprises can not be quickly and efficiently and find emergency document the problem. Through the document structured analysis, extracted emergency document knowledge element. RDF to integrate the knowledge element analysis, the semantics associated with the completion of the extraction of the relationship between the knowledge element. Finally, the knowledge element linking design plans for emergency management of forestry enterprises. The one hand, the results of this study can guide forestry enterprises emergency practice, on the other hand has laid a foundation for the visual design of the emergency management of forestry enterprises

Aimed at problems of large amoumt of calculation and too many charts existed in the field of machine design, a kind of neural-fuzzy reasoning system was applied to carry out mapping discrimination in order to raise computing speed and processing ability of programming and eliminate the phenomenon of “combined explosion” of knowledge. It is shown by means of practical application to concrete data and charts that the relation of primary charts and data can be nicely expressed by the use of this kind of method. At the same time, this method has more high precission and more quick speed on data recognition and can satisfy the requirement of designing computation when compared with the method of using merely artificial neural net.

At present, cooperation between manufacturing enterprises is becoming important. Many manufacturing enterprises establish their own manufacturing resource domain ontologies. In order to achieve knowledge reuse and sharing among different manufacturing resource domain ontologies, the method of manufacturing resource domain ontology integration was introduced. Three cases-equivalence, parallel, and intersection were proposed. Based on analyzing the characteristics and functions of manufacturing resource domain ontology, a new concept similarity computation algorithm was introduced for the manufacturing resource domain ontology. Both information content-based similarity computation and distance -based similarity computation were considered in this algorithm. At the end, an instance about integration of two part of different manufacturing resource domain ontologies was shown.

The Occupant Protection of Vehicle Side Impact (GB20071-2006) is a compulsive and executable standard that all M1 and N1 style vehicle must obey to. Simulation for vehicle side impact is one of important test methods to study vehicle impact safety. And the model of moving deformable barrier (MDB) is the basic tool of the test. According to the requirements in regulation for the performance and test verification method of MDB, the finite element model of MDB is built and the simulation test is carried out to verify the correctness of the model. And the simulation results show the model is accurate and effective.

Quality of organization decision making can be increased greatly by group decision, and more and more managers begin pay attention to this method. But there are still some problems existed in group decision, such as in some case, experts' ability, attitude and confidence will greatly affect decision result, and how to find abnormal expert and reduce his decision weight or dismiss him from the decision group is very important for increasing decision quality. For the reason above, this paper tries to find an effective method for avoiding abnormal experts' negative effect. And build a model for recognize abnormal experts based on support vector domain description, and the model take the experts' decision activities as input and the model will automatically find out the outlier expert according his abnormal decision activities.

The similarity computation for domain ontology is a critical problem. On the basis of analysis characteristics and functions of manufacturing resource domain ontology, a similarity computation algorithm for the manufacturing resource domain ontology is introduced. It depends on the process capability of manufacturing resource and is quite easier to understand. Then, two applications of this similarity computation algorithm are described, which are manufacturing resource domain ontology integration and virtual manufacturing cell (VMC) search. Both of them are quite helpful for the manufacturing scheduling system.

With the popularization of networks, digitalization and automation, demand for enterprise informatization becomes more urgent. There are many factors leading to the demand for EIS. Some of the factors are from enterprise development, while others are from policy driven. In this paper, we present a relationship model by using system dynamics method to characterize the cause-result (C-R) of demand drivers of enterprise informatization. Based on the empirical studies, we reveal how the factors affect the demand for enterprise informatization, which form a cluster of causation to be used as the cause variables in our model. This procedure is to settle on the interim variables and result variables, which formed systematic dynamics C-R charts. Questionnaires and interviews from dozens of enterprises, Delphi Expert Decision method are made and analyzed, which verified the relationship between variables. The results presented in this paper provide good insights for the enterprise managers’ optimal decisions.

Infinite nature of sensor data poses a serious challenge for query processing even in a cloud infrastructure. Model-based sensor data approximation reduces the amount of data for query processing, but all modeled segments need to be scanned, in the worst case. In this paper, we propose an innovative index for modeled segments in key-value stores, namely KVI-index. KVI-index has an in-memory tree component and a secondary structure materialized in the key-value store that maps the tree nodes to the modeled data segments. Then, we introduce a KVI-index-Scan-MapReduce hybrid approach to perform efficient query processing. As proved by a series of experiments in a real private cloud infrastructure, our approach outperforms in query response time and index updating efficiency both Hadoop-based parallel processing of the raw sensor data and multiple alternative indexing approaches of model-view data.

As the number of sensors that pervade our lives increases (e.g., environmental sensors, phone sensors, etc.), the efficient management of massive amount of sensor data is becoming increasingly important. The infinite nature of sensor data poses a serious challenge for query processing even in a cloud infrastructure. Traditional raw sensor data management systems based on relational databases lack scalability to accommodate large-scale sensor data efficiently. Thus, distributed key-value stores in the cloud are becoming a prime tool to manage sensor data. Model-view sensor data management, which stores the sensor data in the form of modeled segments, brings the additional advantages of data compression and value interpolation. However, currently there are no techniques for indexing and/or query optimization of the model-view sensor data in the cloud; full table scan is needed for query processing in the worst case. In this paper, we propose an innovative index for modeled segments in key-value stores, namely KVI-index. KVI-index consists of two interval indices on the time and sensor value dimensions respectively, each of which has an in-memory search tree and a secondary list materialized in the key-value store. Then, we introduce a KVI-index-Scan-MapReduce hybrid approach to perform efficient query processing upon modeled data streams. As proved by a series of experiments at a private cloud infrastructure, our approach outperforms in query-response time and index-updating efficiency both Hadoop-based parallel processing of the raw sensor data and multiple alternative indexing approaches of model-view data.

The availability of massive volumes of data and recent advances in data collection and processing platforms have motivated the development of distributed machine learning algorithms. In numerous real-world applications large datasets are inevitably noisy and contain outliers. These outliers can dramatically degrade the performance of standard machine learning approaches such as regression trees. To this end, we present a novel distributed regression tree approach that utilizes robust regression statistics, statistics that are more robust to outliers, for handling large and noisy data. We propose to integrate robust statistics based error criteria into the regression tree. A data summarization method is developed and used to improve the efficiency of learning regression trees in the distributed setting. We implemented the proposed approach and baselines based on Apache Spark, a popular distributed data processing platform. Extensive experiments on both synthetic and real datasets verify the effectiveness and efficiency of our approach. The data and software related to this paper are available at https://github.com/weilai0980/DRSquare_tree/tree/master/.

With increasing availability of mobile sensing devices including smartphones, online mobile data segmentation becomes an important topic in reconstructing and understanding mobile data. Traditional approaches like online time series segmentation either use a fixed model or only apply an adaptive model on one dimensional data; it turns out that such methods are not very applicable to build online segmentation for multiple dimensional mobile sensor data (e.g., 3D accelerometer or 11 dimension features like ‘mean’, ‘vari- ance’, ‘covariance’, ‘magnitude’, etc). In this paper, we design an adaptive model for segment- ing real-time accelerometer data from smartphones, which is able to (a) dynamically select suitable dimensions to build a model, and (b) adaptively pick up a proper model. In addition to using the traditional residual-style regression errors to evaluate time series segmentation, we design a rich metric to evaluate mobile data segmentation results, including (1) traditional regression error, (2) Information Retrieval style measurements (i.e., precision, recall, F-measure), and (3) segmentation time delay.

Time series forecasting for streaming data plays an important role in many real applications, ranging from IoT systems, cyber-networks, to industrial systems and healthcare. However the real data is often complicated with anomalies and change points, which can lead the learned models deviating from the underlying patterns of the time series, especially in the context of online learning mode. In this paper we present an adaptive gradient learning method for recurrent neural networks (RNN) to forecast streaming time series in the presence of anomalies and change points. We explore the local features of time series to automatically weight the gradients of the loss of the newly available observations with distributional properties of the data in real time. We perform extensive experimental analysis on both synthetic and real datasets to evaluate the performance of the proposed method.