Search

This work proposes a sub-optimal method based on a two-layer structured meta-deep reinforcement learning (MDRL) approach to address the hardware impairment (HWI) optimization issue in large intelligent surface (LIS) systems. This method, designed for distributed LIS systems with reflection matrices, effectively enhances the system capacity and performance despite HWIs. Building upon existing techniques of dividing large-area LIS systems into multiple small-area subsystems, the simulated results demonstrate that sub-optimal LIS performance can be achieved with fewer samples in diverse dynamic wireless environments. This innovative approach enhances the adaptability of distributed LIS systems and offers an effective HWI management strategy, paving the way for future LIS system optimization.

Transient simulation in power engineering is crucial as it models the dynamic behavior of power systems during sudden events like faults or short circuits. Electromagnetic transient simulations involve multiple coordinated tasks. Traditional simulations are centralized and struggle to meet scalability requirements. To achieve these goals, distributed electromagnetic transient simulation has emerged as a new trend. Nevertheless, the distributed electromagnetic transient simulation introduces network communication. Achieving real-time simulation across distributed nodes poses the challenge of minimizing communication costs. In this paper, our proposal focuses on optimizing the task orchestration to reduce communication costs. Specifically, in the electromagnetic transient simulation, these tasks has certain communication pattern where the communicated objects of each task are pre-defined. We represent the pattern as a graph, with tasks represented as nodes and communications as edges. Furthermore, we propose to use graph partition with the objective of minimal communication costs and fine tune the partitions with the resource requirements of each distributed node. The experimental results demonstrate that our proposal has strength in achieving high-performance electromagnetic transient simulation.

Hyper-Dimensional (HD) computing is a brain-inspired learning approach for efficient and fast learning on today’s embedded devices. HDC first encodes all data points to high-dimensional vectors called hypervectors and then efficiently performs the classification task using a well-defined set of operations. Although HDC achieved reasonable performances in several practical tasks, it comes with huge memory requirements since the data point should be stored in a very long vector having thousands of bits. To alleviate this problem, we propose a novel HDC architecture, called StrideHD. By utilizing the window striding in image classification, StrideHD enables HDC system to be trained and tested using binary hypervectors and achieves high accuracy with fast training speed and significantly low hardware resources. StrideHD encodes data points to distributed binary hypervectors and eliminates the expensive Channel item Memory (CiM) and item Memory (iM) in the encoder, which significantly reduces the required hardware cost for inference. Our evaluation also shows that compared with two popular HD algorithms, the singlepass StrideHD model achieves a 27.6 $\times$ and 8.2 $\times$ reduction in inference memory cost without hurting the classification accuracy, while the iterative mode further provides 8.7 $\times$ memory efficiency. Under the same inference memory cost, our single-pass mode StrideHD averagely achieves 13.56% accuracy improvement in comparison with the single-pass baseline HD, which is a similar performance even in comparison with the costly iterative baseline HD models. As an extension, the iterative retraining mode of StrideHD averagely provides 11.33% accuracy improvement to its single-pass mode, which can be accomplished in fewer iterations in comparison with the baseline HD algorithms. The hardware implementation also demonstrates that StrideHD achieves over 9.9 $\times$ and 28.8 $\times$ reduction compared with baseline in area and power, respectively.

Distance relays are critical components in protection systems of power grids that can be attacked by cyber-attackers. Indeed, a cyber-attacker injects fake data into a distance relay to pretend a fault has happened, and the distance relay must be tripped. Thus, a new powerful approach, named Multi-Agent Distributed Deep Learning (MADDL) method is proposed to tackle cyber-attacks in distance relays. Unlike centralized methods, the protection system with several distance relays is mapped to a multi-agent distributed system by employing the graph theory, in which the distance relays are considered as the agents of the multi-agent system or the nodes of the considered graph. Each agent is only connected to the neighboring agents to exchange voltage and current data. Then, a deep neural network as a cyber-attack detection structure is assumed for each agent that utilizes the local voltage and current data and the received data from the neighboring agents to detect the attacks. Hence, the considered detection structures are tuned by employing train data, obtained by simulating the grid in different types of faults. Then, the tuned detection structures are evaluated by a test dataset, including data from the grid under various faults and the normal situation by injecting fake data as cyber-attacks. The developed method has been employed for three different case studies, including IEEE 6-bus, IEEE 14-bus, and IEEE 118-bus power grids. According to the simulation results, the proposed algorithm has succeeded in identifying more than 99.88% of faults and cyber-attacks.

The split computing approach, where the head and tail models are respectively distributed between the IoT device and cloud, suffers from high network latency especially when the cloud is located far from the IoT device. To mitigate this problem, we introduce a distributed split computing system (DSCS) where an IoT device (called split computing requester) broadcasts a split computing request to its neighboring IoT devices. After receiving the request, the neighboring IoT devices (i.e., requestees) distributively determine whether or not to accept the split computing request by taking into account the unnecessary energy consumption and computation time. To minimize energy consumption while maintaining a specified probability of on-time computing completion, we develop a constrained stochastic game model. Then, a best-response dynamics-based algorithm is used to obtain the Nash equilibrium. The evaluation results demonstrate that the DSCS consumes can reduce more than 20% energy consumption compared to a probabilistic-based acceptance scheme, where the IoT devices accept a split computing request based on a predefined probability, while providing high on-time computing completion probability.

As a specific category of artificial intelligence (AI), generative artificial intelligence (GenAI) generates new content that resembles what humans create. The rapid development of GenAI systems has created a huge amount of new data on the Internet, posing new challenges to current computing and communication frameworks. Currently, GenAI services rely on the traditional cloud computing framework due to the need for large computation resources. However, such services will encounter high latency because of data transmission and a high volume of user requests. On the other hand, edge-cloud computing can provide adequate computation power and low latency at the same time through the collaboration between edges and the cloud. Thus, it is attractive to build GenAI systems at scale by leveraging the edge-cloud computing paradigm. In this overview paper, we review recent developments in GenAI and edge-cloud computing, respectively. Then, we use two exemplary GenAI applications to discuss technical challenges in scaling up their solutions using edge-cloud collaborative systems. Finally, we list design considerations for training and deploying GenAI systems at scale and point out future research directions.

The pre-trained model (PTM) is revolutionizing Artificial Intelligence (AI) technology. However, the hardware requirement of PTM training is prohibitively high, making it a game for a small proportion of people. Therefore, we proposed PatrickStar system to lower the hardware requirements of PTMs and make them accessible to everyone. PatrickStar uses the CPU-GPU heterogeneous memory space to store the model data. Different from existing works, we organize the model data in memory chunks and dynamically distribute them in the heterogeneous memory. Guided by the runtime memory statistics collected in a warm-up iteration, chunks are orchestrated efficiently in heterogeneous memory and generate lower CPU-GPU data transmission volume and higher bandwidth utilization. Symbiosis with the Zero Redundancy Optimizer, PatrickStar scales to multiple GPUs on multiple nodes. The system can train tasks on bigger models and larger batch sizes, which cannot be accomplished by existing works. Experimental results show that PatrickStar extends model scales 2.27 and 2.5 times of DeepSpeed, and exhibits significantly higher execution speed. PatricStar also successfully runs the 175B GPT3 training task on a 32 GPU cluster. Our code is available at https://github.com/Tencent/PatrickStar. [ABSTRACT FROM AUTHOR]

Eventual consistency is a consistency model that favors liveness over safety. It is often used in large-scale distributed systems where models ensuring a stronger safety incur performance that are too low to be deemed practical. Eventual consistency tends to be uniformly applied within a system, but we argue a demand exists for differentiated eventual consistency, e.g. in blockchain systems. We propose update-query consistency with primaries and secondaries (UPS) to address this demand. UPS is a novel consistency mechanism that works in pair with our novel two-phase epidemic broadcast protocol gossip primary-secondary (GPS) to offer differentiated eventual consistency and delivery speed. We propose two complementary analyses of the broadcast protocol: a continuous analysis and a discrete analysis based on compartmental models used in epidemiology. Additionally, we propose the formal definition of a scalable consistency metric to measure the consistency trade-off at runtime. We evaluate UPS in two simulated worldwide settings: a one-million-node network and a network emulating that of the Ethereum blockchain. In both settings, UPS reduces inconsistencies experienced by a majority of the nodes and reduces the average message latency for the remaining nodes. [ABSTRACT FROM AUTHOR]

The smart city system, which is a type of enterprise management system (EMS), automatically manages cities and schedules resources efficiently based on spatial data generated by devices, such as the Internet of Things and mobile. However, with the increasing deployment of technologies, including sensor and location-based services, their ever-growing spatial data are no longer managed efficiently by traditional EMS. To overcome this issue, we present SeFrame, which is a spatially enabled framework for improving the efficiency of smart city EMS based on a distributed architecture. The framework supports a set of spatial queries, including: The range query, k-nearest neighbors query, and spatial join query. It benefits greatly from using the buffer-enabled partition method to eliminate duplicate results. In each partition, the local index based on combination of the quad-tree and grid index (CQG) significantly improves the spatial query efficiency in memory. CQG manages complex spatial objects, including a point, polygon, and polyline. By taking full advantage of the local index, SeFrame accesses skewed spatial data in constant time. In experiments, we demonstrated that the proposed method delivered superior performance in terms of scalability and query efficiency, in most cases. [ABSTRACT FROM AUTHOR]

Deduplication technology can obtain higher space utilization by keeping only one duplicate. But in a distributed storage system, the overall deduplication ratio will be limited due to redundancy elimination across nodes. The traditional deduplication methods usually utilize data similarity and data locality to improve the deduplication ratio. However, higher system overhead is caused by frequent similarity calculations. To deal with this problem, this paper proposes a new Feature-Aware Stateful Routing method (FASR), aiming to reduce the system overhead and keep a high deduplication ratio in the distributed environment. Firstly, we design a feature-aware nodes selection strategy to choose similar nodes by extracting data feature and data distribution characteristics. This strategy will save the similarity calculation with the nodes that are not similar to the data. Then, we present a stateful routing algorithm to determine the target node using super-chunk and handprint technology. Meanwhile, the algorithm maintain load balance of the entire distributed system. Finally, the data is deduplicated locally based on similarity index and fingerprint cache. Extensive experiments demonstrate that FASR can reduce system overhead by around 30% at most and also effectively obtain a higher deduplication ratio.

With the widespread use of mobile and sensing devices, and the popularity of online map-based services, such as navigation services, the volume of spatio-temporal data is growing rapidly. Conventional big data technologies in existing distributed systems cannot effectively process spatio-temporal big data with temporal continuity and spatial proximity. How to construct an effective index for the application requirements of spatio-temporal data in a distributed environment has become one of the hotspots of spatio-temporal big data research. Many spatio-temporal indexing methods have been proposed to support efficient query processing of spatio-temporal data. In this article, the various spatio-temporal big data indexing methods proposed by domestic and foreign researchers from 2010 to 2020 are classified and summarized according to the distributed environment and application background, and the hot issues that need to be paid attention to in the future are proposed according to the changes in application requirements

By means of a simple buffer design, we show that active versus passive and push versus pull settings in asynchronous communication protocols, also known as handshake protocols, can be controlled by initialization. We advocate postponing initialization until run time and show that postponement simplifies the design and design process and serves test, debug, and analysis. We design the buffer as a network of communication channels with storage, called Links, and storage-free computation modules, called Joints. We describe the behaviors of Links and Joints using a shared variable model presented here for the first time. [ABSTRACT FROM AUTHOR]

To improve the accuracy and efficiency of 3D LiDAR mapping, real-time cooperative SLAM has been considered to explore large and complex areas. To merge the individual maps from multiple robots, it is crucial to identify the common areas and obtain alternative matches between them. However, data transmission, especially in sparse networks with narrow bandwidth and limited range, is a challenging issue for the above problem. Since the distribution manner is suitable for limited communication, we proposed a common framework of 3D real-time distributed cooperative SLAM to fill the community gap. Assuming that each robot can communicate with others, the presented framework consists of four key modules: place recognition, relative pose estimation, distributed graph optimization, and communication. Meanwhile, we developed a complete real-time distributed cooperative SLAM system, called RDC-SLAM, by integrating state-of-the-art components into the framework. For computation and data transmission efficiency, descriptor-based registration is used instead of the conventional point cloud matching. An intensity-based descriptor is developed to perform the place recognition and obtain the alternative matches, while an eigenvalue-based segment descriptor is applied to further refine the relative pose estimations between these alternative matches. A distributed graph optimization method is utilized to obtain the maximum likelihood of multi-trajectory estimation. A communication protocol is also designed to associate data among robots that are easy to deploy and have low network requirements. The RDC-SLAM is validated by real-world experiments and exhibits superior performance concerning accuracy, computation efficiency, and data efficiency. [ABSTRACT FROM AUTHOR]

For the rapid growth computation requirements in big data and artificial intelligence area, CPU-GPU heterogeneous clusters can provide more powerful computing capacity compared to CPU clusters. The high parallel computing capabilities of GPUs greatly accelerate computation-intensive applications. And the number of GPUs on single computing node is scalable, which greatly improves the computing capacity of the cluster under the condition of limited cluster size. However, there is a lack of the effective load-balancing scheduling model in multi-GPU hardware environment. This article proposes AEML, an acceleration engine for multi-GPU load-balancing in distributed heterogeneous environment. AEML can effectively integrate GPUs into distributed processing framework and achieve great load-balance among multiple heterogeneous GPUs. We propose a heterogeneous task execution model based on multiple GPUs and multiple streams (MGMS), which can effectively balance the workload of multiple GPUs. MGMS model utilizes four core techniques: a fine-grained task mapping mechanism, a device resource unified management scheme, a novel resource-aware GPU task scheduling strategy, and a feedback-based streams adjustment scheme. The implementation of AEML system is based on Spark 3.0.0 and NVIDIA CUDA 10.0. We comprehensively evaluate the performance of AEML with multiple typical benchmarks. Experimental results show that AEML can fully exploit the computing power of GPUs and achieve great load-balance among multiple heterogeneous GPUs. [ABSTRACT FROM AUTHOR]

The mathematical properties of generative adversarial networks (GANs) are presented via opinion dynamics, in which the discriminator is regarded as an agent while the generator is regarded as the principal (or another type of agent) in a GAN. In some cases, the overall convergence must be achieved through establishing a local information interaction between one agent and its neighbors via the multiagent system consistency theory and algorithms. So the goals of the multiagent consistency algorithm and the Nash equilibrium of GANs are virtually identical. Then, the existence of the Degroot model solution proves that the generator and discriminator in GANs can reach a consensus on the distribution function. Therefore, a new sufficient and necessary condition for the existence of a Nash equilibrium in GANs is obtained. Furthermore, a novel multiagent distributed GAN (MADGAN) is proposed to address the multiagent cognitive consistency problem in large-scale distributed network, based on the social group wisdom and the influence of the network structure on the agent. The nodes of a multiagent network are regarded as discriminators and generators, the discriminator with a relatively large degree of influence as a leader, and the generator as a follower, the lines with the degree denote the influence between the agents. The conditions of consensus are presented for a multigenerator and multidiscriminator (multiagent) distributed GAN by analyzing the existence of stationary distribution to the Markov chain of multiple agent states. Finally, theoretical results are verified with simulation based on the MNIST dataset. [ABSTRACT FROM AUTHOR]

Recent advances on Edge computing, Network Function Virtualization (NFV) and 5G are stimulating the interest of the industrial sector to satisfy the stringent and real-time requirements of their applications. Digital Twin is a key piece in the industrial digital transformation and its benefits are very well studied in the literature. However, designing and implementing a Digital Twin system that integrates all the emerging technologies and meets the connectivity requirements (e.g., latency, reliability) is an ambitious task. Therefore, prototyping the system is required to gradually validate and optimize Digital Twin solutions. In this work, an Edge Robotics Digital Twin system is implemented as a prototype that embodies the concept of Digital Twin as a Service (DTaaS). Such system enables real-time applications such as visualization and remote control, requiring low-latency and high reliability. The capability of the system to offer potential savings by means of computation offloading are analyzed in different deployment configurations. Moreover, the impact of different wireless channels (e.g., 5G, 4G and WiFi) to support the data exchange between a physical device and its virtual components are assessed within operational Digital Twins. Results show that potentially 16% of CPU and 34% of MEM savings can be achieved by virtualizing and offloading software components in the Edge. In addition, they show that 5G connectivity enables remote control of 20 ms, appearing as the most promising radio access technology to support the main requirements of Digital Twin systems.

Convolutional dictionary learning (CDL) estimates shift invariant basis adapted to represent signals or images. CDL has proven useful for image denoising or inpainting, as well as for pattern discovery on multivariate signals. Contrarily to standard patch-based dictionary learning, patterns estimated by CDL can be positioned anywhere in signals or images. Optimization techniques consequently face the difficulty of working with extremely large inputs with millions of pixels or time samples. To address this optimization problem, we propose a distributed and asynchronous algorithm, employing locally greedy coordinate descent and a soft-locking mechanism that does not require a central server. Computation can be distributed on a number of workers which scales linearly with the size of the data. The parallel computation accelerates the parameter estimation and the distributed setting allows our algorithm to be used with data that do not fit into a single computer's RAM. Experiments confirm the theoretical scaling properties of the algorithm. This allows to demonstrate an improved pattern recovery as images grow in size, and to learn patterns on images from the Hubble Space Telescope containing tens of millions of pixels. [ABSTRACT FROM AUTHOR]

While the Internet is open by design, it is still the case that users can be subject to censorship by governments or enterprises in accessing Web services and data. In this paper we propose SEnD, a fully-distributed censorship circumvention system built upon an overlay, where users have peer-to-peer virtual private IP tunnels to proxies within their social network. With SEnD, users in an uncensored area can act as proxy servers for their social friends in a censored area, allowing them to bypass the censorship. SEnD is able to outperform the current censorship techniques, such as IP address blocking and active probing attacks. We assessed the effectiveness of SEnD through extensive simulations based on a synthetic dataset, as well as experiments based on a prototype implementation. We built our synthetic dataset based on parameters obtained from questionnaires administered both inside and outside China (we consider China as a case study of censorship area). The results show that SEnD is feasible, efficient and scalable. For example, when the proportion of concurrent active users is less than 60, 99.9 percent of these users are able to find proxy servers. [ABSTRACT FROM AUTHOR]

This paper presents an interfaced wind energy conversion system (WECS) using the switched reluctance generator (SRG). A bidirectional DC-DC converter is adopted as an interface stage to achieve optimal excitation voltage at the SRG side, while a voltage source converter regulates the DC-bus voltage. The SRG works with superior performance, which is characterized by a compromise between efficiency and torque ripple, using the optimal excitation parameters. In addition, the interface stage adds a filtering behavior, mitigating the voltage ripple produced by the SRG, thus improving the power quality reflected in the grid current. A comparison of optimized SRG systems, with direct connection and with interfaced connection, is conducted. Experimental results show that the interface converter improves the performance of the SRG in WECS. Although the extra converter efficiency is approximately 88%, the overall system efficiency is only reduced by up to 2.5% due to the SRG optimal parameters. Nonetheless, as advantages of the method, the torque oscillations are reduced by up to 9.5% in high speeds. Moreover, due to the converter filtering behavior, the grid current harmonic content is reduced by up to 6% in the comparison. Therefore, this research indicates that the proposed approach is a viable option for SRG in WECS, especially for scenarios where the DC-DC converter is already necessary, such as in a microgrid application. [ABSTRACT FROM AUTHOR]

The distributed monitoring framework is undoubtedly more suitable for large-scale complex industrial systems. However, most existing distributed monitoring methods ignored the information interaction between the local system and its neighbors. In this article, an improved distributed fault detection framework that considering the communication between subsystems is present. The system decomposition is optimized based on the monitoring performance with mechanism knowledge as constraints. The integration of mechanism and data is helpful to find the appropriate common variables between subsystems. The distributed partial least squares (DPLSs) algorithm is proposed to address the local monitoring challenges caused by the propagation of a common variable. The local monitoring model takes full advantage of the information from neighbors to reduce the uncertainty of the local system. Bayesian fusion performance metrics strategy is implemented to detect system status. The simulation results of the Tennessee Eastman process verify the effectiveness of the proposed scheme. Note to Practitioners—This article attempted to tackle an issue derived from distributed process monitoring of industrial processes. Even in an era of big industrial data, the fusion idea of process data and mechanism knowledge also provides a solution to the process decomposition monitoring strategy. It reduces the computational complexity, corrects the misdirection caused by the false information hidden in the measurements, and further increases the monitoring accuracy. Considering the information flowing and spreading along with the process equipment, common variables are used to describe the interaction between different subsystems. Then, the pretrained monitoring model and the online monitoring strategy are given to promote automatic implementation. The operability and monitoring accuracy of the proposed method is verified. It is suitable for process monitoring of large-scale complex industrial systems. [ABSTRACT FROM AUTHOR]

In task offloading systems, it is imperative to guarantee that an offloaded task is completed within a pre-specified deadline. In this paper, we propose a distributed device-to-device (D2D) offloading system (DDOS) in which a task owner opportunistically broadcasts an offloading request that includes its mobility level and task completion deadline. After receiving the request, mobile devices in the vicinity of the task owner employ a constraint stochastic game to decide, in a distributed manner, whether to accept the request or not. We devise a best response dynamics-based algorithm (BRDA) to obtain a multi-policy constrained Nash equilibrium. Evaluation results demonstrate that DDOS can guarantee a high on-time task completion probability, as well as a low energy consumption. [ABSTRACT FROM AUTHOR]

To protect user privacy and meet law regulations, federated (machine) learning is obtaining vast interests in recent years. The key principle of federated learning is training a machine learning model without needing to know each user's personal raw private data. In this article, we propose a secure matrix factorization framework under the federated learning setting, called FedMF. First, we design a user-level distributed matrix factorization framework where the model can be learned when each user only uploads the gradient information (instead of the raw preference data) to the server. While gradient information seems secure, we prove that it could still leak users' raw data. To this end, we enhance the distributed matrix factorization framework with homomorphic encryption. We implement the prototype of FedMF and test it with a real movie rating dataset. Results verify the feasibility of FedMF. We also discuss the challenges for applying FedMF in practice for future research. [ABSTRACT FROM AUTHOR]

We present new results for uplink distributed MIMO with two-stage beamforming (BF). Two- and four-cluster base station layouts are considered with four BF strategies. Stage one considers analog (A) or digital (D) maximal ratio combining (MRC), with digital MRC or digital zero forcing (ZF) in stage two. We denote the resulting BF schemes as A-D, D-D, A-ZF and D-ZF. Using new analytical results for the expected signal-to-interference-plus-noise-ratio (SINR) of A-D and D-D, we show that the performance gap between these schemes increases with the number of clusters. Using novel simplifications for the A-ZF and D-ZF SINR, we show that the interference reduction of the first stage processing causes a delay of the MRC/ZF performance cross-over relative to a single-stage system. We analyze the asymptotics of A-D and D-D strategies, extending existing results for single-stage co-located processing to a distributed two-stage system. We show how the relative performance of A-D and D-D varies significantly depending on the distribution of interference among the base stations and that A-D can outperform D-D. [ABSTRACT FROM AUTHOR]

IoT devices have some special characteristics, such as mobility, limited performance, and distributed deployment, which makes it difficult for traditional centralized access control methods to support access control in current large-scale IoT environment. To address these challenges, this paper proposes an access control system in IoT named fabric-iot, which is based on Hyperledger Fabric blockchain framework and attributed based access control (ABAC). The system contains three kinds of smart contracts, which are Device Contract (DC), Policy Contract (PC), and Access Contract (AC). DC provides a method to store the URL of resource data produced by devices, and a method to query it. PC provides functions to manage ABAC policies for admin users. AC is the core program to implement an access control method for normal users. Combined with ABAC and blockchain technology, fabric-iot can provide decentralized, fine-grained and dynamic access control management in IoT. To verify the performance of this system, two groups of simulation experiments are designed. The results show that fabric-iot can maintain high throughput in large-scale request environment and reach consensus efficiently in a distributed system to ensure data consistency.

Blockchain technology is an emerging distributed ledger technology that has exploratory applications in many areas. The consensus algorithm, as the core module of the blockchain, has an important impact on the security, scalability, and efficiency of the blockchain network. The consensus algorithm is also a popular topic in current blockchain technology research. In existing consortium blockchains and public blockchains, the blockchain has low efficiency or poor fault tolerance because of the limitations of the consensus algorithm. To both ensure the fault tolerance of the blockchain and improve the scalability and throughput, in this paper, we propose a new type of consensus algorithm: mixed Byzantine fault tolerance (MBFT). MBFT uses sharding and layered technology. MBFT functionally partitions the nodes that participate in the consensus process and improves the scalability and efficiency without sacrificing security. MBFT also introduces a random node selection mechanism and a credit mechanism to improve security and fault tolerance. We analyze the security and experiment on transaction throughput in a real network environment. The results prove that MBFT has good security and scalability and high throughput.

Distributed ledger technology faces scalability problems due to a long commit time despite recent successes for cryptocurrency. Small group consensus studies have improved this scalability of distributed ledgers. However, they still have problems of the consensus process itself. For example, most blockchain systems perform serialized block proposal and consensus processing, guarantee the finality with high overhead, and handle byzantine nodes inefficiently. To address these problems, we propose a consensus system, named graph learning byzantine fault tolerance (GL BFT), which offers high parallelism and low latency under Byzantine fault. To do this, we enable a parallel pipelined agreement by separating the block proposal and the consensus process. Second, we devise two techniques of merging blocks and commit learning to guarantee the finality with little overhead. Finally, we present a path learning approach which chooses optimal paths to handle Byzantine fault. The proposed GL BFT can achieve instant finality with low message overhead among a small group of nodes even if Byzantine nodes exit. Also, we evaluate its performance on an open source blockchain protocol. Experimental results show that our design reduces data traffic required by the consensus up to 30%, one transaction is finalized within a few seconds, and optimal performance is maintained.

Big data refers to numerous forms of complex and large datasets which need distinctive computational platforms in order to be analyzed. Hadoop is one of the popular frameworks for analytics of big data. In Hadoop, a big job is split into multiple small tasks and then they are distributed to worker nodes in a parallel way using MapReduce to speed up computational processes. In this aspect, it is important how to improve throughput performance. MapReduce jobs require quick responses from the worker nodes to complete them under their deadlines. The existing scheduling schemes for Hadoop such as FIFO, fair, and capacity schedulers cannot guarantee the quick response requirement satisfying a prior deadline. Thus, Hadoop system needs to improve response time and completion time for the heterogeneous MapReduce jobs. In this paper, we propose an efficient preemptive deadline constraint scheduler based on least slack time and data locality. In order for better allocation of tasks and load balancing, we first analyze the task scheduling behaviors of the Hadoop platform. Based on that, we propose a novel preemptive approach which considers the remaining execution time of the job being executed in deciding preemption. The experimental results show that the proposed scheme significantly reduces the job execution time and queue waiting time, compared to existing schemes.

The existing distributed operation schemes for microgrids lack the ability to determine the power selling to the grid during normal operation mode and are unable to provide service reliability to critical loads, during islanded operation mode. In order to overcome these issues, in this study, we have proposed a distributed operation method for both grid-connected and islanded modes of microgrids. Unlike the existing studies, where the utility grid is considered as a dispatchable generator, the bi-directional flow of power with the grid is considered in this study. Similarly, different load agents are considered for different priority loads to assure the service reliability to the critical loads during islanding. A two-step operation method is proposed for both grid-connected and islanded mode operations. During the first step, each agent in the network shares information with its neighboring agents to determine the total load and available renewable power in the network. Whereas, in the second step, each agent in the network determines the optimal operation points based on the local information received from the neighboring agents. Moreover, a modified cost function for the battery is also proposed in this study, which utilizes the information of market price and load to enhance the battery operation. A comparison is made between the centralized method, conventional distributed method, and the proposed distributed operation method. Simulation results have proved the effectiveness of the proposed method for realizing distributed operation for microgrids in both grid-connected and islanded modes.

In the microservice system, microservices are generally deployed as many microservice instances connected by a network system, and there are several composite services (CSs) in the system, each of which consists of several tasks that need to be executed on instances and has some Quality of Service (QoS) constraints to represent the user requirement for QoS. At runtime, it is necessary to select the microservice instance for each task under constraints and form an execution path of the CS. However, traditional service selection methods focus on the QoS optimization of a single CS, while ignoring sharing and competition of instances. To address this problem, this paper presents a Microservice Service Selection algorithm (MSS) based on list scheduling. Firstly, a workflow model is employed to describe the CS and to analyze the processing speed of instances, the transferring speed of the network and the degree of task concurrency, to calculate the sub-deadline of each task. Then, according to the sub-deadline and other information, the scheduling urgency of each task is calculated and updated in real-time. Finally, two service selection strategies based on the sub-deadline and urgency are proposed to complete the microservice instances selection and constitute the service path. Experiments have been carried out using the standard workflow examples of real-world applications. The results show that the proposed selection strategies can effectively improve the performance of service selection in network systems.

With the development of earth observation technologies and the construction of earth observation systems, an increasing amount of remote sensing data are being obtained. These provide the datasets required for research on remote sensing monitoring across large areas. To compensate for the shortcomings of global and large-area temporal monitoring data, synergized computing using multi-source remote sensing data can improve the accuracy and temporal resolution of remote sensing monitoring. However, remote sensing data are drawn from multiple sources and multiple scales, and have a complex structure and large volume; in addition, the nested system architecture of multi-source synergized remote sensing products makes the design of large-scale multi-source synergized remote sensing monitoring systems difficult. In this paper, we describe the design and implementation of a distributed parallel processing system for multi-source data synergized quantitative remote sensing based on a distributed cluster platform. The system integrates the algorithms normalizing more than 30 kinds of data sources and producing 40 quantitative remote sensing products. The system also connects a number of centers for satellite data, serves for several applications, and implements dynamic expansion integration for highly efficient quantitative remote sensing products. The system has produced approximately 50 TB of quantitative remote sensing products, and the application of these data to agriculture, forestry, the environment, and water conservation has resulted in very positive effects.

This study aims to develop a system of device-to-device (D2D) communications for information collection, distribution and sharing among local communities. It establishes a D2D testbed with low power consumption and solar power supply to evaluate the performance of the proposed system. Devices in the system can communicate with each other autonomously without a centralized control device when they are within a certain distance. A single device does not result in a dominant effect for the whole system, which ensures system robustness. To validate the performance of the designed system, a simulation is used to confirm the effectiveness of the system. Moreover, in cooperation with local communities, the system has been developed and deployed in some areas of Tokyo and Kyoto, Japan. The devices were deployed on buses and fixed places along the bus routes. The experimental results show that the proposed system can work with moderate delay latency. Specifically, the latency performance obtained in the field experiments meets the system design criterion quite well, which validates the feasibility and effectiveness of the proposed system. In summary, the proposed D2D local information sharing system can work independently from the cellular system and any other existing wireless network systems, which is expected to provide more proprietary and different wireless services. The paper also gives some ideas for further system improvement.

The dramatically increasing of mobile applications make it convenient for users to complete complex tasks on their mobile devices. However, the latency brought by unstable wireless networks and the computation failures caused by constrained resources limit the development of mobile computing. A popular approach to solve this problem is to establish a mobile service provisioning system based on a mobile edge computing (MEC) paradigm. In the MEC paradigm, plenty of machines are placed at the edge of the network so that the performance of applications can be optimized by using the involved microservice instances deployed on them. In this paper, we explore the deployment problem of microserivce-based applications in the MEC environment and propose an approach to help to optimize the cost of application deployment with the constraints of resources and the requirement of performance. We conduct a series of experiments to evaluate the performance of our approach. The result shows that our approach can improve the average response time of mobile services. [ABSTRACT FROM AUTHOR]

Cyber-physical systems (CPSs) are nowadays an important component of most industrial infrastructures and materialize the integration of control systems with advanced information technologies. Commonly, they aggregate distinct communication platforms and networked devices or nodes with different capabilities and goals. The resulting increase in complexity has inevitably brought into playing new challenges, namely, on the physical world and cyber space. In these systems, the development and gaining of state awareness and how to deal with the inherent vulnerabilities of the overall system are essential and also challenging. In this paper, the problem of resilience enhancement in CPSs is addressed based on a hierarchical multi-agent framework that is implemented over a distributed middleware, in which each agent carries out specific tasks. Physical and cyber vulnerabilities are taken into account, and state and context awarenesses of the whole system are targeted. The proposed framework ensures a minimum level of acceptable performance in case of physical disturbances and malicious attacks, as demonstrated by experiments on an IPv6-based test-bed.

In distributed computing systems slow working nodes, known as stragglers, can greatly extend finishing times. Coded computing is a technique that enables straggler-resistant computation. Most coded computing techniques presented to date provide robustness by ensuring that the time to finish depends only on a set of the fastest nodes. However, while stragglers do compute less work than non-stragglers, in real-world commercial cloud computing systems (e.g., Amazon’s Elastic Compute Cloud (EC2)) the distinction is often a soft one. In this paper, we develop hierarchical coded computing that exploits the work completed by all nodes, both fast and slow, automatically integrating the potential contribution of each. We first present a conceptual framework to represent the division of work amongst nodes in coded matrix multiplication as a cuboid partitioning problem. This framework allows us to unify existing methods and motivates new techniques. We then develop three methods of hierarchical coded computing that we term bit-interleaved coded computation (BICC), multilevel coded computation (MLCC), and hybrid hierarchical coded computation (HHCC). In this paradigm, each worker is tasked with completing a sequence (a hierarchy) of ordered subtasks. The sequence of subtasks, and the complexity of each, is designed so that partial work completed by stragglers can be used, rather than ignored. We note that our methods can be used in conjunction with any coded computing method. We illustrate this by showing how we can use our methods to accelerate all previously developed coded computing techniques by enabling them to exploit stragglers. Under a widely studied statistical model of completion time, our approach realizes a 66% improvement in the expected finishing time. On Amazon EC2, the gain was 27% when stragglers are simulated. [ABSTRACT FROM AUTHOR]

Software-defined storage (SDS) is an abrupt technology that aims to address the explosive data proliferation and storage management complexity through the separation of control and data path, which are tightly bounded in the traditional storage area network (SAN) model. In SDS, the centralized controller manages and controls the overall storage services and operations. However, the centralized controller approach raises the scalability and single point of failure flag. In such a paradigm, the controller placement is an important activity. A random placement of controllers suffers from unpredictable delay in a large enterprise. Multiple SDS controllers mitigate the single point of failure and provide redundancy. We propose a method called controller placement based on the center of gravity (CPCG) to solve the optimal placement controllers’ locations to reduce the delay. CPCG incorporates different coordinates and analyzes their density (the number of attached devices to a switch) to suggest new optimal sites. For the least number of SDS controllers, we propose an areawise optimal placement location strategy; the areawise approach mitigates workload processing from the central controller and also provides a scalable and reliable storage system. Experiments show that the optimal placement drastically reduces delay and, hence, boosts scalability. [ABSTRACT FROM AUTHOR]

With the development of technologies, an increasing number of wearable devices that are currently at the heart of the development of the Internet of Things are used around the world. The concerns about privacy, in particular healthcare wearable devices, are exacerbated as the requirement for self-health monitoring increases. In addition, many sources of data are geographically separated and might not be allowed to release due to patients or regulatory constraints. Accordingly, in this paper, a distributed hierarchical deep learning system is proposed. The proposed system applying a distributed hierarchical neural network over a cloud server and smartphones. The system enables multiple smartphones to train a shared consensus model collaboratively while keeping the private data locally to protect data privacy, and the system takes advantage of the abundant computational resources on the cloud server to lower the computational overhead on smartphones. The proposed system is demonstrated by a fall detection study which is the common healthcare issue among human beings. The patients’ data are collected from multiple wearable devices including the smartphone, the smartwatch, and the smart insoles. The experimental results show that the distributed hierarchical deep learning system can reproduce the accuracy, specificity, precision, and sensitivity of centralized machine learning while preserving privacy. [ABSTRACT FROM AUTHOR]

In Software Defined Networking (SDN), network programmability is enabled through a logically centralized control plane. Production networks deploy multiple controllers for scalability and reliability reasons, which in turn rely on distributed consensus protocols to operate in a logically centralized manner. However, bugs in distributed control plane can have disastrous effects on the data plane, e.g., losing traffic by installing paths containing blackholes. In this paper we study the prevalence of issues in state-of-the-art distributed frameworks in SDN, by analyzing 500+ issues reported in two of the largest open source SDN controller platforms: Open Network Operating System (ONOS) and OpenDaylight (ODL), during the period between 2014-2019. We identify system vulnerabilities, localize dependability bottlenecks, and provide stochastic models for a holistic assessment of system dependability. [ABSTRACT FROM AUTHOR]

With the development of Internet of Things (IoT) and cloud technologies, numerous IoT devices and sensors transmit huge amounts of data to cloud data centers for further processing. While providing us considerable convenience, cloud-based computing and storage also bring us many security problems, such as the abuse of information collection and concentrated web servers in the cloud. Traditional intrusion detection systems and web application firewalls are becoming incompatible with the new network environment, and related systems with machine learning or deep learning are emerging. However, cloud-IoT systems increase attacks against web servers, since data centralization carries a more attractive reward. In this article, based on distributed deep learning, we propose a web attack detection system that takes advantage of analyzing URLs. The system is designed to detect web attacks and is deployed on edge devices. The cloud handles the above challenges in the paradigm of the Edge of Things. Multiple concurrent deep models are used to enhance the stability of the system and the convenience in updating. We implemented experiments on the system with two concurrent deep models and compared the system with existing systems by using several datasets. The experimental results with 99.410% in accuracy, 98.91% in true positive rate (TPR), and 99.55% in detection rate of normal requests (DRN) demonstrate the system is competitive in detecting web attacks. [ABSTRACT FROM AUTHOR]

With the recent success of deep learning, the amount of data and computation continues to grow daily. Hence a distributed deep learning system that shares the training workload has been researched extensively. Although a scale-out distributed environment using commodity servers is widely used, not only is there a limit due to synchronous operation and communication traffic but also combining deep neural network (DNN) training with existing clusters often demands additional hardware and migration between different cluster frameworks or libraries, which is highly inefficient. Therefore, we propose TensorLightning which integrates the widely used data pipeline of Apache Spark with powerful deep learning libraries, Caffe and TensorFlow. TensorLightning embraces a brand-new parameter aggregation algorithm and parallel asynchronous parameter managing schemes to relieve communication discrepancies and overhead. We redesign the elastic averaging stochastic gradient descent algorithm with pruned and sparse form parameters. Our approach provides the fast and flexible DNN training with high accessibility. We evaluated our proposed framework with convolutional neural network and recurrent neural network models; the framework reduces network traffic by 67% with faster convergence.

Spaceborne radar sounders are nadir-looking sensors operating in the high frequency (HF) or very high frequency (VHF) bands with subsurface sensing capabilities. Due to technological limitations, this type of sensors often deploys omnidirectional antennas. This results in undesired artifacts such as off-nadir clutter which could hinder data interpretation. Very recent technological advancements open up the possibility of synthesizing very large antenna apertures in HF/VHF band by using small satellites array deployed in suitable orbital formation flying. Accordingly, in this study, we propose a novel concept of distributed radar sounder system. The proposed concept is complemented with a mathematical model for performance prediction which takes into account the uncertainty on the position of the sensors. Moreover, we discuss possible orbital solutions for the problem of the deployment of the distributed radar sounder system. The results show that a distributed radar sounder operating in small satellites formation flying is particularly appealing as it can: 1) substantially reduce the impact of surface clutter; 2) increase the across-track resolution; 3) increase the signal-to-noise ratio (SNR) (or, alternatively, decrease the overall required transmitted power with respect to a traditional single configuration radar sounder design); and 4) provide large flexibility in the data processing of the signals acquired by the different sensors. [ABSTRACT FROM AUTHOR]

Wireless sensor networks are composed of nodes that monitor the environment and send their measurements to a base station. Due to the reduced computational and energetical resources of the sensor nodes, network organization must take into account those constraints to reduce energy consumption and, in turn, to prolong the lifetime of the network. Thus, collaborative management techniques, like clustering, are usually implemented. This paper proposes a novel unequal distributed clustering algorithm for wireless sensor networks that employs a new set of input variables. Compared with previous works, these variables help for a more accurate estimation of the convenience of a node to be a cluster head. The developed algorithm relies on local information of the sensors to diminish the data exchange, reduce the interference and the consumption of the transmission processes. Because of the underlying uncertainty of the local data, a type-2 fuzzy logic system constitutes the basis of the proposed clustering algorithm whose knowledge base is sampled to allow its feasible implementation in a node. In addition, it is used an unequal scheme for the cluster sizes because it has been demonstrated that those methods are more convenient in terms of energy consumption. The simulation results demonstrate the effectiveness of the algorithm to extend the network lifetime. [ABSTRACT FROM AUTHOR]

Large-scale systems with all-flash arrays have become increasingly common in many computing segments. To make such systems resilient, we can adopt erasure coding such as Reed-Solomon (RS) code as an alternative to replication because erasure coding incurs a significantly lower storage overhead than replication. To understand the impact of using erasure coding on the system performance and other system aspects such as CPU utilization and network traffic, we build a storage cluster that consists of approximately 100 processor cores with more than 50 high-performance solid-state drives (SSDs), and evaluate the cluster with a popular open-source distributed parallel file system, called Ceph. Specifically, we analyze the behaviors of a system adopting erasure coding from the following five viewpoints, and compare with those of another system using replication: (1) storage system I/O performance; (2) computing and software overheads; (3) I/O amplification; (4) network traffic among storage nodes, and (5) impact of physical data layout on performance of RS-coded SSD arrays. For all these analyses, we examine two representative RS configurations, used by Google file systems, and compare them with triple replication employed by a typical parallel file system as a default fault tolerance mechanism. Lastly, we collect 96 block-level traces from the cluster and release them to the public domain for the use of other researchers. [ABSTRACT FROM AUTHOR]

While the intersection of blockchain and Industrial Internet of Things (IIoT) has received considerable research interest lately, the conflict between the high resource requirements of blockchain and the generally inadequate performance of IIoT devices has not been well tackled. On one hand, due to the introductions of mathematical concepts, including Public Key Infrastructure, Merkle Hash Tree, and Proof of Work (PoW), deploying blockchain demands huge computing power. On the other hand, full nodes should synchronize massive block data and deal with numerous transactions in peer-to-peer network, whose occupation of storage capacity and bandwidth makes IIoT devices difficult to afford. In this paper, we propose a lightweight blockchain system called LightChain, which is resource-efficient and suitable for power-constrained IIoT scenarios. Specifically, we present a green consensus mechanism named Synergistic Multiple Proof for stimulating the cooperation of IIoT devices, and a lightweight data structure called LightBlock to streamline broadcast content. Furthermore, we design a novel Unrelated Block Offloading Filter to avoid the unlimited growth of ledger without affecting blockchain's traceability. The extensive experiments demonstrate that LightChain can reduce the individual computational cost to 39.32% and speed up the block generation by up to 74.06%. In terms of storage and network usage, the reductions are 43.35% and 90.55%, respectively. [ABSTRACT FROM AUTHOR]

Modern computing systems especially cloud-based and cloud-centric systems always consist of a mass of components running in large distributed environments with complicated interactions. They are vulnerable to performance problems due to the highly dynamic runtime environment changes (e.g., overload and resource contention) or software bugs (e.g., memory leak). Unfortunately, it is notoriously difficult to diagnose the root causes of these performance problems in a fine granularity due to complicated interactions and a large cardinality of potential cause set. In this paper, we build an automated, black-box and end-to-end cause inference system named CauseInfer to pinpoint the root causes or at least provide some hints. CauseInfer can automatically map a distributed system to a two-layer hierarchical causality graph and infer the root causes along the causal paths in the causality graph. CauseInfer models the fault propagation paths in an explicit way and works without instrumentation to the running production system, which makes CauseInfer more effective and practical than previous approaches. The experimental evaluations in two benchmark systems show that CauseInfer can identify the root causes in a high accuracy. Compared to several state-of-the-art approaches, CauseInfer can achieve over 10 percent improvement. Moreover, CauseInfer is lightweight and flexible enough to readily scale out in large distributed systems. With CauseInfer, the mean time to recovery (MTTR) of the cloud systems can be significantly reduced. [ABSTRACT FROM AUTHOR]

Orthogonal frequency-division multiplexing with index modulation (OFDM-IM) has become a high-profile modulation scheme for the fifth-generation (5G) wireless communications and has thus been extended to multi-hop scenarios in order to improve the network coverage and energy efficiency. However, the extension of OFDM-IM to multi-relay cooperative networks is not trivial, since it is required that a complete OFDM block should be received and decoded as an entity in one node. This requirement prevents the employment of multiple relays to forward fragmented OFDM blocks on individual subcarriers. In this regard, we propose a distributed processing scheme for multi-relay assisted OFDM-IM, by which multiple relays are selected to forward signals in a per-subcarrier manner to provide optimal error performance for two-hop decode-and-forward (DF) OFDM-IM systems. Specifically, a single selected relay only needs to decode partial information carried on certain active subcarriers and forward just as for traditional OFDM systems without IM. After receiving all signals on active subcarriers forwarded by different relays, the destination can reconstruct the complete OFDM block and retrieve the full information. We analyze the average block error rate and modulation capacity of the two-hop OFDM-IM system employing the proposed distributed DF protocol and verify the analysis by numerical simulations. [ABSTRACT FROM AUTHOR]

The ever increasing of display resolution on mobile devices raises high demand for GPU rendering details. However, the challenge of poor hardware support but fine-grained rendering details often makes user unsatisfied especially in calling for high frame rate scenarios, e.g., game. To resolve such issue, we propose ButterFly, a novel system which collaboratively utilizes mobile GPUs to process high-quality rendering details for on-the-go mobile users. In particular, ButterFly achieves three technical contributions for the collaborative design: (1) a mobile device can migrate GPU workloads in buffer queue to peers, (2) the collaborative rendering mechanism benefits user high quality details while significant power saving performance, and (3) unnecessary 3D texture rendering can be clipped for further optimization. All the techniques are compatible with the OpenGL ES standards. Furthermore, a 40-person survey perceives that ButterFly can provide excellent user experience of both rendering details and frame rate over Wi-Fi network. In addition, our comprehensive trace-driven experiments on Android prototype reveal the benefits of Butterfly have more superior performance over state-of-the-art systems, which achieves more than 28.3 percent power saving. [ABSTRACT FROM AUTHOR]

This paper presents a nondestructive method to examine the condition of objects, such as brick wall, flood embankments, landfills, air brick, and similar constructions. We used a setup made of specially built a distributed system with measuring devices to determine the moisture level of flood embankments and to test brick walls. It is an innovative solution for the evaluation buildings and tanks, both in terms of the measuring method the reconstruction algorithm. The application of modern tomographic techniques in conjunction with topological algorithms allows us to perform a noninvasive and very accurate spatial assessment of the dampness level. Electrical tomography under various names includes many techniques for tomographic imaging of the electrical parameters of an object placed in an examination area. We propose a new hybrid solution utilizing imaging techniques together with surface electrodes. Although many methods of evaluating dampness and damage exist, there is no universal solution that is optimal under a wide range of measurement conditions. Our proposed solution achieves this using the topology method for optimization. The smart tomographic measurement systems used in our solution were developed by authors belonging to the Netrix Research and Development Laboratory. These systems included measuring devices, distributed systems, algorithms, and applications for image reconstruction. Several types of reconstruction algorithms and models are explored in this paper. The solution of this optimization problem is obtained by combining the finite element method and topological algorithms. Reconstruction of 2-D examples using numerical and experimental data is shown. The proposed tomographic system consists of a central processing unit, a set of sensors (devices) and a software solution that leverages cloud computing and a big data cluster for processing, visualizing, and analyzing data (a cyber-physical system). [ABSTRACT FROM PUBLISHER]

The ITER neutral beam test facility, currently under the construction at Padua, Italy, consists of two experiments: source for production of ions of deuterium extracted (SPIDER) to test the ITER size ion source and megavolt iter injector and concept advancement (MITICA) to test the full ITER neutral beam injector. While the former represents an intermediate step, also finalized to the ITER Diagnostic Neutral Beam, the latter will represent the final design and implementation of the full prototype of the ITER heating neutral beam (HNB) injectors. Consequently, for the control and data acquisition system (CODAS) of SPIDER, there are no requirements for compliance with ITER, whereas the MITICA HNB Plant System will comply with the ITER directives for plant instrumentation and control because final system will be eventually integrated in ITER. SPIDER CODAS is currently under commissioning and the integration of the different frameworks that operate in an integrated way is proven to be successful. The experience and solutions gained in SPIDER CODAS development will be reused as far as possible in MITICA and to provide at the same time the required ITER plant system compatibility, a set of ITER-like layers (networks) will be defined, using exactly the same protocol defined for ITER interfaces. [ABSTRACT FROM PUBLISHER]