Future Generation Computer Systems-The International Journal of Escience最新文献

Mobile Crowdsourcing (MCS) has emerged as a significant edge-cloud computing paradigm in which workers perceive data at the network edge and report it to cloud-based computing services for processing, enabling the construction of various applications. Consequently, it is imperative to achieve Bilateral Location Privacy-Preserving (BLPP) to protect the privacy of both Data Requester (DR) and workers, as disclosing location information entails many sensitive details that can result in losses for DR and workers alike. The Local Differential Privacy (LDP) approach is widely employed in Privacy-Preserving (PP) techniques due to its inherent advantages, wherein owners release data with added noise, allowing for proactive customization of privacy strength without relying on any third party. However, the current state of LDP methods presents a dilemma: when privacy protection is strong, introducing excessive location noise can lead to a decrease in the accuracy of task-worker matching, while a high rate of task-worker matching necessitates the compromise of privacy strength. In this paper, an intelligent Virtual Location Replacement based enhanced Bilateral Privacy-Preserving (VLR-BPP) architecture is proposed to improve privacy protection strength and matching accuracy in MCS simultaneously. Within the VLR-BPP architecture, a Bipartite-Graph-based Matrix Completion (BGMC) model is employed to establish the spatiotemporal correlations among data. Then, a Virtual Location Replacement (VLR) strategy is proposed to obfuscate the locations of tasks or workers to their highly correlated virtual location before publishing. Based on VLR, three preemptive location virtualization approaches are introduced: Only Task Location Virtual (OTLV), Only Workers Location Virtual (OWLV), and Both Task and Workers Location Virtual (BTWLV). For workers and DR, Randomized Response (RR) techniques and Random Matrix Multiplication Mechanism (RMM) are used to implement LDP independently. A greedy algorithm is adopted to recruit workers for tasks. In response to the data submitted by workers, BGMC imputation mechanism is utilized to enhance data quality. Finally, simulations based on real-world datasets demonstrate that the performance of our architecture surpasses existing state-of-the-art methods in privacy protection and data collection quality by 18.92∼38.17% and 15.49∼50.77%, respectively.

As the containers are lightweight in resource usage, they are preferred for cloud and edge computing service deployment. Containers serve the requests whenever a user sends a query; however, they remain idle when no user request comes. Again, improving the consolidation ratio of container deployments is essential to ensure fewer servers are used in a cloud data center with an optimal resource balance. To increase the consolidation ratio of a cloud data center, in this paper, we propose a system called Container State Management for Deployment (CSMD) to manage the container states. CSMD uses an algorithm to checkpoint the idle containers so that their resources can be released. The new containers are deployed using the released resources in a server. In addition, CSMD uses an algorithm to check the container status periodically, and the containers are resumed from the checkpoint state when the user requests them. We evaluate CSMD in Amazon Elastic Compute Cloud (Amazon EC2) by performing efficient state management of the containers. The experiments in the Amazon cloud show that the proposed CSMD system is superior to the existing algorithms as the proposed system increases the consolidation ratio of data centers.

Large-scale graph processing poses challenges due to its size and irregular memory access patterns, causing performance degradation in common architectures, such as CPUs and GPUs. Recent research includes accelerating graph processing using Field Programmable Gate Arrays (FPGAs). FPGAs can provide very efficient acceleration thanks to reconfigurable on-chip resources. Although limited, these resources offer a larger design space than CPUs and GPUs.

We propose an approach in which data are preprocessed in small chunks with an optimized graph partitioning technique for execution on FPGA accelerators. The chunks, located on the host, are streamed directly into a customized memory layer implemented in the FPGA, which is tightly coupled with the processing elements responsible for the graph algorithm execution. This improves application memory access latency, which is crucial in large-sale graph computing performance.

This work presents a hardware design that, combined with graph partitioning, enables us to achieve high-performance and potentially scalable handling of large graphs (i.e., graphs with millions of vertices and billions of edges in current scenarios) while using popular graph algorithms. The proposed framework accelerates performance 56 times compared with CPU (multicore with 16 logical cores in our reference experiments), 2.5 times and 4 times faster compared to state-of-the-art FPGA and GPU solutions (FPGA has 15 compute units, and GPU reference has 128 streaming-multiprocessors in our experiments), respectively, when using the PageRank algorithm. For the Single-Source-Shortest-Past (SSSP) algorithm, we achieve speedups of up to 65x, 26x, and 18x compared to CPU, GPU, and FPGA works, respectively. Lastly, in the context of the Weakly Connected Component (WCC) algorithm, our framework achieves a speedup of up to 403 times compared to the CPU, 7.4x against the GPU, and it is faster than the FPGA alternatives up to 10.3x.

Developing an annual and global high-resolution land cover map is one of the most ambitious tasks in remote sensing, with increasing importance due to the continual rise in validated data and satellite imagery. The success of land cover classification models largely hinges on the data quality, coupled with the application of Big Data techniques and distributed computing. This is essential for efficiently processing the extensive volume of available satellite data. However, maintaining the lifecycle of several annual Machine Learning models presents a complex challenge. The rise of Machine Learning Operations offers an opportunity to automate the maintenance of these models, a feature particularly beneficial in systems that require generating new models each year alongside the continuous integration of validated data. This article details the development of an end-to-end MLOps workflow, meticulously integrating land cover classification models that employ Big Data strategies for processing large-scale, high-resolution spatial data. The workflow is designed within a Kubernetes environment, achieving on-demand auto-scaling, distributed computing, and load balancing. This integration demonstrates the practicality and efficiency of managing and deploying models that treat satellite imagery in an automated, scalable framework, thus marking a significant advancement in remote sensing and MLOps.

This paper presents an innovative IIoT multi-level platform tailored to address the specific needs of the mining domain. The platform has been conceptualized and built in the context of the illuMINEation European project. For this purpose, mining specific use cases have been designed such as promoting underground safe areas, performing efficient mining operations or boosting predictive maintenance approaches. Then, specific requirements have been identified and, as a result, the platform has been developed. It consists of four-level layered platform: (1) edge devices layer to manage several sensors deployed in the mines; (2) edge box layer to provide in-mine operations such as filtering, streaming and processing; (3) fog layer which offers an overall perspective of each mine; and (4) cloud layer to centralize the data of all the mines and to provide powerful processing capabilities. In addition, the platform is robustly secured in terms of protecting communications confidentiality and access control and also provides a toolbox aimed at manipulating 3D complex images to obtain operable mine-domain novel user interfaces. Finally, a platform validation is proposed where three different use cases are explained to better show and demonstrate the capabilities of the platform.

Geo-distributed Bigdata processing is increasing day by day, resulting in the origins of data that are geographically distributed in different countries and hold datacenters (DCs) across the globe, and also the applications that use different sites to increase reliability, security, and processing performances. Most popular frameworks like Hadoop and Spark are re-designed to process geographically distributed data at their locations. However, these methods still suffer from a large amount of data transfer over the Internet, which prohibits a high processing time and cost for many applications, and in several cases, the output results of the computation are smaller than its inputs. In this paper, we keep the data locality principle for processing data at different locations but ignore the principle of transferring the entire intermediate results to a single global reducer. We propose Geo-MR, an intelligent geo-distributed MapReduce-based framework across federated cloud based on two heuristic algorithms: (i) chosen the best clusters as global reducers to reduce the communication and optimize the transfer on the bandwidth, GResearch. (ii) The second, Geo-MR, ensures the scheduling of only the relevant data to selected global reducers that process the final results. As a baseline, we propose an exact MapReduce scheduling model for benchmarking and to compare and discuss the Geo-MR heuristic algorithm results. The experimental results show that the proposed algorithm Geo-MR can improve resource (bandwidth and VMs of clusters) utilization of the cloud federation and consequently reduce cost and job response time.

Edge intelligence, network autonomy, broadband satellite connectivity, and other concepts for private 6G networks are enabling new applications for public safety authorities, e.g., for police and rescue personnel. Enriched situational awareness, group communications with high-quality video, large scale IoT, and remote control of vehicles and robots will become available in any location and situation. We analyze cybersecurity in intelligent tactical bubbles, i.e., in autonomous rapidly deployable mobile networks for public safety operations. Machine learning plays major roles in enabling these networks to be rapidly orchestrated for different operations and in securing these networks from emerging threats, but also in enlarging the threat landscape. We explore applicability of different threat and risk analysis methods for mission-critical networked applications. We present the results of a joint risk prioritization study. We survey security solutions and propose a security architecture, which is founded on the current standardization activities for terrestrial and non-terrestrial 6G and leverages the concepts of machine learning-based security to protect mission-critical assets at the edge of the network.

Linear wireless sensor networks (LWSNs) are a specialized topology of wireless sensor networks (WSNs) widely used for environmental monitoring. Traditional WSNs rely on batteries for energy supply, limiting their performance due to battery capacity constraints, while renewable energy harvesting technology is an effective approach to alleviating the battery capacity bottleneck. However, the stochastic nature of renewable energy makes designing an efficient energy management scheme for network performance improvement a compelling research problem. In this paper, we investigate the problem of maximizing throughput over a finite-horizon time period for an energy harvesting-based linear wireless sensor network (EH-LWSN). The solution to the original problem is very complex, and this complexity mainly arises from two factors. First, the optimal energy allocation scheme has temporal coupling, i.e., the current optimal strategy relies on the energy harvested in the future. Second, the optimal energy allocation scheme has spatial coupling, i.e., the current optimal strategy of any node relies on the available energy of other nodes in the network. To address these challenges, we propose an iterative energy allocation algorithm for EH-LWSN. Firstly, we theoretically prove the optimality of the algorithm and analyze the time complexity of the algorithm. Next, we design the corresponding distributed version and consider the case of estimating the energy harvest. Finally, through experiments using a real-world renewable energy dataset, the results show that the proposed algorithm outperforms the other two heuristics energy allocation schemes in terms of network throughput.

In recent vigorous developments, digital twin edge networks (DITEN) have emerged as a network paradigm to improve network communication efficiency. Given that Web 3.0 technologies promise secure decentralized data storage and effective information exchange, it is feasible to construct a wireless edge intelligence-enabled Web 3.0 physical infrastructure through DITEN. However, DITEN encounters various security threats related to communication and authentication, and establishing a secure and cost-effective authentication scheme for confidential access to physical entities poses a significant challenge. To tackle this issue, in this article, we introduce READ, a provably secure multi-factor user authentication scheme tailored for DITEN in industrial applications. Using designed ASCON cryptography primitive cipher suite, physical unclonable functions, extended Chebyshev chaotic maps, one-way secure collision-resistant hash functions, and lightweight bitwise exclusive-or operations, READ enables mutual authentication and session key negotiation among mobile users, smart gateways, and smart industrial devices. Rigorous security assessments, conducted through the real-or-random (ROR) model, the automated validation of internet security-sensitive protocols and applications (AVISPA) simulation tool, and heuristic informal security analysis, confirm that READ meets all 13 security evaluation criteria. Furthermore, compared to other seven advanced multi-factor user authentication schemes, READ excels in security and efficiency, making it ideal for practical multi-factor user authentication scenarios.

Federated Learning (FL) enhances Artificial Intelligence (AI) applications by enabling individual devices to collaboratively learn shared models without uploading local data with third parties, thereby preserving privacy. However, implementing FL in real-world scenarios presents numerous challenges, especially with IoT devices with limited memory, diverse communication conditions, and varying computational capabilities. The research community is turning to lightweight FL, the new solutions that optimize FL training, inference, and deployment to work efficiently on IoT devices. This paper reviews lightweight FL, systematically organizing and summarizing the related techniques based on its workflow. Finally, we indicate potential problems in this area and suggest future directions to provide valuable insights into the field.