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

{"title":"GPartition-store: A multi-group collaborative parallel data storage mechanism for permissioned blockchain sharding","authors":"Lin Qiu ,&nbsp;Bo Yi ,&nbsp;Xingwei Wang ,&nbsp;Fei Gao ,&nbsp;Kaimin Zhang ,&nbsp;Yanpeng Qu ,&nbsp;Min Huang","doi":"10.1016/j.future.2025.107731","DOIUrl":"10.1016/j.future.2025.107731","url":null,"abstract":"<div><div>The problem of insufficient storage space caused by the full-replication mechanism, which is commonly employed in existing blockchains, poses an obstacle to system scalability. Moreover, existing storage sharding mechanisms are confronted with the risk of data tampering by reason of the existence of Byzantine nodes. To address the above problems, the storage partition mechanisms, integrating Erasure Coding with Byzantine Fault Tolerance consensus protocol, are proposed such as BFT-Store and PartitionChain. While promising, these solutions still encounter three significant challenges. First, the substantial computational complexity associated with encoding during data storage and decoding during data recovery will impede the efficiency (e.g., latency and throughput) of the permissioned blockchain. Second, the signature schemes employed for verifying the completeness and correctness of encoded data on each node lead to massive communication over the network, thereby further limiting the system efficiency. Third, the process of system re-initialization, which necessitates the participation of all nodes, degrades the system stability. This paper proposes a Multi-group Collaborative Parallel Data Storage Mechanism for Permissioned Blockchain Sharding called GPartition-Store to alleviate the above problems, where the nodes are divided into multiple Storage Groups (SGs). First, the original block is partitioned into <span><math><mi>g</mi></math></span> sub-blocks (assuming <span><math><mi>g</mi></math></span> is the number of SGs), with each sub-block being further partitioned and encoded into smaller encoded-blocks or recovered by decoding in parallel across all SGs. Hence, the computational complexity of coding (i.e., encoding and decoding) can be decreased by about <span><math><msup><mrow><mi>g</mi></mrow><mrow><mn>2</mn></mrow></msup></math></span> and <span><math><msup><mrow><mi>g</mi></mrow><mrow><mn>3</mn></mrow></msup></math></span> times respectively. Second, the bloom filter is utilized to generate the verification proofs of the sub-blocks and encoded-block sets, which simultaneously avoids the heavy amount of transmitted messages, while liberating the requirement for dependence on any trusted third party. Third, the re-initialization process is launched exclusively within a specific SG when a node joins/quits the system or a single crashed node needs repair, thereby enhancing the system stability. Compared with the full-replication mechanism, BFT-Store and PartitionChain, the experimental results illustrate that GPartition-Store can improve the scalability, efficiency and stability of the dynamic blockchain network while maintaining the availability of the blocks.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107731"},"PeriodicalIF":6.2,"publicationDate":"2025-01-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143077827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"FedDSHAR: A dual-strategy federated learning approach for human activity recognition amid noise label user","authors":"Ziqian Lin ,&nbsp;Xuefeng Jiang ,&nbsp;Kun Zhang ,&nbsp;Chongjun Fan ,&nbsp;Yaya Liu","doi":"10.1016/j.future.2025.107724","DOIUrl":"10.1016/j.future.2025.107724","url":null,"abstract":"<div><div>Federated learning (FL) has recently achieved successes in privacy-sensitive health-care applications like medical analysis. Most previous studies suppose that collected user data are well-annotated, however, it is a strong assumption in practice. For instance, human activity recognition (HAR) task aims to train a model which predicts a certain person’s activity based on sensor data series collected from a given period of time. Due to diverse and incomplete annotation approaches, user-side data inevitably contain significant label noise, which greatly degrade model convergence and performance. In this work, we propose a novel FL framework FedDSHAR, which partitions the user-side data into the clean data subset and noisy data subset. Two strategies are utilized on two subsets to further exploit extra effective information from data, where strategic time-series augmentation is adopted on the clean subset and the semi-supervised learning scheme is used for the noisy subset. Extensive experiments conducted on three public real-world HAR datasets demonstrate that FedDSHAR outperforms six state-of-the-art methods, particularly in addressing extreme label noise in real-world distributed noisy HAR scenarios. Our code is available at <span><span>https://github.com/coke2020ice/FedDSHAR</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107724"},"PeriodicalIF":6.2,"publicationDate":"2025-01-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143077828","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An efficient blockchain for decentralized ABAC policy decision point","authors":"Qiwei Hu ,&nbsp;Miguel Correia ,&nbsp;Tao Jiang","doi":"10.1016/j.future.2025.107732","DOIUrl":"10.1016/j.future.2025.107732","url":null,"abstract":"<div><div>Blockchain-enabled Policy Decision Point (PDP) has been a promising solution to the centralization concern in practical deployment of Attribute-Based Access Control (ABAC). However, existing blockchain systems cannot support PDP adequately since PDP functionalities introduce extra latency to blockchain’s execution process and limits system throughput. This paper proposes an efficient PDP Blockchain (PDPB) by exploiting a minimum-redundancy execution paradigm. Concretely, we design a novel Echo-Based Execution Conclude (EBEC) mechanism to enable minimum redundancy request evaluation while ensure blockchain safety and liveness. Two optimization techniques, Echo Compacting (EC) and Load Balancing (LB), are proposed to reduce the communication and computation overhead of PDPB and further enhance its performance. We implement a prototype of PDPB and evaluate it on Amazon Web Services (AWS) servers. The results show that PDPB achieves more than 35.6% performance improvement over existing methods.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107732"},"PeriodicalIF":6.2,"publicationDate":"2025-01-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143077826","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Leveraging local and global relationships for corrupted label detection","authors":"Phong Lam,&nbsp;Ha-Linh Nguyen,&nbsp;Xuan-Truc Dao Dang,&nbsp;Van-Son Tran,&nbsp;Minh-Duc Le,&nbsp;Thu-Trang Nguyen,&nbsp;Son Nguyen,&nbsp;Hieu Dinh Vo","doi":"10.1016/j.future.2025.107729","DOIUrl":"10.1016/j.future.2025.107729","url":null,"abstract":"<div><div>The performance of the Machine learning and Deep learning models heavily depends on the quality and quantity of the training data. However, real-world datasets often contain a considerable percentage of noisy labels, ranging from 8.0% to 38.5%. This could significantly reduce model accuracy. To address the problem of corrupted labels, we propose <span>Cola</span>, a novel data-centric approach that leverages both <em>local</em> neighborhood similarities and <em>global</em> relationships across the entire dataset to detect corrupted labels. The main idea of our approach is that similar instances tend to share the same label, and the relationship between clean data can be learned and utilized to distinguish the correct and corrupted labels. Our experiments on four well-established datasets of image and text demonstrate that <span>Cola</span> consistently outperforms state-of-the-art approaches, achieving improvements of 8% to 21% in F1-score for identifying corrupted labels across various noise types and rates. For visual data, <span>Cola</span> achieves improvements of up to 80% in F1-score, while for textual data, the average improvement reaches about 17% with a maximum of 91%. Furthermore, <span>Cola</span> is significantly more effective and efficient in detecting corrupted labels than advanced large language models, such as <em>Llama3</em>, with improvements of up to 112% in Precision and a 300X reduction in execution time.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107729"},"PeriodicalIF":6.2,"publicationDate":"2025-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143077829","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Smart contract-based public integrity auditing for cloud storage against malicious auditors","authors":"Hui Tian ,&nbsp;Nan Gan ,&nbsp;Fang Peng ,&nbsp;Hanyu Quan ,&nbsp;Chin-Chen Chang ,&nbsp;Athanasios V. Vasilakos","doi":"10.1016/j.future.2025.107709","DOIUrl":"10.1016/j.future.2025.107709","url":null,"abstract":"<div><div>Cloud storage, a vital component of cloud computing, faces significant challenges in ensuring data integrity, which hinders its widespread adoption. Public auditing models, which rely on third-party auditors (TPAs), have been developed to address these issues by offloading computation from users. However, maintaining the consistent trustworthiness of TPAs remains a major challenge, especially in preventing dishonest behaviors, such as collusion, procrastination, and forgery. In this paper, we propose a novel smart contract-based public integrity auditing scheme for cloud storage, introducing a transparent, non-black-box auditing process. This scheme adopts certificateless authentication, significantly reducing the overhead associated with traditional key management and certificate handling. To mitigate TPA dishonesty, we introduce a blockchain-based challenge generation algorithm and an auditing process preservation mechanism. The challenge algorithm ensures fair random sampling by leveraging blockchain’s immutability, reducing the risk of collusion between TPAs and cloud service providers (CSPs). The auditing process preservation mechanism prevents procrastination by recording task completion times and preserving metadata, ensuring full traceability and accountability. We also present a post-auditing validation mechanism that enhances the verifiability of auditing results, comprising two components: auditing computation proof, which verifies the correctness of computationally intensive steps, and auditing process replay, which replays the entire auditing using preserved metadata. Finally, we formally prove the security of our scheme and conduct a comprehensive performance comparison with existing solutions. The results demonstrate that our approach offers strong security, reduces computational overhead, and maintains comparable communication overhead to other schemes.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107709"},"PeriodicalIF":6.2,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143049873","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Generating hard Ising instances with planted solutions using post-quantum cryptographic protocols","authors":"Salvatore Mandrà ,&nbsp;Humberto Munoz-Bauza ,&nbsp;Gianni Mossi ,&nbsp;Eleanor G. Rieffel","doi":"10.1016/j.future.2025.107721","DOIUrl":"10.1016/j.future.2025.107721","url":null,"abstract":"<div><div>In this paper we present a novel method to generate hard instances with planted solutions based on the public–private McEliece post-quantum cryptographic protocol. Unlike other planting methods rooted in the infinite-size statistical analysis, our cryptographic protocol generates instances which are <em>all</em> hard (in cryptographic terms), with the hardness tuned by the size of the private key, and with a guaranteed unique ground state. More importantly, because of the private–public key protocol, planted solutions cannot be easily recovered by a direct inspection of the planted instances without the knowledge of the private key used to generate them, therefore making our protocol suitable to test and evaluate quantum devices without the risk of “backdoors” being exploited.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107721"},"PeriodicalIF":6.2,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143077831","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Using binary hash tree-based encryption to secure a deep learning model and generated images for social media applications","authors":"Soniya Rohhila,&nbsp;Amit Kumar Singh","doi":"10.1016/j.future.2025.107722","DOIUrl":"10.1016/j.future.2025.107722","url":null,"abstract":"<div><div>Deep learning (DL) plays a vital role in identifying critical features and patterns in digital images. Deep learning models and generated records, particularly digital images, are highly effective in media and other applications but pose privacy and security challenges. For example, healthcare professionals must understand how Artificial Intelligence (AI) makes decisions to trust and fully incorporate its findings into medical practice. This research addresses the security and privacy challenges associated with a DL model and generated records for social media applications. In this work, we propose a binary hash tree-based encryption that encrypts a customised model and generated images to minimise data leakage. The proposed method includes three parts. First is a customised autoencoder that minimises the size of digital images ensuring the security of generated images with a Henon chaotic map and ephemeral keys derived from a binary hash tree (BHT) for encryption in a Galois field (GF). Further, we encrypt the fewest possible weight parameters of the customised model with the same ephemeral key to preserve privacy. By doing so, our method reduces data leakage and further improves the model security at the same time. Extensive experiments reveal that the proposed method is more secure against attacks than state-of-the-art methods, which could be helpful in media and several other applications. To the best of our knowledge, we are the first to explore a secure system that protects both the model and the generated media at the same time using an encryption technique.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107722"},"PeriodicalIF":6.2,"publicationDate":"2025-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143077830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Distributed and heterogeneous tensor–vector contraction algorithms for high performance computing","authors":"Pedro J. Martinez-Ferrer ,&nbsp;Albert-Jan Yzelman ,&nbsp;Vicenç Beltran","doi":"10.1016/j.future.2024.107698","DOIUrl":"10.1016/j.future.2024.107698","url":null,"abstract":"<div><div>The tensor–vector contraction (TVC) is the most memory-bound operation of its class and a core component of the higher-order power method (HOPM). This paper brings distributed-memory parallelization to a native TVC algorithm for dense tensors that overall remains oblivious to contraction mode, tensor splitting, and tensor order. Similarly, we propose a novel distributed HOPM, namely dHOPM<span><math><msub><mrow></mrow><mrow><mn>3</mn></mrow></msub></math></span>, that can save up to one order of magnitude of streamed memory and is about twice as costly in terms of data movement as a distributed TVC operation (dTVC) when using task-based parallelization. The numerical experiments carried out in this work on three different architectures featuring multicore and accelerators confirm that the performances of dTVC and dHOPM<span><math><msub><mrow></mrow><mrow><mn>3</mn></mrow></msub></math></span> remain relatively close to the peak system memory bandwidth (50%–80%, depending on the architecture) and on par with STREAM benchmark figures. On strong scalability scenarios, our native multicore implementations of these two algorithms can achieve similar and sometimes even greater performance figures than those based upon state-of-the-art CUDA batched kernels. Finally, we demonstrate that both computation and communication can benefit from mixed precision arithmetic also in cases where the hardware does not support low precision data types natively.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107698"},"PeriodicalIF":6.2,"publicationDate":"2025-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143077832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"EC5: Edge–cloud collaborative computing framework with compressive communication","authors":"Jingwei Tan ,&nbsp;Fagui Liu ,&nbsp;Bin Wang ,&nbsp;Qingbo Wu ,&nbsp;C.L. Philip Chen","doi":"10.1016/j.future.2025.107715","DOIUrl":"10.1016/j.future.2025.107715","url":null,"abstract":"<div><div>With an increasing number of deep neural network (DNN)-based applications being deployed at the edges, edge–cloud collaborative computing has emerged as a promising solution to alleviate the burden on resource-constrained edges by collaborative inference. However, simply offloading part of DNN to the cloud introduces significant communication overhead during inference. In this paper, we propose EC5, an Edge–Cloud Collaborative Computing framework with Compressive Communication. The compression of the intermediate feature is formulated using information theory and jointly optimized with the DNN through end-to-end multi-task learning. By decomposing DNN parameters into a new space, EC5 enables efficient storage and update of models across various compression levels. An Adaptive Exit scheme is designed to retain high-confidence inputs on the edge for fast inference, reducing reliance on the cloud. Experimental comparisons with baseline methods prove that EC5 significantly conserves network bandwidth and reduces communication instances, with low latency and acceptable accuracy loss, showing flexibility across different communication scenarios.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107715"},"PeriodicalIF":6.2,"publicationDate":"2025-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143049875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ENNigma: A framework for Private Neural Networks","authors":"Pedro Barbosa ,&nbsp;Ivone Amorim ,&nbsp;Eva Maia ,&nbsp;Isabel Praça","doi":"10.1016/j.future.2025.107719","DOIUrl":"10.1016/j.future.2025.107719","url":null,"abstract":"<div><div>The widespread use of the Internet and digital services has significantly increased data collection and processing. Critical domains like healthcare rely on this data, but privacy and security concerns limit its usability, constraining the performance of intelligent systems, particularly those leveraging Neural Networks (NNs). NNs require high-quality data for optimal performance, but existing privacy-preserving methods, such as Federated Learning and Differential Privacy, often degrade model accuracy. While Homomorphic Encryption (HE) has emerged as a promising alternative, existing HE-based methods face challenges in computational efficiency and scalability, limiting their real-world application.</div><div>To address these issues, we introduce ENNigma, a novel framework employing state-of-the-art Fully Homomorphic Encryption techniques. This framework introduces optimizations that significantly improve the speed and accuracy of encrypted NN operations. Experiments conducted using the CIC-DDoS2019 dataset — a benchmark for Distributed Denial of Service attack detection — demonstrate ENNigma’s effectiveness. A classification performance with a maximum relative error of 1.01% was achieved compared to non-private models, while reducing multiplication time by up to 59% compared to existing FHE-based approaches. These results highlight ENNigma’s potential for practical, privacy-preserving neural network applications.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"166 ","pages":"Article 107719"},"PeriodicalIF":6.2,"publicationDate":"2025-01-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143077833","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}