Pub Date : 2026-06-01Epub Date: 2026-02-26DOI: 10.1016/j.sysarc.2026.103750
Qiqi Gu , Chenpeng Wu , Dong Zhang , Bingheng Yan , Junwei Zhou , Jianguo Yao
The emergence of byte-addressable Non-Volatile Main Memory (NVMM) with near-DRAM latency, high endurance, and true persistence has enabled a new generation of high-performance file systems. However, existing NVMM-based file systems exhibit poor scalability under intensive concurrent access to a shared file, primarily due to coarse-grained file-level lock mechanism. To address this bottleneck, we propose three design principles as guidelines for developing a scalable file system. Based on these principles, three design strategies, including block-level lock, log pre-allocation, and page-level garbage collection, are introduced and implemented in a scalable NVMM-based file system called ISFS, which supports intensive concurrent requests for a shared file. Compared with NVMM-based kernel-level file systems, ISFS achieves up to 22.4 and 2.2 performance improvement for intensive shared-file read and write workloads, respectively. By narrowing the performance gap between private and shared file access, ISFS demonstrates that native, scalable shared-file support is achievable on byte-addressable persistent memory.
{"title":"Scaling NVMM-based file system on intensive shared file access","authors":"Qiqi Gu , Chenpeng Wu , Dong Zhang , Bingheng Yan , Junwei Zhou , Jianguo Yao","doi":"10.1016/j.sysarc.2026.103750","DOIUrl":"10.1016/j.sysarc.2026.103750","url":null,"abstract":"<div><div>The emergence of byte-addressable Non-Volatile Main Memory (NVMM) with near-DRAM latency, high endurance, and true persistence has enabled a new generation of high-performance file systems. However, existing NVMM-based file systems exhibit poor scalability under intensive concurrent access to a shared file, primarily due to coarse-grained file-level lock mechanism. To address this bottleneck, we propose three design principles as guidelines for developing a scalable file system. Based on these principles, three design strategies, including block-level lock, log pre-allocation, and page-level garbage collection, are introduced and implemented in a scalable NVMM-based file system called ISFS, which supports intensive concurrent requests for a shared file. Compared with NVMM-based kernel-level file systems, ISFS achieves up to 22.4<span><math><mo>×</mo></math></span> and 2.2<span><math><mo>×</mo></math></span> performance improvement for intensive shared-file read and write workloads, respectively. By narrowing the performance gap between private and shared file access, ISFS demonstrates that native, scalable shared-file support is achievable on byte-addressable persistent memory.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103750"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386517","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}
Pub Date : 2026-06-01Epub Date: 2026-03-04DOI: 10.1016/j.sysarc.2026.103760
Zeming Wei, Guanzhang Yue, Yihao Zhang, Meng Sun
Recently, Large Language Models (LLMs) have achieved tremendous success in various tasks. In particular, In-context Learning (ICL) has emerged as a popular inference paradigm for eliciting the reasoning capability of LLMs. In ICL systems, LLMs can efficiently learn new tasks during inference without modifying their parameters by adding only a few input–output example pairs demonstrating the task. Such mysterious ability of LLMs has attracted great research interests in understanding, formatting, and improving the in-context demonstrations, while still suffering from drawbacks like the sensitivity to the selection and organization of examples, making reliable evaluations for ICL systems crucial. Inspired by the foundations of adopting testing techniques in machine learning (ML) systems, we study the mutation testing technique for ICL systems, aiming to characterize the quality and effectiveness of their test data. First, we propose a general mutation testing framework for ICL systems, as well as the mutation operators and scores that are specialized for ICL demonstrations. Then, we implement this testing pipeline under various types of ICL systems, including classification, regression, and generation tasks. By adapting the generalized mutation operators and scores, we demonstrate the framework’s versatility and effectiveness for different ICL applications. With comprehensive experiments across multiple models and benchmarks, we show the effectiveness of our framework in evaluating the reliability and quality of ICL test suites under various scenarios, contributing new insights and techniques for ICL system evaluation. Our code is available at https://github.com/weizeming/MILE.
{"title":"On Mutation Testing of In-Context Learning Systems","authors":"Zeming Wei, Guanzhang Yue, Yihao Zhang, Meng Sun","doi":"10.1016/j.sysarc.2026.103760","DOIUrl":"10.1016/j.sysarc.2026.103760","url":null,"abstract":"<div><div>Recently, Large Language Models (LLMs) have achieved tremendous success in various tasks. In particular, In-context Learning (ICL) has emerged as a popular inference paradigm for eliciting the reasoning capability of LLMs. In ICL systems, LLMs can efficiently learn new tasks during inference without modifying their parameters by adding only a few input–output example pairs demonstrating the task. Such mysterious ability of LLMs has attracted great research interests in understanding, formatting, and improving the in-context demonstrations, while still suffering from drawbacks like the sensitivity to the selection and organization of examples, making reliable evaluations for ICL systems crucial. Inspired by the foundations of adopting testing techniques in machine learning (ML) systems, we study the mutation testing technique for ICL systems, aiming to characterize the quality and effectiveness of their test data. First, we propose a general mutation testing framework for ICL systems, as well as the mutation operators and scores that are specialized for ICL demonstrations. Then, we implement this testing pipeline under various types of ICL systems, including classification, regression, and generation tasks. By adapting the generalized mutation operators and scores, we demonstrate the framework’s versatility and effectiveness for different ICL applications. With comprehensive experiments across multiple models and benchmarks, we show the effectiveness of our framework in evaluating the reliability and quality of ICL test suites under various scenarios, contributing new insights and techniques for ICL system evaluation. Our code is available at <span><span>https://github.com/weizeming/MILE</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103760"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386520","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}
Pub Date : 2026-06-01Epub Date: 2026-02-21DOI: 10.1016/j.sysarc.2026.103741
Antônio Augusto Fröhlich, Leonardo Passig Horstmann, Jozimar Custódio Xavier
Power management is a cornerstone for many Cyber-Physical Systems (CPSs), which relies on low-power circuits, dynamic power management algorithms and energy-aware software to match their requirements in terms of energy. As CPSs evolve towards data-centric designs to more promptly accommodate AI models and integration, traditional power management techniques must also be improved. In this paper, we build on SmartData to introduce a data-centric Power Manager (PM) framework that allows CPSs to model energy in terms of data. SmartData defines a high-level interface for sensing, actuation, and control in data-centric CPSs. It abstracts the myriad of features of modern embedded platforms related to processing, scheduling, synchronization, and communication. These Energetic SmartData encapsulate the components of a CPS, which interact in a publish–subscribe fashion, declaring interest on other SmartData and responding to other SmartData interests. We introduce an algorithm to extract a Directed Acyclic Graph (DAG) from these Interest relationships, with vertices representing the involved components and edges representing the associated cost in terms of energy. We also introduce a Power Manager that uses such DAGs to monitor the state of the system, eventually overriding low-priority Interests to reach the specified lifetime. We evaluated the proposed framework through a case study with Ocean-Bottom Nodes (OBNs) under realistic, dynamic energy conditions. Results show that without any power management, the system fails 12 days before its target operational lifetime. The proposed data-driven PM was then benchmarked against a fixed-schedule Static PM and a reactive Threshold PM. Our approach was the only strategy to guarantee a 365-day lifetime in all scenarios. With an ideal initial battery capacity of 260 Ah, it achieved a high utility of 23.1%. It also proved its adaptability in an energy-deficit scenario with an initial capacity of 257 Ah, where it reduced utility to 2.8% to survive, a condition in which the other strategies failed.
{"title":"Energetic SmartData: A data-driven power management approach for cyber–physical systems","authors":"Antônio Augusto Fröhlich, Leonardo Passig Horstmann, Jozimar Custódio Xavier","doi":"10.1016/j.sysarc.2026.103741","DOIUrl":"10.1016/j.sysarc.2026.103741","url":null,"abstract":"<div><div>Power management is a cornerstone for many Cyber-Physical Systems (CPSs), which relies on low-power circuits, dynamic power management algorithms and energy-aware software to match their requirements in terms of energy. As CPSs evolve towards data-centric designs to more promptly accommodate AI models and integration, traditional power management techniques must also be improved. In this paper, we build on SmartData to introduce a data-centric Power Manager (PM) framework that allows CPSs to model energy in terms of data. SmartData defines a high-level interface for sensing, actuation, and control in data-centric CPSs. It abstracts the myriad of features of modern embedded platforms related to processing, scheduling, synchronization, and communication. These <em>Energetic SmartData</em> encapsulate the components of a CPS, which interact in a publish–subscribe fashion, declaring interest on other SmartData and responding to other SmartData interests. We introduce an algorithm to extract a Directed Acyclic Graph (DAG) from these Interest relationships, with vertices representing the involved components and edges representing the associated cost in terms of energy. We also introduce a Power Manager that uses such DAGs to monitor the state of the system, eventually overriding low-priority Interests to reach the specified lifetime. We evaluated the proposed framework through a case study with Ocean-Bottom Nodes (OBNs) under realistic, dynamic energy conditions. Results show that without any power management, the system fails 12 days before its target operational lifetime. The proposed data-driven PM was then benchmarked against a fixed-schedule Static PM and a reactive Threshold PM. Our approach was the only strategy to guarantee a 365-day lifetime in all scenarios. With an ideal initial battery capacity of 260 Ah, it achieved a high utility of 23.1%. It also proved its adaptability in an energy-deficit scenario with an initial capacity of 257 Ah, where it reduced utility to 2.8% to survive, a condition in which the other strategies failed.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103741"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386647","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}
Pub Date : 2026-06-01Epub Date: 2026-02-24DOI: 10.1016/j.sysarc.2026.103753
Ishrak Jahan Ratul , Zhishan Guo , Kecheng Yang
Real-time object classification on edge devices with constrained computing resources involves a trade-off between computation workload and classification accuracy. Existing classifier models are typically developed either for fast inference with compromised accuracy or for high accuracy with heavy computational cost. A recently proposed concept, called IDK (“I don’t know”) classifiers, enables cascading multiple existing classifiers to achieve high accuracy while reducing average inference time. In this work, we compose IDK classifier cascades by fine-tuning existing models. We use the Tiny ImageNet and CIFAR-100 datasets with modified final layers for 200- and 100-class classification, respectively. We select a confidence threshold for each model to declare either a successful classification or an IDK decision. With a cascade of IDK classifiers, each input can be examined by more than one classifier, ordered from faster, less accurate ones to slower, more accurate ones. When an upstream classifier returns a class with high confidence, downstream classifiers are not executed, improving inference time. More accurate downstream classifiers are needed when upstream ones lack confidence. Our experimental results demonstrate that IDK classifier cascades reduce average inference time while maintaining high classification accuracy, making them suitable for real-time AI applications on edge devices.
{"title":"Accuracy-Aware IDK Cascades for Real-Time Object Classification at the Edge","authors":"Ishrak Jahan Ratul , Zhishan Guo , Kecheng Yang","doi":"10.1016/j.sysarc.2026.103753","DOIUrl":"10.1016/j.sysarc.2026.103753","url":null,"abstract":"<div><div>Real-time object classification on edge devices with constrained computing resources involves a trade-off between computation workload and classification accuracy. Existing classifier models are typically developed either for fast inference with compromised accuracy or for high accuracy with heavy computational cost. A recently proposed concept, called IDK (“I don’t know”) classifiers, enables cascading multiple existing classifiers to achieve high accuracy while reducing average inference time. In this work, we compose IDK classifier cascades by fine-tuning existing models. We use the Tiny ImageNet and CIFAR-100 datasets with modified final layers for 200- and 100-class classification, respectively. We select a confidence threshold for each model to declare either a successful classification or an IDK decision. With a cascade of IDK classifiers, each input can be examined by more than one classifier, ordered from faster, less accurate ones to slower, more accurate ones. When an upstream classifier returns a class with high confidence, downstream classifiers are not executed, improving inference time. More accurate downstream classifiers are needed when upstream ones lack confidence. Our experimental results demonstrate that IDK classifier cascades reduce average inference time while maintaining high classification accuracy, making them suitable for real-time AI applications on edge devices.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103753"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386652","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}
Pub Date : 2026-06-01Epub Date: 2026-02-28DOI: 10.1016/j.sysarc.2026.103758
Amel Faiza Tandjaoui
This paper addresses the joint optimization of Phasor Measurement Unit (PMU) and wireless communication sensor placement to minimize the installation cost of Wide Area Measurement Systems (WAMS) while ensuring both complete power grid observability and communication network connectivity. Mathematical models are first provided for the optimal PMU placement and optimal sensor placement problems independently, with the latter formulated as a Steiner tree problem. These are then integrated into a unified mixed integer linear program for joint optimization. Due to its computational complexity for large-scale networks, two efficient solution approaches are proposed: GlobalGrid, which is exhaustive, and SubSteiner, which balances optimality with computational efficiency. Numerical experiments demonstrate that networks with higher nodal degrees require fewer sensors for connectivity. Results show that incorporating Zero Injection Buses (ZIBs) reduces the required number of PMUs by up to 23%, leading to a corresponding reduction in sensor deployment costs of up to 12%. The proposed SubSteiner approach achieves near-optimal solutions with an optimality gap of less than 6% compared to GlobalGrid while significantly reducing computational time.
{"title":"PMU and wireless sensor placement in WAMS: A joint resolution","authors":"Amel Faiza Tandjaoui","doi":"10.1016/j.sysarc.2026.103758","DOIUrl":"10.1016/j.sysarc.2026.103758","url":null,"abstract":"<div><div>This paper addresses the joint optimization of Phasor Measurement Unit (PMU) and wireless communication sensor placement to minimize the installation cost of Wide Area Measurement Systems (WAMS) while ensuring both complete power grid observability and communication network connectivity. Mathematical models are first provided for the optimal PMU placement and optimal sensor placement problems independently, with the latter formulated as a Steiner tree problem. These are then integrated into a unified mixed integer linear program for joint optimization. Due to its computational complexity for large-scale networks, two efficient solution approaches are proposed: GlobalGrid, which is exhaustive, and SubSteiner, which balances optimality with computational efficiency. Numerical experiments demonstrate that networks with higher nodal degrees require fewer sensors for connectivity. Results show that incorporating Zero Injection Buses (ZIBs) reduces the required number of PMUs by up to 23%, leading to a corresponding reduction in sensor deployment costs of up to 12%. The proposed SubSteiner approach achieves near-optimal solutions with an optimality gap of less than 6% compared to GlobalGrid while significantly reducing computational time.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103758"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386516","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}
Pub Date : 2026-06-01Epub Date: 2026-02-27DOI: 10.1016/j.sysarc.2026.103727
Juhyeong Han , Jalel Ben-Othman , Hyunbum Kim
In urban and disaster environments, non-line-of-sight (NLoS) blockage and strict end-to-end latency constraints jointly degrade multi-UAV connectivity, especially when stable backhaul is unavailable. We present a RIS-augmented distributed crowdsourcing multi-agent reinforcement learning (MARL) framework in which UAV agents and RIS agents are modeled as independent learners under centralized training and distributed execution (CTDE). Each UAV learns motion and single-link selection (direct-only or direct + one selected RIS) with PPO, while each RIS learns a discrete phase/codebook policy with a categorical PPO backend.
Our learning objective explicitly internalizes (i) step-wise deadline-exceedance penalties (ReLU of delay above deadline), (ii) a priced bandwidth-sharing budget for control/neighbor messages via an online dual variable, and (iii) a non-negative diversity loss that discourages traffic collapse onto a single RIS. Tail metrics (delay ) and deadline miss rate (DMR) are used strictly as evaluation KPIs and are not backpropagated through.
In simulation under matched urban/disaster settings, we observe that RIS-augmented MARL improves service-level reliability and tail behavior under contention: success rate increases while delay decreases compared with heuristic baselines. The average SNR improvement is modest (e.g., dB vs. dB under the common logging schema), so our claims focus on tail-aware robustness and deadline feasibility rather than large mean-SNR gains. Overall, these results demonstrate that coupling learnable RIS control with distributed MARL and explicit overhead/deadline pricing yields a practical design point for edge AI-enabled crowdsourcing UAV communications under NLoS and time-critical constraints.
{"title":"RIS-augmented distributed crowdsourcing multi-agent reinforcement learning with edge AI-enabled UAV communications","authors":"Juhyeong Han , Jalel Ben-Othman , Hyunbum Kim","doi":"10.1016/j.sysarc.2026.103727","DOIUrl":"10.1016/j.sysarc.2026.103727","url":null,"abstract":"<div><div>In urban and disaster environments, non-line-of-sight (NLoS) blockage and strict end-to-end latency constraints jointly degrade multi-UAV connectivity, especially when stable backhaul is unavailable. We present a <em>RIS-augmented distributed crowdsourcing multi-agent reinforcement learning (MARL)</em> framework in which UAV agents and RIS agents are modeled as independent learners under centralized training and distributed execution (CTDE). Each UAV learns motion and <em>single-link selection</em> (direct-only or direct + one selected RIS) with PPO, while each RIS learns a discrete phase/codebook policy with a categorical PPO backend.</div><div>Our learning objective explicitly internalizes (i) step-wise <em>deadline-exceedance</em> penalties (ReLU of delay above deadline), (ii) a priced bandwidth-sharing budget for control/neighbor messages via an online dual variable, and (iii) a non-negative <em>diversity loss</em> that discourages traffic collapse onto a single RIS. Tail metrics (delay <span><math><mrow><mi>p</mi><mn>95</mn><mo>/</mo><mi>p</mi><mn>99</mn></mrow></math></span>) and deadline miss rate (DMR) are used strictly as <em>evaluation KPIs</em> and are not backpropagated through.</div><div>In simulation under matched urban/disaster settings, we observe that RIS-augmented MARL improves service-level reliability and tail behavior under contention: success rate increases while delay <span><math><mrow><mi>p</mi><mn>95</mn></mrow></math></span> decreases compared with heuristic baselines. The average SNR improvement is modest (e.g., <span><math><mrow><mo>−</mo><mn>4</mn><mo>.</mo><mn>848</mn></mrow></math></span> dB vs. <span><math><mrow><mo>−</mo><mn>4</mn><mo>.</mo><mn>74</mn></mrow></math></span> dB under the common logging schema), so our claims focus on <em>tail-aware robustness and deadline feasibility</em> rather than large mean-SNR gains. Overall, these results demonstrate that coupling learnable RIS control with distributed MARL and explicit overhead/deadline pricing yields a practical design point for edge AI-enabled crowdsourcing UAV communications under NLoS and time-critical constraints.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103727"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386515","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}
Pub Date : 2026-06-01Epub Date: 2026-03-05DOI: 10.1016/j.sysarc.2026.103773
Zejia Li , Lunzhi Deng , Xidan Xiao , Na Wang , Lanlan Liu , Siwei Li
In fog-based smart grids, smart meters collect users’ electricity data in real time and transmit it to fog node (FN) for aggregation. Then FN sends the aggregated data to control center for in-depth analysis. Throughout this process, data privacy preservation must be implemented, for which privacy-preserving data aggregation (PPDA) serves as a viable solution. However, most existing PPDA schemes do not support multidimensional data aggregation and fault-tolerance(i.e. scheme can still achieve data aggregation even when some devices malfunction), and these schemes are vulnerable to collusion attacks and have high computational and communicational overheads. To address these issues, this paper proposes a certificateless privacy-preserving multidimensional data aggregation scheme with fault-tolerance. The scheme first employs Chinese Remainder Theorem to integrate multidimensional data into a single data, followed by data obfuscation using the updated blind factors, and subsequently encrypts the data via the Paillier homomorphic encryption. Security analysis indicates that the scheme achieves data confidentiality and authenticity while resisting collusion attacks. Performance comparisons indicate its advantages in both communicational and computational overheads.
{"title":"Certificateless privacy-preserving multidimensional data aggregation scheme with fault-tolerance for fog-based smart grids","authors":"Zejia Li , Lunzhi Deng , Xidan Xiao , Na Wang , Lanlan Liu , Siwei Li","doi":"10.1016/j.sysarc.2026.103773","DOIUrl":"10.1016/j.sysarc.2026.103773","url":null,"abstract":"<div><div>In fog-based smart grids, smart meters collect users’ electricity data in real time and transmit it to fog node (FN) for aggregation. Then FN sends the aggregated data to control center for in-depth analysis. Throughout this process, data privacy preservation must be implemented, for which privacy-preserving data aggregation (PPDA) serves as a viable solution. However, most existing PPDA schemes do not support multidimensional data aggregation and fault-tolerance(i.e. scheme can still achieve data aggregation even when some devices malfunction), and these schemes are vulnerable to collusion attacks and have high computational and communicational overheads. To address these issues, this paper proposes a certificateless privacy-preserving multidimensional data aggregation scheme with fault-tolerance. The scheme first employs Chinese Remainder Theorem to integrate multidimensional data into a single data, followed by data obfuscation using the updated blind factors, and subsequently encrypts the data via the Paillier homomorphic encryption. Security analysis indicates that the scheme achieves data confidentiality and authenticity while resisting collusion attacks. Performance comparisons indicate its advantages in both communicational and computational overheads.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103773"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386523","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}
Pub Date : 2026-06-01Epub Date: 2026-02-17DOI: 10.1016/j.sysarc.2026.103738
Quentin Dariol , Sébastien Le Nours , Sébastien Pillement , Ralf Stemmer , Domenik Helms , Kim Grüttner
Deploying Artificial Neural Networks (ANNs) on embedded multi-core platforms requires precise models for estimating and optimizing timing and energy, which is crucial for enabling novel Artificial Intelligence (AI) applications. However, predicting non-functional properties (timing, power) is challenging due to degrees of parallelism in ANNs and complex effects in execution platforms (e.g. contentions at shared resources, dynamic power management). This article presents an Electronic System-Level (ESL) timing and energy modeling flow and the associated calibration methodology for optimizing ANN deployment on multi-core platforms. The proposed flow leverages SystemC simulation to offer both speed and accuracy while ensuring high scalability in many dimensions, such as platform resources modeling. Analytical models are used for ANN layer computation and communication delays as well as power consumption and energy cost. We propose a measurement-based calibration approach to these models which enables high prediction accuracy while guaranteeing high re-usability. The calibrated models can be used across different settings without the need to re-perform a calibration phase. We validate our flow against real measurements of ANN implementations on a prototype multi-core platform. Results demonstrate over 97% accuracy in timing and 93% in energy for 54 mappings of different ANNs tested with and without the use of power management on the platform, with an evaluation time under 2s per mapping. Furthermore, we illustrate that our flow is suitable for Design Space Exploration (DSE), allowing up to 24% improvement in inference time and 16% in energy compared to baseline implementation.
{"title":"A measurement-based calibration approach for highly scalable timing and energy modeling of EdgeAI multi-core systems","authors":"Quentin Dariol , Sébastien Le Nours , Sébastien Pillement , Ralf Stemmer , Domenik Helms , Kim Grüttner","doi":"10.1016/j.sysarc.2026.103738","DOIUrl":"10.1016/j.sysarc.2026.103738","url":null,"abstract":"<div><div>Deploying Artificial Neural Networks (ANNs) on embedded multi-core platforms requires precise models for estimating and optimizing timing and energy, which is crucial for enabling novel Artificial Intelligence (AI) applications. However, predicting non-functional properties (timing, power) is challenging due to degrees of parallelism in ANNs and complex effects in execution platforms (e.g. contentions at shared resources, dynamic power management). This article presents an Electronic System-Level (ESL) timing and energy modeling flow and the associated calibration methodology for optimizing ANN deployment on multi-core platforms. The proposed flow leverages SystemC simulation to offer both speed and accuracy while ensuring high scalability in many dimensions, such as platform resources modeling. Analytical models are used for ANN layer computation and communication delays as well as power consumption and energy cost. We propose a measurement-based calibration approach to these models which enables high prediction accuracy while guaranteeing high re-usability. The calibrated models can be used across different settings without the need to re-perform a calibration phase. We validate our flow against real measurements of ANN implementations on a prototype multi-core platform. Results demonstrate over 97% accuracy in timing and 93% in energy for 54 mappings of different ANNs tested with and without the use of power management on the platform, with an evaluation time under 2s per mapping. Furthermore, we illustrate that our flow is suitable for Design Space Exploration (DSE), allowing up to 24% improvement in inference time and 16% in energy compared to baseline implementation.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103738"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386646","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}
Pub Date : 2026-06-01Epub Date: 2026-02-20DOI: 10.1016/j.sysarc.2026.103733
Lide Xue , Mingzheng Wang , Xin Wang
Blockchain technology is emerging as the foundational trust infrastructure for Web 3.0, albeit faced with significant scalability challenges. While Layer-2 solutions offer high throughput, they diminish decentralization, increase security risks, or require users to continuously monitor the blockchain network—a practice unfriendly to mobile and IoT devices.
To address these issues, this paper introduces VWchain, a scalable end-to-end blockchain designed for lightweight devices. It employs a novel Value Witness (VW) data structure and a series of on-chain designs that fully leverage the storage and computational resources on the edge side (user-side), achieving a rational dispersion of global state data storage and transaction validation computation. Thanks to this approach, VWchain can be effectively deployed on consumer-grade hardware, providing the system with strong decentralization and robustness. Meanwhile, user usability is also ensured as users can freely manage their online time. Furthermore, we formally analyze and prototype VWchain, experimental results demonstrate that VWchain exhibits balanced and outstanding performance in terms of performance, device cost, and security.
{"title":"VWchain: A lightweight scalable edge-side blockchain based on Value-Witness","authors":"Lide Xue , Mingzheng Wang , Xin Wang","doi":"10.1016/j.sysarc.2026.103733","DOIUrl":"10.1016/j.sysarc.2026.103733","url":null,"abstract":"<div><div>Blockchain technology is emerging as the foundational trust infrastructure for Web 3.0, albeit faced with significant scalability challenges. While Layer-2 solutions offer high throughput, they diminish decentralization, increase security risks, or require users to continuously monitor the blockchain network—a practice unfriendly to mobile and IoT devices.</div><div>To address these issues, this paper introduces VWchain, a scalable end-to-end blockchain designed for lightweight devices. It employs a novel Value Witness (VW) data structure and a series of on-chain designs that fully leverage the storage and computational resources on the edge side (user-side), achieving a rational dispersion of global state data storage and transaction validation computation. Thanks to this approach, VWchain can be effectively deployed on consumer-grade hardware, providing the system with strong decentralization and robustness. Meanwhile, user usability is also ensured as users can freely manage their online time. Furthermore, we formally analyze and prototype VWchain, experimental results demonstrate that VWchain exhibits balanced and outstanding performance in terms of performance, device cost, and security.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103733"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386648","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}
Pub Date : 2026-06-01Epub Date: 2026-02-27DOI: 10.1016/j.sysarc.2026.103751
Abdelhamid Garah, Nader Mbarek, Sergey Kirgizov
Ensuring the integrity of Internet of Things (IoT) objects is challenging due to their limited energy and processing resources, as well as their exposure to security threats. Remote Attestation (RA) is a widely used technique that enables a trusted entity, such as a gateway, to verify the integrity of constrained IoT devices remotely. However, applying RA in constrained environments introduces challenges, including redundant attestations, high energy consumption, and vulnerabilities, such as Time-of-Check-Time-of-Use (TOCTOU) attacks. To address these limitations, this paper proposes a novel autonomic IoT framework for self-managing the integrity of IoT objects using a lightweight remote attestation mechanism and the Autonomic Computing paradigm. The proposed approach uses a DBSCAN model to determine when attestation is required, along with a fuzzy-logic system that dynamically selects an appropriate lightweight hash function based on the device state. Meanwhile, the attestation process uses a lightweight HMAC scheme to ensure device integrity. Our proposed framework reduces redundant attestations, optimizes energy consumption, and extends the lifetime of IoT systems, making it suitable for resource-constrained environments.
{"title":"Enhancing IoT object integrity and energy efficiency through DBSCAN and fuzzy Logic-based self-management framework","authors":"Abdelhamid Garah, Nader Mbarek, Sergey Kirgizov","doi":"10.1016/j.sysarc.2026.103751","DOIUrl":"10.1016/j.sysarc.2026.103751","url":null,"abstract":"<div><div>Ensuring the integrity of Internet of Things (IoT) objects is challenging due to their limited energy and processing resources, as well as their exposure to security threats. Remote Attestation (RA) is a widely used technique that enables a trusted entity, such as a gateway, to verify the integrity of constrained IoT devices remotely. However, applying RA in constrained environments introduces challenges, including redundant attestations, high energy consumption, and vulnerabilities, such as Time-of-Check-Time-of-Use (TOCTOU) attacks. To address these limitations, this paper proposes a novel autonomic IoT framework for self-managing the integrity of IoT objects using a lightweight remote attestation mechanism and the Autonomic Computing paradigm. The proposed approach uses a DBSCAN model to determine when attestation is required, along with a fuzzy-logic system that dynamically selects an appropriate lightweight hash function based on the device state. Meanwhile, the attestation process uses a lightweight HMAC scheme to ensure device integrity. Our proposed framework reduces redundant attestations, optimizes energy consumption, and extends the lifetime of IoT systems, making it suitable for resource-constrained environments.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"175 ","pages":"Article 103751"},"PeriodicalIF":4.1,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147386514","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号