Journal of Systems Architecture最新文献_第9页

DynaNet: A dynamic BFT consensus framework 动态网络：一个动态BFT共识框架

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-03-04 DOI: 10.1016/j.sysarc.2025.103386

Jiarui Li , Guoyan Zhang , Duan Liu , Fei Zhou , Leibo Li , Weizhi Meng

引用次数: 0

Beyond VABlock: Improving Transformer workloads through aggressive prefetching 超越VABlock：通过主动预取改善Transformer工作负载

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-03-04 DOI: 10.1016/j.sysarc.2025.103389

Jane Rhee , Ikyoung Choi , Gunjae Koo , Yunho Oh , Myung Kuk Yoon

{"title":"Beyond VABlock: Improving Transformer workloads through aggressive prefetching","authors":"Jane Rhee , Ikyoung Choi , Gunjae Koo , Yunho Oh , Myung Kuk Yoon","doi":"10.1016/j.sysarc.2025.103389","DOIUrl":"10.1016/j.sysarc.2025.103389","url":null,"abstract":"<div><div>The memory capacity constraint of GPUs is a major challenge in running large deep learning workloads with their ever increasing memory requirements. To run a large Transformer model with limited GPU memory, programmers need to manually allocate and copy data between CPU and GPUs. This programming burden is eased by Unified Virtual Memory (UVM), which automatically manages data transfer through its demand paging scheme. However, using UVM can cause performance degradation, especially under memory oversubscription. In this paper, we analyze the memory behavior of inference in large Transformer models using real hardware and the open-source NVIDIA UVM driver. The default Tree-Based Neighborhood (TBN) prefetcher in the UVM driver supports page prefetching within a 2MB virtual address block (VABlock), but it only detects locality within a VABlock, limiting its effectiveness for large models. Our analysis reveals that this locality extends beyond the VABlock, which the default prefetcher cannot exploit. To address this, we propose a block-aware prefetcher that prefetches multiple contiguous VABlocks with greater aggressiveness. Our evaluation shows that this approach delivers an average 2.7x performance improvement over the default TBN prefetcher when GPU memory is oversubscribed.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"162 ","pages":"Article 103389"},"PeriodicalIF":3.7,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143610428","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}

引用次数: 0

A server-aided secure interval test and comparison protocol with OPRF 一个带有OPRF的服务器辅助安全间隔测试和比较协议

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-03-03 DOI: 10.1016/j.sysarc.2025.103383

Qiming Hu , Chenyu Hu , Ziqing Wang , Fagen Li

{"title":"A server-aided secure interval test and comparison protocol with OPRF","authors":"Qiming Hu , Chenyu Hu , Ziqing Wang , Fagen Li","doi":"10.1016/j.sysarc.2025.103383","DOIUrl":"10.1016/j.sysarc.2025.103383","url":null,"abstract":"<div><div>As mobile and edge devices become more prevalent, applications handling sensitive data, such as health monitoring, age verification for online services, and vulnerability scanning tools, are becoming increasingly common. The comparison of critical data (e.g., health metrics, ages, or software versions) is essential for informed decision-making. However, securely processing this sensitive information in resource-limited environments presents significant challenges. To address these challenges, we propose a server-aided secure interval test and comparison protocol based on point and range encoding. The proposed protocol allows two parties to verify if a value belongs to an interval or is less than another value without revealing their private inputs. To handle the set membership problem in the encoding, we introduce an innovative oblivious updatable key pseudorandom function. Leveraging this function, our protocol achieves efficient delegation to a cloud server, ensures that online costs grow linearly with input length, and maintains exceptionally low amortized preprocessing costs. We formally prove the semi-honest security of our protocol and compare it to the other semi-honest protocols. On the client side, our protocol requires only tens of microseconds for computational time and less than 1 KB for communication cost for inputs up to 40 bits. Compared to the other secure interval test protocols, it significantly reduces both computational and communication costs, especially for larger inputs. Moreover, it achieves the lowest computation costs while maintaining similar communication costs to the other secure comparison protocols.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"162 ","pages":"Article 103383"},"PeriodicalIF":3.7,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143580761","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}

引用次数: 0

MIFS: A low overhead and efficient mixture file index management method in flash file system MIFS: flash文件系统中一种低开销、高效的混合文件索引管理方法

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-03-03 DOI: 10.1016/j.sysarc.2025.103387

Jingjing Jiang , Mengfei Yang , Lei Qiao , Tingyu Wang , Xi Chen

{"title":"MIFS: A low overhead and efficient mixture file index management method in flash file system","authors":"Jingjing Jiang , Mengfei Yang , Lei Qiao , Tingyu Wang , Xi Chen","doi":"10.1016/j.sysarc.2025.103387","DOIUrl":"10.1016/j.sysarc.2025.103387","url":null,"abstract":"<div><div>Emerging resource-scarce embedded systems, such as Leo satellites, face significant challenges in file system design due to the highly asymmetric sizes of main memory and storage devices. Adopting an in-storage file system to manage raw flash memory can effectively minimize data transfer latency. Existing works focus on reducing the memory footprints of file management but ignore the performance of file accesses. In this paper, we propose a new design of file index taking advantage of the continuity of data pages in the flash memory. The proposed file index manages a series of contiguous physical spaces with a <span><math><msup><mrow><mi>B</mi></mrow><mrow><mo>+</mo></mrow></msup></math></span> tree, which ensures a deterministic time to locate the file data with small memory footprints. Furthermore, we propose a two-mode page allocation mechanism to improve the performance of write operations. Based on these designs, we implement a file system called MIFS in a real embedded system. Experimental results show that MIFS can reduce 36.1% of memory footprint than YAFFS2, the typical embedded flash file system with a similar file access performance. Furthermore, MIFS can achieve 2.23<span><math><mo>×</mo></math></span>, 1.35<span><math><mo>×</mo></math></span>, 214<span><math><mo>×</mo></math></span>, and 84% performance improvement over ELOFS, the state-of-the-art file system for resource-scarce embedded systems, for write, read, update, and file system unmount operations, respectively.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"162 ","pages":"Article 103387"},"PeriodicalIF":3.7,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621502","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}

引用次数: 0

Mobile Crowdsensing Model: A survey 移动众感模型：一项调查

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-03-03 DOI: 10.1016/j.sysarc.2025.103384

Abderrafi Abdeddine, Youssef Iraqi, Loubna Mekouar

{"title":"Mobile Crowdsensing Model: A survey","authors":"Abderrafi Abdeddine, Youssef Iraqi, Loubna Mekouar","doi":"10.1016/j.sysarc.2025.103384","DOIUrl":"10.1016/j.sysarc.2025.103384","url":null,"abstract":"<div><div>Mobile Crowdsensing (MCS) is a community detection method in which a person selects a large group of individuals with mobile devices capable of detecting the physical environment and performing various sensing tasks. Thanks to the growth of the Internet of Things, it has recently become the most used paradigm to retrieve sensing data from a dynamic environment due to the users’ mobility and involvement. Indeed, compared to other sensing methods, MCS offers extensive coverage and more precise sensing performance. Optimized with specific models and parameters, it can effectively address challenges and limitations often encountered in traditional methods. To fully leverage the benefits of MCS, an in-depth understanding of its components is essential. This ensures the development of efficient strategies that aptly address the inherent challenges of MCS. Much research has converged on topics such as task allocation, incentivization, and privacy concerns. However, this has inadvertently led to confusion due to varied interpretations of models and overlapping terminology, leaving gaps in knowledge and understanding for newcomers. Our work addresses these gaps by providing a comprehensive representation of the MCS model, seeking to unify the prevailing terminologies.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"162 ","pages":"Article 103384"},"PeriodicalIF":3.7,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143571483","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}

引用次数: 0

Detection and defence against thermal and timing covert channel attacks in multicore systems 多核系统中热定时隐蔽信道攻击的检测与防御

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-02-27 DOI: 10.1016/j.sysarc.2025.103380

Parisa Rahimi , Amit Kumar Singh , Xiaohang Wang , Seyedali Pourmoafi

引用次数: 0

A cache-aware DAG scheduling method on multicores: Exploiting node affinity and deferred executions 多核上的缓存感知DAG调度方法：利用节点亲和性和延迟执行

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-02-25 DOI: 10.1016/j.sysarc.2025.103372

Huixuan Yi , Yuanhai Zhang , Zhiyang Lin , Haoran Chen , Yiyang Gao , Xiaotian Dai , Shuai Zhao

{"title":"A cache-aware DAG scheduling method on multicores: Exploiting node affinity and deferred executions","authors":"Huixuan Yi , Yuanhai Zhang , Zhiyang Lin , Haoran Chen , Yiyang Gao , Xiaotian Dai , Shuai Zhao","doi":"10.1016/j.sysarc.2025.103372","DOIUrl":"10.1016/j.sysarc.2025.103372","url":null,"abstract":"<div><div>With increasingly complex functionalities being implemented in emerging applications, multicores are widely adopted with a layered cache hierarchy, and Directed Acyclic Graphs (DAGs) are commonly employed to model the execution dependencies between tasks. For such systems, scheduling methods can be designed to effectively leverage the cache to accelerate the system execution. However, the traditional methods either do not consider DAGs, or rely on sophisticated static analysis to produce fixed scheduling solutions that require additional hardware support (<strong>e.g.</strong>, cache partitioning and colouring), which undermines both the applicability and flexibility of these methods. Recently, an online cache-aware DAG scheduling method has been presented that schedules DAGs using an execution time model with caching effects considered, eliminating the need for static analysis and additional hardware support. However, this method relies on simple heuristics with limited considerations on both the allocatable cores and the competition between nodes, resulting in intensive inter-node contention that undermines cache performance. This paper proposes CADE, a cache-aware scheduling method for DAG tasks that leverages the cache to reduce DAG makespan. To achieve this, an affinity-aware priority assignment is first constructed that mitigates the competition among nodes for their preferred cores to hit the cache. Then, a contention-aware allocation mechanism is constructed, which (i) accounts for the impact of an allocation decision on the speed-up of other nodes; and (ii) includes the busy cores for allocation by enabling the deferred execution, effectively enhancing the cache performance to accelerate the DAG execution. Experiments show that compared to the state-of-the-art, the CADE significantly reduces the DAG makespan by 24.02% on average (up to 33%) with the cache miss rate reduced by 22.06% on average.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"161 ","pages":"Article 103372"},"PeriodicalIF":3.7,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143521220","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}

引用次数: 0

Subversion-resistant public-key searchable encryption for data sharing in IIoT 用于工业物联网数据共享的抗颠覆公钥可搜索加密

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-02-22 DOI: 10.1016/j.sysarc.2025.103370

Rang Zhou , Yongkang He , Wanpeng Li , Yanping Wang , Xiaojun Zhang

{"title":"Subversion-resistant public-key searchable encryption for data sharing in IIoT","authors":"Rang Zhou , Yongkang He , Wanpeng Li , Yanping Wang , Xiaojun Zhang","doi":"10.1016/j.sysarc.2025.103370","DOIUrl":"10.1016/j.sysarc.2025.103370","url":null,"abstract":"<div><div>The rapid growth of the Industrial Internet of Things (IIoT) has driven companies to leverage its potential for enhancing manufacturing efficiency, relying heavily on data analysis. This has led to the collection of vast IIoT data volumes, often stored securely on encrypted cloud servers. To facilitate streamlined data retrieval, keyword search technology is widely employed. Traditional keyword searchable schemes focus on ensuring keyword privacy within the trapdoor and index structures. However, advanced backdoor attacks have emerged as a significant threat. These attacks exploit hidden backdoors in software or hardware, allowing random number manipulation that undermines the security of existing keyword-searchable encryption scheme. To mitigate these attacks, we propose a novel subversion-resistant public-key searchable encryption scheme that incorporates cryptographic reverse firewalls (CRFs). In our scheme, we designed a trust zone to deploy these CRFs. Moreover, the CRFs in the trust zone is randomly chosen to re-randomize the trapdoor and index values, thereby enhancing security against backdoor attacks. Additionally, our scheme facilitates data sharing in the model of one-to-one sending and receiving between different users. Performance testing demonstrates that our scheme requires only lightweight operations for the ReTrapdoor and RePEKS algorithms, making it well-suited for deployment on resource-constrained IIoT devices.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"161 ","pages":"Article 103370"},"PeriodicalIF":3.7,"publicationDate":"2025-02-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143474754","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}

引用次数: 0

Checkpointing for single core energy-neutral real-time systems 单核能量中性实时系统的检查点

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-02-20 DOI: 10.1016/j.sysarc.2025.103371

Houssam-Eddine Zahaf, Pierre-Emmanuel Hladik, Sébastien Faucou, Audrey Queudet

引用次数: 0

Optimizing encoding and repair for wide-stripe minimum bandwidth regenerating codes in in-memory key-value stores 在内存键值存储中优化宽条最小带宽再生码的编码和修复

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-02-19 DOI: 10.1016/j.sysarc.2025.103369

Xuzhe Liu , Yuchong Hu , Weichun Wang , Dan Feng , Hai Zhou

{"title":"Optimizing encoding and repair for wide-stripe minimum bandwidth regenerating codes in in-memory key-value stores","authors":"Xuzhe Liu , Yuchong Hu , Weichun Wang , Dan Feng , Hai Zhou","doi":"10.1016/j.sysarc.2025.103369","DOIUrl":"10.1016/j.sysarc.2025.103369","url":null,"abstract":"<div><div>In-memory key–value (KV) stores are essential for databases and large-scale websites. While recent studies deploy wide-stripe erasure coding in such systems to ensure data reliability and achieve extreme storage savings, they also introduce a repair penalty. A class of erasure codes, Minimum Bandwidth Regenerating (MBR) codes, offers optimal single-chunk repair bandwidth. However, deploying wide-stripe MBR codes in this context results in two types of additional traffic: (i) <em>encoding traffic</em> incurred by transmitting large amounts of raw data between nodes; (ii) <em>repair traffic</em> from retrieving unnecessary data to repair failed data.</div><div>This paper proposes MBRWide to optimize encoding and repair performance for wide-stripe MBR codes in in-memory KV stores. MBRWide includes an <em>all-node cooperative encoding scheme</em> (ACES) and a <em>fragmented repair scheme</em> (FRS). ACES selectively encodes raw chunks to reduce encoding traffic. FRS aims to enhance repair efficiency by dynamically fragmenting parity chunks during encoding. This study implements MBRWide in Memcached, a foundational component in real-world in-memory KV services. Experiments show that ACES improves encoding throughput by 16.02% to 72.92% compared to traditional encoding methods. FRS reduces degraded read latency to failed data and multiple failures repair latency by up to 34.19% and 44.89%, respectively.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"161 ","pages":"Article 103369"},"PeriodicalIF":3.7,"publicationDate":"2025-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143487600","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}

引用次数: 0