Special Issue on selected papers from the 15th International Symposium on Parallel and Distributed Computing

This special issue consists of seven representative research articles presented at the 15th International Symposium on Parallel and Distributed Computing, in Fuzhou, Fujian, China. This annual symposium brings together practitioners, researchers, and scholars from the field of parallel and distributed computing to facilitate the exchange of ideas, enable collaborations, and promote the development of new research directions. The symposium had a highly selective program composed of papers describing original and unpublished research advancing the state of the art in the field of parallel and distributed computing. For this special issue, we selected the seven best papers presented at the symposium. These papers reflect the broad nature of the field, addressing both theoretical and practical issues in cloud computing, services computing, high performance I/O, parallel algorithms, GPU computing, and multi-agent systems. Agrawal et al,1 propose a scalable model and a mechanism for automatic anomaly detection in large-scale cloud computing systems. The proposed mechanism relies on principal component analysis of system's performance metrics and achieves an accuracy of more than 80%. Alshareef and Grigoras2 present the design of a mobile cloud service that uses social media applications, such as Twitter, in emergency and risk management. The proposed service allows users to provide on-the-ground information regarding emergency events, and emergency teams to access the information in a matter of seconds. Ertl et al,3 describe the design and optimization of an efficient HDF5 I/O Kernel for massive parallel fluid flow simulations supporting fast checkpointing, restarting, and selective visualization using a single shared output file. The performance analysis shows that the proposed design achieves bandwidths close to the theoretical peak. Lin et al,4 propose a parallel scheme for computing a planar Delaunay triangulation based on the divide-and-conquer strategy. The proposed scheme is suitable for deployment on cloud computing systems achieving good speedup and efficiency. Lv et al,5 present a task-based approach for finding strongly connected components in large-scale graphs. The approach allows different algorithms to be assigned to different datasets according to the stage of processing or the size of the dataset. Performance evaluation on large real-world graphs shows that the proposed approach outperforms existing external memory solutions. Wang et al,6 propose a selective victim cache design that enables better data locality and achieves high performance. The design consists of a set-associative structure that is equivalent to the original L1D cache and a prediction scheme to avoid costly block interchanges and evictions. Performance analysis shows that the design achieves a high on-chip data cache hit rate. Zhou et al,7 propose a distributed cooperative control algorithm for networked multi-agent systems. The design of the algorithm consists of a pre-filter and a set of distributed adaptive neural network controllers. The proposed algorithm alleviates the negative effects of model uncertainty, unknown external disturbances, and actuator faults.