Logs record valuable data from different software and systems. Execution logs are widely available and are helpful in monitoring, examination, and system understanding of complex applications. However, log files usually contain too many lines of data for a human to deal with, therefore it is important to develop methods to process logs by computers. Logs are usually unstructured, which is not conducive to automatic analysis. How to categorize logs and turn into structured data automatically is of great practical significance. In this paper, LTmatch algorithm is proposed, which implements a log pattern extracting algorithm based on a weighted word matching rate. Compared with our preview work, this algorithm not only classifies the logs according to the longest common subsequence(LCS) but also gets and updates the log template in real-time. Besides, the pattern warehouse of the algorithm uses a fixed deep tree to store the log patterns, which optimizes the matching efficiency of log pattern extraction. To verify the advantages of the algorithm, we applied the proposed algorithm to the open-source data set with different kinds of labeled log data. A variety of state-of-the-art log pattern extraction algorithms are used for comparison. The result shows our method is improved by 2.67% in average accuracy when compared with the best result in all the other methods.

In recent years, High Performance Computing (HPC) has developed rapidly in China. From Chinese Academy of Sciences (CAS) level, Scientific Computing Grid (ScGrid), is a general-purpose computing platform started from 2006 in CAS, which provided a problem solving environment for computing users through grid computing and cloud computing technologies. Then ScGrid becomes Supercomputing Cloud, an important port of Chinese Science Cloud from 2011. From national level, China National Grid (CNGrid) has integrated massive HPC resources from several national supercomputing centers and other large centers distributed geographically, and been providing efficient computing services for users in diverse disciplines and research areas. During more than 10 years, CNGrid and ScGrid has integrated tens of HPC resources distributed geographically across China, comprising 6 National Supercomputer Centers of Tianjin, Jinan, Changsha, and Shenzhen, Guangzhou, Wuxi, and also dozens of teraflops-scale HPC resources belong to universities and institutes. In total, the computing capability is more than 200PF and the storage capacity is more than 160PB in CNGrid. As worked in the operation and management center of CNGrid and ScGrid for many years, we notice that users prefer to manage their jobs at different supercomputers and clusters via a global account on different remote clients such as science gateways, desktop applications and even scripts. And they don’ t like to apply for an account to each supercomputer and login into the supercomputer in specific way. Therefore, we described Unified Account Management as a Service (UAMS) to access and use all HPC resources via a global account for each user in this paper. We addressed and solved challenges for mapping a global account to many local accounts, and provided unified account registration, management and authentication for different collaborative web gateways, command toolkits and other desktop applications. UAMS was designed in accordance with the core rules of simplicity, compatibility and reusability. In architecture design, we focused on loosely-coupled style to acquire good scalability and update internal modules transparently. In implementation, we applied widely accepted knowledge for the definitions of the RESTful API and divided them into several isolated microservices according to their usages and scenarios. For security, all sensitive data transferred in wide-network is protected by HTTPS with transport layer security outside of CNGrid and secure communication channels provided by OpenSSH inside of CNGrid. In addition, all parameters submitted to RESTful web services are strictly checked in format and variable type. By providing these frequently important but always challenging capabilities as a service, UAMS allows users to use tens of HPC resources and clients via only an account, and makes it easy for developers to implement clients and services related HPC with advantages of numerous users and single sign-on capability. Based on UAMS, representative clients are introduced and reviewed combined with different authentication schemes. Finally, analysis and test of UAMS shows that it can support authentication in milliseconds level and has good scalability. In future, we plan to implement federated account service that enable a local HPC account similar to a global account to login the national HPC environment, access and use all HPC resources in CNGrid.

As scientific workflows grow more complex and green computing becomes a priority, efficient multi-objective scheduling is essential to optimize makespan, cost, and energy consumption for large task graphs. However, existing methods often suffer from scalability bottlenecks and insufficient modeling of task dependencies, leading to degraded performance on large-scale workflows. This paper proposes HiGAT-AC, a framework that combines a hierarchical graph attention network with actor-critic reinforcement learning for scalable workflow scheduling in heterogeneous systems. HiGAT-AC splits large workflows into subgraphs via spectral clustering and uses a three-level hierarchy to capture local task dependencies, coordinate inter-subgraph information, and conduct global resource allocation. The actor-critic model employs Chebyshev scalarization to balance the three conflicting objectives. Experimental results show that HiGAT-AC achieves competitive composite scores across workflow scales from 500 to 1000 tasks, with scores reaching 0.954 on 500-task workflows and 1.000 on 1000-task workflows, while remaining stable above 0.70 across all scales. Compared with traditional and representative learning-based methods, HiGAT-AC exhibits favorable overall performance and relatively stable scalability on large task graphs, providing a promising solution for scientific workflow scheduling that balances performance and sustainability.