For the high-performance computing in a WAN environment, the geographical locations of national supercomputing centers are scattered and the network topology is complex, so it is difficult to form a unified view of resources. To aggregate the widely dispersed storage resources of national supercomputing centers in China, the team led by Zhisheng Huo have previously proposed a global virtual data space named GVDS. However, the GVDS suffers from performance bottlenecks in data migration and access across WANs.
To solve the problems, the team led by Zhisheng Huo published their new research on 15 December 2024 in Frontiers of Computer Science co-published by Higher Education Press and Springer Nature.
The team proposes a performance optimization framework of GVDS including the multitask-oriented data migration method and the request access-aware IO proxy resource allocation strategy. The experimental results show that the framework can effectively reduce the average data access delay of GVDS while improving the performance of the application greatly.
In the research, they analyze performance bottlenecks in data migration and access across WANs, they propose a performance optimization framework of the GVDS, which includes the multitask-oriented data migration method and the request access-aware IO proxy resource allocation strategy.
Firstly, the multitask-oriented data migration method is named MODM, which has taken consideration of the current information of the WAN bandwidth, and can meet the performance requirements of the data migration tasks by making full use of the idle bandwidth of the WAN. Secondly, a request access-aware IO proxy resource allocation strategy is named RAAS, which can measure the highest pressure of users accessing the corresponding data space, and can solve the performance bottleneck caused by the inefficient IO proxy resource allocation of the GVDS.
DOI: 10.1007/s11704-023-3087-8