Research Perspectives Toward Autonomic Optimization of In Situ Analysis and Visualization
Zhe Wang, Matthieu Dorier, Manish Parashar · 2022
Due to the growing gap between computation and I/O capabilities of large-scale computing systems, in situ processing has become a widely adopted approach complementary to post hoc processing of data generated by scientific simulations. For scientific workflows composed of simulations and analysis/visualization (ana/vis) applications, in situ approaches enable performing data ana/vis tasks close to the data source and running them on the same system. However, variations in the simulation data and the diversity of underlying High-Performance Computing (HPC) environments increase the difficulty of adjusting the in situ processing configurations, such as when and how to execute in situ tasks, during the workflow execution. Autonomic computing has proven successful in responding to dynamic behaviors of software systems according to the high-level objectives of its users. In the context of in situ processing, triggers are an emerging strategy that follows the autonomic computing paradigm to optimize when and how to execute in situ ana/vis tasks. By inspecting particular indicators, the trigger can flexibly issue customized control instructions to optimize the execution of in situ ana/vis tasks in real time. This position paper formalizes the elements of the trigger mechanism according to the definition of autonomic computing. It then uses this formalization as a guideline to summarize the research status of different aspects of the trigger mechanism for in situ processing. The scenarios adopting a trigger mechanism to optimize in situ processing discussed in this paper include (1) where to execute ana/vis tasks, (2) resource allocation of ana/vis tasks, and (3) when to execute ana/vis tasks. Finally, this paper provides suggestions for future research directions that rely on autonomic computing to facilitate scientific in situ processing.