Multi-Intention-Aware Configuration Selection for Performance Tuning

Abstract

Automatic configuration tuning helps users who intend to improve software performance. However, the auto-tuners are limited by the huge configuration search space. More importantly, they fo-cus only on performance improvement while being unaware of other important user intentions (e.g., reliability, security). To re-duce the search space, researchers mainly focus on pre-selecting performance-related parameters which requires a heavy stage of dynamically running under different configurations to build per-formance models. Given that other important user intentions are not paid attention to, we focus on guiding users in pre-selecting performance-related parameters in general while warning about side-effects on non-performance intentions. We find that the con-figuration document often, if it does not always, contains rich in-formation about the parameters' relationship with diverse user intentions, but documents might also be long and domain-specific. In this paper, we first conduct a comprehensive study on 13 representative software containing 7,349 configuration parame-ters, and derive six types of ways in which configuration parame-ters may affect non-performance intentions. Guided by this study, we design SAFETUNE, a multi-intention-aware method that pre-selects important performance-related parameters and warns about their side-effects on non-performance intentions. Evaluation on target software shows that SAFETUNE correctly identifies 22–26 performance-related parameters that are missed by state-of-the-art tools but have significant performance impact (up to 14.7x). Furthermore, we illustrate eight representative cases to show that SAFETUNE can effectively prevent real-world and critical side-effects on other user intentions.