Self-driving database systems: a conceptual approach
- PDF / 1,034,504 Bytes
- 23 Pages / 439.37 x 666.142 pts Page_size
- 66 Downloads / 186 Views
Self‑driving database systems: a conceptual approach Jan Kossmann1 · Rainer Schlosser1
© The Author(s) 2020
Abstract Challenges for self-driving database systems, which tune their physical design and configuration autonomously, are manifold: Such systems have to anticipate future workloads, find robust configurations efficiently, and incorporate knowledge gained by previous actions into later decisions. We present a component-based framework for self-driving database systems that enables database integration and development of self-managing functionality with low overhead by relying on separation of concerns. By keeping the components of the framework reusable and exchangeable, experiments are simplified, which promotes further research in that area. Moreover, to optimize multiple mutually dependent features, e.g., index selection and compression configurations, we propose a linear programming (LP) based algorithm to derive an efficient tuning order automatically. Afterwards, we demonstrate the applicability and scalability of our approach with reproducible examples. Keywords Database systems · Self-driving · Recursive tuning · Workload prediction · Robustness
1 Self‑driving database systems The topic of database systems that change their configuration autonomously came to recent popularity in academia [19, 23, 31] and industry [10, 30]. According to Chaudhuri and Weikum [5], the costs for database personnel are a major factor in the TCO of database systems. These costs can be further increased by a higher complexity of configuration and tuning tasks which is caused, e.g., by non-stable workloads that change over time, a lack of domain knowledge and application context [10] in cloud environments, and more available dependent configuration options [39].
* Jan Kossmann [email protected] Rainer Schlosser [email protected] 1
Hasso Plattner Institute, University of Potsdam, Potsdam, Germany
13
Vol.:(0123456789)
Distributed and Parallel Databases
Therefore, systems that are capable of autonomously adjusting their configuration could save costs and lead to more efficient configurations. However, such systems face a multitude of challenges, e.g., finding efficient solutions for configuration problems in a scalable fashion [39], predicting future workloads [23], and learning from past self-management decisions. The topic also offers interesting questions from a database system integration and architecture perspective since systems were usually not designed with such capabilities in mind. We conducted interviews with industry database architects. These showed that low overhead, a maximum of 1% of additional runtime introduced by such capabilities, and minimally invasive changes to the architecture are mandatory. However, most of the work in this area mainly focuses on single aspects while holistic approaches remain unexplored. Contribution This paper is an extended version of [18]. The main contributions of [18] are the following. We present the concept of a component-based framework for self-drivi
Data Loading...
