Authors: Ana Carolina Almeida, Daniel Schwabe, Fernanda Baiao, Maria Luiza M. Campos, Rafael P. de Oliveira, Sergio Lifschitz
Tags: 2019, conceptual modeling
Database tuning is a complex task, involving technology-specific concepts. Although they seem to share a common meaning, there are very specific implementations across different DBMSs vendors and particular releases. Database tuning also involves parameters that are often adjusted empirically based on rules of thumb. Moreover, the intricate relationships among these parameters often pose a contradictory impact on the overall performance improvement goal. Nevertheless, the literature – and practice – on this topic defines a set of heuristics followed by DBAs, which are implemented by the available tuning tools in different ways for specific DBMSs. In this paper, we argue that a semantic support for the implementation of tuning heuristics is crucial for providing DBAs with a higher-level conceptualization, unburdening them from worrying about internal implementations of data access structures in distinct platforms. Our proposal encompasses a set of formally-defined rules based on an ontology, enabling DBAs to define new configuration parameters and to assess the application of tuning heuristics at a conceptual level. We illustrate this proposal with two use case scenarios that show the advantages of this semantic support for the definition and execution of sophisticated DB tuning heuristics, involving hypothetical indexes and what-if situations for relational databases.Read the full paper here: https://link-springer-com.proxy2.hec.ca/chapter/10.1007/978-3-030-33223-5_20
