An economical query cost model in the cloud

Abstract

The Cloud Computing (CC) brings a new approach of information technology (IT) consumption and is changing the investment manner of enterprises and companies. While the reputation of CC is increasing, a large number of applications managing large amount of data are moving towards the Cloud which incorporates several new dimension: payment, query processing, etc.. The analytical data management applications are an example of those applications. They are intended to the decision support process requiring complex queries. To optimize these queries, optimization structures such as indexes and materialized views are required. In the traditional database infrastructures (centralized, parallel, distributed, etc.), the choice of the optimal configuration of optimization structures is usually guided by mathematical cost models. They are used to quantify the quality of the obtained solutions. The purpose of our work is to develop a cost model to select materialized views in the Cloud. The main characteristic of our cost model is that it considers the payment cost and the query processing paradigm. Intensive experiments were conducted using our cost model and the obtained results are deployed in an assimilated Cloud infrastructure. © Springer-Verlag 2013.