A grid-based approach in answering top-k dominating queries on groups

This study aims to describe the implementation of top-k dominating query processing in the group-based data. In this rapidly developing technology era, the data needed is no longer only single but a group, in some cases. For example, in a data center, a user may need to do clustering or CPU grouping. For that, we need a combination of CPU groups that can beat the performance of other groups. Moreover, with the increment of dataset size, the number of possible data group will increase. Therefore, optimal computing costs are needed so that data can be processed quickly. This study provides a solution in the form of processing a top-k dominating query using an indexing grid. Group data indexing task is performed by placing data in groups on a grid based on the aggregate value. From the grid, we can find which cells are candidate cells using the theory of data dominance. The groups of data that occupy the candidate cells are then processed with a dominating top-k query to get the leading groups. Since the topic is newly discussed problem, our proposed method is compared to the naïve one and reaches a 69.5% better performance.