Execution Planning for Aggregated Search in the Web of Data: A Free-Metadata Approach

An aggregate search system in the web of data makes it possible to search for pieces of information that may not exist entirely in a single source. It provides a unified query interface allowing access to several data sources. One of the major concerns while setting up an aggregate search system is adopting an adequate execution planning strategy, indeed, thanks to execution planning, the system optimizes network traffic while collecting fragments of data, avoids unnecessary executions, and reduces the waiting time for the user. Execution planning strategies in the state of the art often include using metadata or checking the existence of resources before execution. In this paper, we present a solution for execution planning in aggregated search systems without using metadata or consulting the sources. This solution aims to optimize network traffic by avoiding duplicate results and exaggerated size of intermediate data, our method is based only on the analysis of the input query.