Acceleration and Implementation of Database Aggregation Query Based on FPGA

Database is an important carrier to meet information overall management. With the continuous development of global information construction in recent years, the types and quantities of information stored in databases are increasing. Implementing high-speed query response to massive database storage resources through hardware-level optimization design has become an important research direction in related fields. Aiming at the key problems such as low efficiency and high delay existing in the process of performing aggregation operation in the current mainstream database, this paper thorough analysis the principles and key steps of database execution of aggregation queries, and a new FPGA-based database aggregation query accelerator is designed using heterogeneous thinking. The accelerator is designed according to the actual application scenario of relational database aggregation query, and implements the aggregation grouping strategy based on hash algorithm and five commonly used aggregation functions through FPGA; In order to maximize processing efficiency and fully utilize the excellent parallel processing capabilities of FPGA, this study also optimizes the existing aggregation query architecture and proposes a decentralized multi-core collaborative aggregation query architecture to meet the needs of single time period and multithreaded queries. The aggregation query experiments on ZNBase, an open source database platform, show that the proposed database aggregation query accelerator based on FPGA is 21.71 times more efficient than the traditional software algorithm to realize the aggregation query scheme.