Focusing on the issue that the HBase storage without spatio-temporal index degrades the traffic data query performance, some HBase spatio-temporal indexes based on row keys were proposed for massive traffic data. Firstly, the dimensionality reduction method based on Geohash was used to convert two-dimensional spatial position data into a one-dimensional code. Then the code was combined with the temporal dimension. Secondly, four index models were put forward based on combination order, and the structures of the models and their adaption conditions for traffic data query were discussed. Finally, the algorithm of index creation as well as traffic data query algorithm was proposed. Experimental results show that the proposed HBase spatio-temporal index structure can effectively enhance the traffic data query performance. In addition, the query performance of four different spatio-temporal index structures in different data size, different query radius and different query time range were compared, which verified the different adaption scenes of different index structures in traffic data query.

Traffic stream data has characteristics of multi-source, high speed and large volume, etc. When dealing with these data, the traditional methods and systems of data storage have exposed the problems of weak scalability and low real-time storage. To address these problems, this work designed and implemented a HBase-based real-time storage system for traffic streaming data. The system adopted the distributed storage architecture, standardized data through front-end preprocessing, divided different kinds of streaming data into different queues by using multi-source cache structure, and combined the consistent Hash algorithm, multi-thread and row-key optimization strategy to write data into HBase cluster in parallel. The experimental results demonstrate that, compared with the real-time storage system based on Oracle, the storage performance of the system has 3-5 times increment. When compared with the original HBase, it has 2-3 times increment of storage performance and it also has good scalability.

The log-based index update mechanism in flash-based database system has following shortage: low query efficiency, expensive update cost, unreasonable space allocation and merge for the log. In order to solve these problems, a new adaptive index structure named LM-B+TREE was proposed. LM-B+TREE can map the page for index update buffer into corresponding node of traditional B+ TREE. Furthermore, according to the read/write workload and read/write overhead, LM-B+TREE can dynamically maintain the update buffer and adjust the index frame adaptively. The experimental results show that LM-B+ TREE can dynamically adjust the index structure to adapt to the read-write workload, significantly reduce the overhead of index update and improve the query performance.