Vecstra: An Efficient and Scalable Geospatial In-Memory Cache.
Yiwen Wang · Research at the University of Copenhagen (University of Copenhagen) · 2018
Nowadays a huge amount of geospatial data are generated and widely used in a variety of areas. In this way, there is an extensive range of database systems that support geospatial services but fall behind in coping with increasingly demanding high throughput, low latency and efficient resource usage requirements. Meanwhile, in-memory databases have become more prevalence due to increasing capability as well as decreasing price of the memory. To take advantage of the fast I/O speed and overhead reduction of memory-based storage, in-memory geospatial data management systems are created to overcome drawbacks of disk-based geospatial databases. However, to the best of our knowledge, current in-memory caches are not fully-featured and do not completely utilize the benefits of widely used geospatial standards. To fill this gap, we specialize to the API of standard geospatial services and seek to achieve by this method a much higher performing in-memory geospatial cache implementation than possible by combining generic components such as a geospatial application server and SQL database. To that end, an efficient and scalable OGC standard-compliant in-memory geospatial cache Vecstra is built. Furthermore, we conduct experiments on Vecstra and analyze the preliminary results to formulate research opportunities