Master of Science (MS)
NoSQL databases provide an edge when it comes to dealing with big unstructured data. Flexibility, agility, and scalability offered by NoSQL databases become increasingly essential when dealing with geospatial data. The proliferation of geospatial applications has tremendously increased the variety, velocity, and volume of data that the data stores must manage. Such characteristics of big spatial data surpassed the capability and anticipated use cases of relational databases. Because we can choose from an extensive collection of NoSQL databases these days, it becomes vital for organizations to make an informed decision. NoSQL Database benchmarks provide system architects, who shoulder a considerable burden of selecting the right technology for their data stores, with a vital start point and source of information. The major utility of these benchmarks is reproducing experiments on similar experimental data that can verify and optimize the process of selecting an optimum tool for data management needs in the early phases of the development. The goal of this research is to develop a benchmark that can compare the performance of NoSQL databases for querying complex geospatial data. We have analyzed throughputs, latencies, and runtime of MongoDB and Couchbase to identify the correct fit for our use case. This way we have also demonstrated a systematic process that can be followed to make an optimum choice of datastore. This benchmark can be extended easily to any NoSQL database that supports geospatial querying.
Kanwar, Yuvraj Singh, "Benchmarking Scalability of NoSQL Databases for Geospatial Queries" (2019). Master's Projects. 673.
Available for download on Sunday, May 17, 2020