| |
 |
| Paper: |
Evaluating a NoSQL Alternative for Chilean Virtual Observatory Services |
| Volume: |
495, Astronomical Data Analysis Software and Systems XXIV (ADASS XXIV) |
| Page: |
409 |
| Authors: |
Antognini, J.; Araya, M.; Solar, M.; Valenzuela, C.; Lira, F. |
| Abstract: |
Currently, the standards and protocols for data access in the Virtual Observatory architecture (DAL) are generally implemented with
relational databases based on SQL. In particular, the Astronomical Data Query Language (ADQL), language used by IVOA to represent
queries to VO services, was created to satisfy the different data access protocols, such as Simple Cone Search. ADQL is based in
SQL92, and has extra functionality implemented using PgSphere. An emergent alternative to SQL are the so called NoSQL databases,
which can be classified in several categories such as Column, Document, Key-Value, Graph, Object, etc.; each one recommended for
different scenarios. Within their notable characteristics we can find: schema-free, easy replication support, simple API, Big Data,
etc. The Chilean Virtual Observatory (ChiVO) is developing a functional prototype based on the IVOA architecture, with the following
relevant factors: Performance, Scalability, Flexibility, Complexity, and Functionality. Currently, it's very difficult to compare
these factors, due to a lack of alternatives. The objective of this paper is to compare NoSQL alternatives with SQL through the
implementation of a Web API REST that satisfies ChiVO's needs: a SESAME-style name resolver for the data from ALMA. Therefore,
we propose a test scenario by configuring a NoSQL database with data from different sources and evaluating the feasibility of
creating a Simple Cone Search service and its performance. This comparison will allow to pave the way for the application of
Big Data databases in the Virtual Observatory. |
|
|
|
|
|
