Indexación semántica de documentos: una perspectiva distribuida

Abstract

The following research is based on a study to improve the functionality of the Latent Semantic Index (LSI) methodology as a search engine in document management from a distributed point of view with the use of micro-services, applied in a file repository called Corpus. For this, software tools have been used to improve the limitations in monolithic applications, reducing the response time offered by these types of applications, avoiding losses in server request. The main software tools that have been used to process the data in applications are: Redis, Docker, Kubernetes. Concurrency and load tests have been carried out in the comparison of the two applications demonstrating how Kubernetes allows the application to remain functional and without losses in the requests, unlike a monolithic application. These applications are implemented in a Google Platform to simulate a test environment with equivalent physical characteristics.