Polyglot Persistence Architecture for Enterprise Content Management

Abstract

The aim of the research is to create and evaluate polyglot persistence architecture for an Enterprise Content Management solution. MS SQL database is used for Current data store that handles the current data while Elasticsearch - for General data store where both current and history data is persisted and queried. The general data store is represented by time (e.g. monthly) spanned indexes on an Elasticsearch cluster of a hot-warm architecture. The proposed architecture is evaluated on a MS Azure cloud hosted Elasticsearch cluster on a several test databases of volume up to 1.14 billion of objects. Various parameter configurations are tested to explore for performance patterns. Results of the performance tests are outlined and suggestions are brought forward on resilience management, performance measurement and management of the cluster in production environment.