Multi-Model Persistent Solution for Healthcare Big Data

Provision of an uniform query interface facade to access the health-care data present in multiple data-stores being used autonomously by various departments of an hospital, will empower the health-care community to make better decisions and conclusions. One naive approach for implementation of such system is to use any one popular database to store and process all the data pertaining to all the departments of an hospital. Nonetheless, any single data-model cannot efficiently store and process multitude of data generated by healthcare institutions. For example, not all data fit well into row-column format of traditional relational databases. However, modern NoSQL data-stores allow data to be stored in a form closer to their actual representation and usage. Storage and retrieval of data from disparate data-stores has its challenges and issues like dealing with multiple querying languages, understanding different data modeling techniques and, creation and maintenance of Knowledge Base of Data (KBoD). We have developed an intelligent information integration solution named Polyglot-persistent Healthcare Information System-PolyglotHIS which makes use of cooperating agents enabling health-care professionals to retrieve data from heterogeneous data-stores. PolyglotHIS uses multiple data-stores for storage and processing of the HIS data. The rationale is to select the most appropriate data storage technology that meets the specific requirements of each module of the HIS. Architecture of PolyglotHIS consists primarily of multiple cooperative agents. Capabilities and contents of data-stores are stored and inferred using Datalog, a declarative logic programming language used to store set of facts and rules. Design and working principle of PolyglotHIS is illustrated with the help of a running example. Performance analysis of the implemented system showcase that very less latency has been induced by the system.