A big data Product Development

Industry

Technology

Amazon EMR Amazon Redshift Apache Hive Apache Spark Scala

Introduction

NashTech converted Zillion’s monolithic architecture to a microservices structure, refactoring and cleaning the code base and implementing deltas for optimised processing. This resulted in a pipeline that could handle five times the load at 1/200th of the original time.

About ZILLION

Zillion is a healthcare company having the goal of improving virtual healthcare. It provides an online interaction management platform that streamlines the entire communication process to create a better participant experience with every interaction.

Challenges

The main challenge for Zillion was to deal with the huge datasets, including the consolidated medical records and participant data from the devices they already use, from smartphones to Fitbits and more. The existing pipelines were processing and loading the whole datasets (in terabytes) each time when the pipeline ran which was taking quite a long time. The challenge was to build the pipeline to implement the deltas so that the data processing and loading occur only on the changed data (inserted/updated) instead of the whole dataset. Another challenge was to reshape the existing unstructured code base using scala best practices and converting the application to microservices-based architecture.

Solution

NashTech quickly converted the whole monolithic architecture to a microservices structure, refactored the code base to make it testable without hindering its actual logic, introduced the test cases, and cleaned the code by using the scala best practices following the Knolway (NashTech software development methodology). Further, NashTech implemented the deltas to process the changed data instead of playing with the large datasets each time, drastically reducing the processing time from hours to seconds. The end result was a pipeline that could handle 5 times the load at 1/200th of the original time.