The Customer Challenge

A recognized industry leader in media cost data has a reporting and analytics application that delivers television rating data to its customers to understand and make better pricing decision as related to TV advertising.

The client had an old client server application built in Foxpro that had to be installed and maintained at 300+ different customer locations with many of them being on different versions and some of them even customized specifically for that site. This application is difficult to manage and nearly impossible to work with – enhancements and data sets are pushed to each installation on a daily basis. Adding new features was tough due to all the product versions, but when it came to adding a new “Big Data” source to the application, there was no choice but to rebuild the application. The main benefit the previous application was speed. Because everything is localized to a customer’s device, the speed of data retrieval was generally not an issue although the existing data volumes were relatively small.

The Intersys Solution

Intersys worked with the client to make the decision to rebuild the application from the ground up on Amazon Web Services (AWS) using modern technologies and frameworks. The UI work is built with Javascript and the services layer with Java. The back-end technology selection was the most difficult because the need for the right combination of price and performance to meet the application usability goals and also meet ROI expectations for the product.

After testing a similar workload on multiple platforms including RedShift, Spark, DynamoDB, and Hive, the best choice to meet our data ingestion and serving requirements was Elasticsearch. It was able to meet or exceed performance expectations across all nine different types of request workloads during maximum expected user concurrency.

Results

  • Lower cost of maintenance and support
  • Ability to buy vendor support
  • Easier to find and attract talent to work on the application
  • Services layer removing all the point to point integration
  • Ability to keep adding new and expanded data sources
  • Consistency of architecture across other applications
Share this:

Ready for your Custom Implementation?

So are we! Let's get started.