Challenge
The Defense Intelligence Agency (DIA) was faced with replacing a 20+ year old database system that housed intelligence data globally for the enterprise and was a slow and manual process for data analysts to provide relevant and accurate data from a multitude of siloed data sources while maintaining referenceable data provenance.
Approach
ITC used a new scalable serverless architecture to expand DIA’s data ingestion and exploitation capabilities. We combined diverse data streams with the power of cloud/high-performance computing (HPC) to enhanced capabilities in Machine Learning (ML) and advanced predictive analytics.
Solution
ITC developed the serverless Hivemind application, a dynamically scalable platform for HPC built around a database design using AWS’ DynamoDB. We extended Hivemind’s functionality to create generalized exploratory HPC architectures to create a space where analysts can develop predictive and machine learning models that are deployed on the server-less HPC architecture. ITC used the AWS SAM framework to integrate both code and architecture into a single model composed in YAML, providing a unified method to define, configure, and deploy components of the Hivemind application.
Value-Add
  • Reduced server costs by 30%+ while providing an easily scalable infrastructure
  • Provided DIA with a unified method to define, configure, and deploy components - saving the team 2+ hours per deployment
  • Provided modern geospatial drive dashboard for visualizing and controlling data
Joel Klein
Director of Data Services