Produced Data Vault, Scaled to 3 Petabytes (circa 2003)

• still growing today
• Sub-second query responses
• Zero re-engineering to add more data, and expand

This project was done to serve the needs of the public.  They captured real-time streams that focused on image processing and image recognition.  The premise is: capture images from multiple angles, determine the interesting points of the images and where they overlap.  Utilizing proprietary hashing algorithms, they were able to make sense of the images and the capture points in the images – effectively tagging them with business keys and other attributes that drive their BI analytics.

A hugely successful implementation which continues to grow today.