Case Study: PingThings cuts grid analytics compute costs by 90% with Blue Orange Digital

A Blue Orange Digital Case Study

Preview of the PingThings Case Study

PingThings cuts grid analytics compute cost 90% with Blue Orange Digital

PingThings was facing a challenge with the high cost and slow performance of running analytic queries on petabytes of electric grid sensor data using AWS EMR. These costs were breaking the unit economics of their business, making customer contracts unprofitable and forcing engineers to focus on infrastructure workarounds instead of product development.

Blue Orange Digital implemented a solution that moved the time-series workload off EMR and onto the Berkeley Tree Database (BTrDB), which is built for high-rate sensor data. This new architecture, which included Kafka, Spark, and TensorFlow models, reduced analytic query times from 30-40 minutes to seconds. As a result, Blue Orange Digital helped PingThings achieve a 90% reduction in monthly compute costs and a 10x or greater reduction in storage footprint.


View this case study…

Blue Orange Digital

49 Case Studies