Beet Analytics is a software company that provides a transformational solution called ENVISION to manufacturing companies. ENVISION gathers granular data from the PLC devices across an entire assembly line and provides insights into where the assembly line can be optimized.
As Beet began to gain traction at major manufacturers, they were running into the following challenges:
- Deploying their server stack inside a manufacturing facility was time consuming and complicated.
- Aggregating data across manufacturing facilities to derive cross-facility analytics was extremely difficult because the data sits on local servers.
- Scaling the existing solution was becoming problematic.
- Performing machine learning to drive increased value for the customers was not possible with the existing servers and architecture.
Maven Wave worked with Beet to address these challenges by helping Beet modernize their solution using a next generation, GCP powered platform.
Phase 1 was executed very quickly. Maven Wave helped Beet migrate the on-premise application components, which consisted of a Microsoft .NET front end, a custom coded data collector, and a MongoDB database, to run on virtual machines on Google Compute Engine. This drove the total install time down from 3-4 months to 3-4 weeks and allowed Beet to scale the existing platform more easily.
A key step in this transition was enabling all of the PLC devices to stream data real time via Google’s Pub/Sub product to Beet’s Google Cloud Platform-based version. Maven Wave helped architect the solution, recommend data compression enhancements to minimize bandwidth usage, and ran multiple network scenario tests to ensure the new solution would not require additional bandwidth within the manufacturing plant.
In Phase 2, the .NET application was moved to Google App Engine, the data collection components were rewritten using Cloud Dataflow and the MongoDB database was replaced with BigTable and BigQuery. This further increased the ease at which the ENVISION solution could scale and also set the solution up for more advanced analytics and machine learning.
The end result of the move to Google Cloud Platform has reduced Beet’s hardware and software costs by 75%, substantially reduced maintenance costs and downtime. Additionally, the move decreased the amount of time needed for deployment from months to weeks and allowed for increased analytics that add greater value for Beet’s customers.