Edge Computing: Chick-fil-A Style

Chick-fil-A’s tech team recently put out a detailed post on their extensive edge computing deployments across their stores. Their overarching thesis is that more data at the restaurant level will help them scale better. In their own words:

Our hypothesis: By making smarter kitchen equipment we can collect more data. By applying data to our restaurant, we can build more intelligent systems. By building more intelligent systems, we can better scale our business.

One example they reference is building a more accurate forecasting model for the number of fries that should be cooked over every minute of the day. The status quo is a forecasting model running in the cloud  that uses transaction level sales data from many restaurants. The issue is that it is not reliable enough to drive food production due to the local factors that influence a certain restaurant (weather, sports, traffic). Instead, an edge deployment coupled with IoT-enabled devices in the store could POS system keystrokes and data from fryers about the status of what’s cooking to make actionable predictions.

Their early bet on data and investment in infrastructure is prescient as new startups enter the food space with high hopes. Robotic restaurants such as Spyce, Creator, and Zume Pizza are hoping that automation will enable them to scale quickly and profitably. One of the few advantages that incumbents have on these new startups is the sheer number of stores and customer interactions. Chick-fil-A understands that value and importance of investing to capture and use that data in a meaningful way.

In terms of the infrastructure itself, they’re running 2,000 small clusters with tens of containers per cluster compared to the well-known players like AWS, Google Cloud and Azure who run 100x-1,000x more containers in a few large containers. Unsurprisingly, they’re using Kurbenetes for orchestration. They’re using commodity hardware for $1k/restaurant. They’re cloud-first when possible with fallback onto edge.

And at the end of the day, all this infrastructure is in service of a single goal: more customers “Eating More Chicken.”

Leave a Reply