Small is Beautiful: Using ClickHouse Data Warehouse for Edge Analytics
September 29 @ 12:15 pm - 12:45 pm
With many edge applications now generating billions of measurements daily, it’s important to have tools that offer enterprise-class analytics on constrained resources. ClickHouse is a popular open-source data warehouse that runs anywhere Linux does. This talk starts with a short intro to ClickHouse features including near-real-time SQL query response and aggregation. After touching on Intel and ARM support, we’ll demonstrate ClickHouse cost-efficiency with a demo of a 500B row dataset running on Intel NUC. We’ll next show how ClickHouse runs on Kubernetes using its own operator. Finally, we’ll show how to roll all of these capabilities into an application design that offers quick local response and efficient aggregation at the edge combined with large-scale analytics on ClickHouse clusters running in the cloud.