Jon Hoffman, Foursquare - MongoDB Days 2013 - #MDBDays #theCUBE
summary
Foursquare moves from a SQL engine to MongoDB about three-and-a-half years ago because check‑ins grow exponentially and require sharding; MongoDB’s auto-sharding and auto‑balancer removes much of the manual hashing and rebalance tooling. Infrastructure team builds higher‑level services and developer tools while breaking a monolith into a service‑oriented architecture. Online data stays in Mongo; regular snapshots are pushed into Hadoop for offline analytics and recommendation signal generation. Predictable database workloads shift off AWS onto owned hardware to cut costs.
AI-written summary · about
