Case Study: Parse achieves sub-100ms latency and scalable MongoDB performance with Amazon Web Services (EBS Provisioned IOPS)

Parse - Customer Case Study

Parse is a San Francisco–based startup that provides cloud-based backend services for mobile app developers, supporting more than 180,000 apps on over 200 million devices. The company needed to improve scalability and consistency for its high-throughput, I/O‑intensive MongoDB clusters because usage can spike dramatically (for example when apps are featured), and it cannot tolerate downtime or periodic latency spikes that would impact thousands of customer apps.

Parse migrated its databases to Amazon Web Services using Amazon EBS Provisioned IOPS with striped volumes on EC2, Elastic Load Balancing, and S3-backed snapshots for fast node rebuilds. As a result, end-to-end latency fell from peaks of ~400 ms to averages under 100 ms, latency spikes were eliminated, memory warm-up time dropped by over 80%, and the platform gained the predictable, low-latency performance and rapid recoverability it required.

Parse

Charity Majors

Manager, Parse