Case Study: eBay achieves an elastic, highly available metadata store for its media platform with MongoDB

An Elastic Metadata Store for eBay’s Media Platform

eBay needed a robust, multi‑tenant, highly available metadata store for its media platform that could meet strict SLAs. Static sharding was insufficient because the service must scale continuously through incremental additions of storage without downtime or manual intervention.

eBay built MStore, an elastic metadata store using MongoDB as the storage engine, with a sharding and orchestration design that supports online resharding and automated scaling. The result is a scalable, highly available storage service for media metadata that can grow incrementally without service interruption and meets the platform’s operational requirements.

eBay

Yuri Finkelstein

Lead Platform Services Architect