Case Study: Orbeus achieves real-time, scalable photo management for millions of images per day with DataStax

Orbeus Delivers “PhotoTime”, A Revolutionary Smart Photo Management App, Built on DataStax

Orbeus built PhotoTime, a smart photo-management app that uses proprietary image-recognition to tag and organize faces, scenes and objects, and faced a rapid influx of users and data. The company needed to ingest and analyze massive volumes of unstructured images—3–5 million photo/video posts per day (300–500 GB/day)—while maintaining high performance and continuous availability so users could find photos in real time.

Orbeus evaluated several databases and chose Apache Cassandra via DataStax Enterprise for its linear scalability, availability and integrated tools. Using DataStax (with OpsCenter and Spark) to store photos and tags and run real-time analytics, PhotoTime now handles peak traffic with sub‑second responses (over 1,000 calls/sec), 100% uptime during peaks, and the capacity to scale as its machine‑learning models and user base grow.

Orbeus

Tianqiang Liu

Co-founder & VP of Engineering