The Challenge
What Mabat Viewer Was Facing
Mabat Viewer needed a SaaS video platform capable of ingesting, processing, and archiving live feeds from hundreds of cameras across geographically distributed US sites. Sending all raw video to the cloud was prohibitively expensive and introduced unacceptable latency for real-time monitoring — but a fully on-premise approach could not give operators the central search and archival capabilities they needed.
The Solution
What We Built
We designed a hybrid edge-cloud architecture. Lightweight edge nodes ran containerized stream processors that performed motion detection and clip extraction locally, sending only relevant segments to the cloud. The cloud platform was built on AWS using Kinesis Video Streams, S3 Glacier for archival, and OpenSearch for metadata indexing. A centralized Prometheus stack monitored stream health across all sites with automated alerting.
Results