AlgoInsight

System design interview question

Design a Distributed Video Transcoding and Adaptive Bitrate Streaming Platform like YouTube at Scale

MediumHLDHLD practice

Design a system that ingests, transcodes, and delivers video content to 500M monthly active users across 190 countries. The platform must handle 1M concurrent viewers, support uploads of videos up to 24 hours long, and transcode each video into 8-12 different bitrate/resolution combinations within 2 hours of upload. The system must serve video segments globally with <200ms latency for 99th percentile viewers, handle regional bandwidth constraints, support live streaming alongside on-demand content, and provide resilience against hardware failures without re-transcoding or viewer interruption. Key constraints include cost optimization for storage and compute, efficient utilization of heterogeneous transcoding infrastructure (GPUs, CPUs), prevention of duplicate transcoding work, and compliance with regional data residency requirements.

Practice this question step by step with an AI interviewer that reviews your requirements, your architecture and your trade-offs, and scores the design.

Distributed SystemsVideo ProcessingMessage QueuesTask SchedulingObject StorageCDN ArchitectureDatabase DesignWorkflow Orchestration

Practice this question