Data Projects: Managing Data Assets at Netflix Scale

Netflix11d ago

Building Service Topology at Scale: Architecture, Challenges, and Lessons Learned

Here's a 3-sentence summary of the engineering blog post "Building Service Topology at Scale: Architecture, Challenges, and Lessons Learned" from Netflix TechBlog: To build a reliable and real-time service dependency map at Netflix scale, engineers adopted a "streaming-first" architecture that continuously ingests flow records and provides near-real-time topology updates with backpressure-enabled reactive pipelines to handle massive scale without data loss. A multi-layer architecture with physically separate topology layers and independent optimization enabled the system to process millions of flow records per second and provide sub-second query responses while maintaining near-real-time freshness. Through lessons learned, Netflix engineers found that backpressure and physical storage isolation are crucial components for building reliable real-time systems at scale, and that solving complex engineering challenges requires embracing system complexity and trade-offs.

StreamingScale

25 min

Data Projects: Managing Data Assets at Netflix Scale

In-House LLM Serving at Netflix

Building Service Topology at Scale: Architecture, Challenges, and Lessons Learned

GenPage: Towards End-to-End Generative Homepage Construction at Netflix