In this short talk, the speaker critiques the notion that public cloud alone can meet the demands of large‑scale WebRTC deployments. He highlights the cost and control limitations of major public cloud providers when handling predictable, global traffic. To address this, he proposes a hybrid cloud architecture: run steady, predictable workloads on owned bare‑metal servers or on lower‑tier IaaS providers to gain cheaper, faster, and more controllable infrastructure, while reserving public clouds (AWS, GCP, Azure) as a burst capacity for viral spikes or datacenter outages. This approach improves resilience, avoids vendor lock‑in, and balances private and public resources. He ends by prompting the audience to consider whether they are truly balancing private and public clouds for their media servers or relying on a single provider.
