Daily's developer platform powers real-time audio and video for millions of people all over the world. Our customers are developers who use our APIs and client SDKs to build audio and video features into applications and websites.
This week we're taking a step back to write about the infrastructure that underpins everything we do at Daily. We’ll cover:
- Our Global Mesh Network, which routes packets with a first-hop latency of 50ms or better over a geographic area that covers five billion people
- How we support Interactive Live Streaming use cases – up to 100,000 participants with 25 presenters (cams/mics on) at real-time (sub-200ms one-way) latencies
- Our simulcast implementation, which we've fine-tuned over the past six years of serving WebRTC traffic globally at production scale
- Daily’s Video Component System, a cloud-native developer framework for creating dynamic recordings and RTMP streams
Our focus is enabling the best video experiences in the world. Here's how we think about that work — read on for more below. You also can listen to our teammate Chad, a Solutions Engineer here at Daily, who's sharing videos on topics throughout the week.
Call Quality: Our Customers and our Customers' Customers
First, call quality and reliability is our top priority, and we have a specific framework for this. The end-user experience — the experience of our customers’ customers — is the driver for everything we do. Our infrastructure’s “job to be done” is delivering the highest possible quality audio and video to everyone in the world, on any network and on any device, all the time.
We are engineers building tools for other engineers, so we think a lot about the developer experience of using our APIs and SDKs. We aim to handle all of the complexity of real-time video automatically, while still giving developers the flexibility to tune and customize everything.
Unlocking New Possibilities and New Use Cases
We also organize our engineering work and our product roadmap around the goal of expanding video use cases — and, relatedly, helping drive the rapid evolution of WebRTC in recent years. The developers and companies we work with are building for a video-first world. Supporting larger and larger numbers of participants at real-time latencies, building more features into our core APIs, and continually pushing costs lower are part of how we contribute to the future of video.
We also hope that what we build has positive ripple effects beyond just our own products. Daily is a member of the W3C WebRTC Working Group; we contribute to webrtc (peerconnection), device (gum, screenshare), and media groups (webcodecs). We also support and contribute to a number of open source projects, including Mediasoup and GStreamer.
Engineering at Scale
Our perspective on how to build infrastructure that supports millions of people, all over the world, and delivers the highest possible video and audio quality is anchored in our experience as network engineers who have been interested in distributed systems, video, and real-time networking for a long time. Many of us at Daily have been writing video-related code since the very early days of video on the Internet. So we have a lot of strong opinions about things like codecs, bandwidth estimation, and UDP packet routing!
We love to write and talk about this stuff. If you’re interested in these topics, too, please check out the rest of our infrastructure posts this week, join us on the peerConnection WebRTC forum, and find us online or IRL at one of the events we host.