Is 2024 the Year of WebRTC?
There has been a lot of talk in the industry over the last few years about real-time streaming. Vendors have tried to claim that delays between broadcast and streaming, resulting in the proverbial “you hear someone cheer before you see the goal,” necessitate an ultra-low–latency approach. I don’t buy it, and neither do the streaming operators. I can’t tell you how many times I’ve heard streaming operators who deliver live sports content say, “Yes, latency is important, but reliability, scalability, and quality are far more important.”
When online sports betting opened up in the U.S., those same vendors cheered: “Yay, real-time streaming is going to be big.” Unfortunately, betting has not quite taken off. In fact, one could argue that it has taken off just as much as interactive streaming. According to Statista, the user penetration rate of online sports betting, which currently stands at 1.9% in 2024, is projected to increase only to 2.3% by 2028.
OK, enough of the naysaying.
I do believe that ultra-low-latency use cases are important. In fact, I’ve had plenty of conversations with sports streamers about it, with some telling me they have parallelized their delivery architectures: part HTTP (for the majority of the audience) and part WebRTC (or another real-time approach for the part of the audience that wants to gamble or interact). The issue with WebRTC, though, has always been one of scale. WebRTC infrastructure is more expensive. It costs more to deliver, and it takes a lot more WebRTC servers to handle ultra-low-latency streaming with high concurrency than it does HTTP servers. But cost isn’t the only issue. WebRTC simply doesn’t have the same functional capabilities as HTTP streaming: no SSAI, no DRM.
I recently spoke with a representative from Phenix Real Time Solutions, a WebRTC streaming provider that has been around for almost 10 years beating the drum of ultra-low-latency streaming. I learned how it has solved the SSAI problem (which it actually announced at IBC 2023, although I’m not sure anyone was really listening at the time). Phenix can now deliver server-side stitched ads in real time. What’s more, the
company, like other DRM providers such as EZDRM and BuyDRM, can supply the security that’s needed to keep content safe.
With large sports-streaming operators, WebRTC provides a real opportunity. But those same operators, which spend billions on licensing rights, can’t afford to just swap the ability to stream content in real time for basic OTT functionality like SSAI and DRM. It would seem, though, that Phenix has finally provided the three key points of the “streaming triangle”: scale, security, and ads. When you think about it, that triangle drives all live streaming. If you can’t scale it, keep it secured, and deliver ads, it’s a losing proposition.
The Phenix rep told me about a commercial customer for which the company delivered WebRTC streams to 500,000 concurrent viewers, but with a very interesting behavior: 3-minute connections followed by 5 minutes of disconnection over the course of 45 minutes or so. This Phenix case study demonstrates that the feasibility of a high concurrency (not the millions) under demanding conditions (creating and destroying sessions repeatedly is not server-friendly) with SSAI and security might finally be the real-world proof point that greenlights streaming operators to embrace WebRTC as a viable protocol for those use cases, like gambling and interactivity, that necessitate ultra-low latency. Of course, it’s still going to come down to cost. With WebRTC being a more expensive protocol through which to deliver, streaming operators are still going to
have to make the hard choice about who gets the WebRTC stream and who gets HTTP content. At least the decision will be made on ARPU, not on technical merits.
$20 that the Steelers make a first down on this play, anyone? Anyone?