WebRTC Still a Journey, Not a DestinationWebRTC Still a Journey, Not a Destination
In 2019, there's still no one-off-and-we're-done approach to WebRTC -- yet.
March 27, 2019
WebRTC enables developers to add live voice and video communications to websites and applications. It’s great, but it’s also still rather new, and thus evolving.
If you plan on using WebRTC for one of your business applications, be aware that WebRTC requires not only an initial development investment but also an ongoing investment -- and this ongoing investment is going to be with you for the foreseeable future.
I’ve got a personal website, written in Hebrew, where I last wrote in August 2018. Before that? August 2016. Before that? February 2016. And before that? August 2013. You can say that my own professional site takes most of my time these days, while my personal site gets a bit neglected. The interesting thing is, however, that whenever I go to my personal website to write something, it just works. There’s nothing that I have to do to get it to work; it just works.
I am hosting this personal website using WordPress, on a managed hosting service, so I don’t have to deal with security, maintenance, backup, or other headaches. The WordPress release is even automatically upgraded for me, as well as PHP and other technologies that are used. When these updates occur, nothing breaks as a result. Somehow, that website has stood the test of time and the software rot that can often take down other systems.
With WebRTC, though, would you expect something you develop and deploy today to work just as well a year or two from today? If you would, then there’s definitely an expectation issue there. Here are a few pointers for you to consider:
WebRTC 1.0 isn’t out yet. It’s coming, but no one knows when exactly
Google has been making and continues to make changes to its WebRTC implementation in Chrome:
A new echo canceller was just recently introduced
Its whole audio threading model got revamped
There’s a new experiment of mDNS going on
Microsoft is replacing the engine used in Edge with Chromium
Down the road, there are thoughts around introducing QUIC to WebRTC as well as using more of WebAssembly in WebRTC
You should add to all that the pace at which browser vendors are shifting their WebRTC APIs to align with the upcoming 1.0 specification (causing more instabilities to existing WebRTC applications along the way).
If you don’t keep pace with all this change, your code rots and becomes irrelevant, and this happens way faster with WebRTC than with many other technologies.
Don’t get me wrong, WebRTC is great -- there’s no other real alternative. But it isn’t a destination. You don’t sweat it out to get WebRTC into your product and then sit back and relax or move on to the next shiny technology. With WebRTC, you need to continue at it. You need to maintain a development team with the role of keeping the lights on, making sure things don’t break with new browser releases, and maybe adding more features as they become available.
WebRTC is a journey, not a destination.
