Interview: Bistri on WebRTC for Video Chat Services

I recently spoke with Arnaud Budkiewicz (AB) of Bistri about their use of HTML5 and WebRTC to build a Web-based video chat service.

IJ: How did you get started on this project?

AB: Two years ago I decided to create a web-based, peer-to-peer video call system based on the standards-track document that was just an early draft at the time. We raised funds and started following the WebRTC Working Group. To help us start building the product while the different editors were working on the implementation of WebRTC, we worked with Flash, using their protocols. Our first focus was on simplicity and interoperability, while we built our technology-agnostic platform.

IJ: Now you've moved to HTML5?

AB: Yes, the Bistri Web application is fully built with HTML5; we only use Flash as a fallback now. As Chrome support for WebRTC has increased, we have seen WebRTC take over the Flash calls.

IJ: What other technologies are you using?

AB: The tool includes text chat, video effects and screen capture as well. So we use XMPP (chat, signaling), HTML5 video, WebRTC, CSS, JavaScript, WebGL (for video effects), Canvas (for screen capture). We plan to use Rich Communication Services (RCS), which is promoted by telcos in Europe in particular.

IJ: What have been the main advantages for you in moving to HTML5?

AB: Of course, WebRTC was a bit part of why we moved to HTML5. But there are other great HTML5 features such as video and canvas.

IJ: If I understand, you have a basic video call service for free. So what is the business model?

AB: The one-to-one capability is free. We also enable people to connect through existing social networks. With a premium account, people will organize teleconferences. Our plans will include enabling more customization and offer persistent file sharing. Right now we're the only B2C provider for video chat using WebRTC.

IJ: Why do you think that is?

AB: The initial reason was "that's our crazy ambition." Beyond that, you need a lot of users to sign up and building a strong user base with an open product is challenging. But I think the timing is right for this, and I expect other companies to follow with similar offerings.

IJ: How long do you think it will take for WebRTC to work interoperably across platforms, including mobile?

AB: I think in 18 months WebRTC will be well-established on laptops and desktops. Mobile is another story due to issues such as memory limitations, chip issues, and battery consumption. As WebRTC matures, we do plan to offer mobile services.

IJ: What about performance?

AB: The network layer raises its own set of issues. Performance on a wifi network is ok, but with 3G and other purely mobile technologies it's more complex. Those networks were not designed with the Internet Protocol (IP) in mind. Telcos need to work on ensuring IP is well-managed on those networks. Ultimately we want a seamless user experience as people move around between wifi and 3G networks.

IJ: You recently joined W3C. What do you observe so far?

AB: We attended TPAC 2012 in Lyon, which was great. As a startup, we have appreciated connecting with people from global organizations. We're also participating in Working Groups. What we need is more practical guidance on participation.

IJ: Whoops! I'm supposed to be working on that project....Thanks Arnaud!