When Google’s Product Manager Serge Lachapelle took center stage to kick off the second day of the WebRTC Conference & Expo, taking place from June 17-19 in Atlanta, Georgia, all eyes were on the thought leader to understand what it took to make WebRTC a reality.
And perhaps unbeknownst to many, much of the innovation began with a group of 10 budding entrepreneurs working at a startup company called Marratech. For Lachapelle, he first got involved in the world of WebRTC in 1995 as a student at the University of Montreal who was interning at Ericsson. During his internship—in which he was in charge of billing runs and testing software—he was introduced to a number of Swedish contractors working in Montreal who convinced him to travel overseas and become an exchange student.
“What I saw when I arrived in Sweden blew my mind,” Lachapelle told a packed keynote room this morning. He instantly connected with their software that allowed groups of people to interact with audio video in real time and set out with a research group to build prototypes around this open source software to improve engineers’ collaborative work over distance.
But quickly, the tools became too complex and popular for the university to handle, according to Lachapelle. There were bugs that needed to be fixed and expertise was needed. And so the researchers formed Marratech in 1998 and built tools to allow large groups of people to interact with video over the Internet. For nine years, the founders of Marratech worked on a wide range of challenges, from core engineering to business plans and investor relations. They also focused on business development and product management. But they ran into problems.
“We were way the hell too early,” Lachapelle said, with a laugh. “We had the ideas, the use cases and a large group of people who wanted to see it happen. We had a great network stack that supported multicast, unicast and end-to-end encryption for large conferences.”
“That was a key asset we had, but as entrepreneurs we hit our first RTC wall, and it was a very tough lesson,” he added.
Specifically, Marratech contended with single processing challenges. It searched to create an immersive audio experience, but it was “just too hard.” The startup did not have the single processing chops to pull it off. Moreover, the company struggled with OS and Java updates. After all, it was just a 10-person company.
While it enjoyed fast success (Apple became its exclusive reseller in North America and a number of prestigious universities began running the company’s software), the speed bumps kept coming.
“Looking back, I have to remind ourselves that we had a lot of users—happy users—and our software was used in the way that it was meant to be used, for collaborative work, but we were too early and solving these issues was exhausting,” Lachapelle explained.
In 2007, Marratech’s software code was sold to Google and key employees joined the search engine company, including Lachapelle. Suddenly, the Marratech team was solving the same challenges it had previously but on a “much larger, much scarier scale.” Engineers weren’t just updating software for hundreds of thousands of users; they were doing it for millions upon millions of users. They did not just have to test on Mac, Windows and Linux; they had to test on Safari, Chrome and Internet Explorer.
The Google team was addressing a myriad of hurdles from integrating separate audio and video stacks to offering seamless plugin updates to creating easy install experiences. It was also grappling with the fact that there was not much video going on across the web and the market was beginning to stagnate.
That’s when Google got a call.
The team at Chrome, which had just launched a year earlier, had performed a gap analysis on the differences between the web platform and native desktop experiences and identified problems with the desktop experience like poorer quality graphics, audio and geolocation, among other things. Perhaps the biggest gap was human communication and the ability to speak and have that voice be heard on the other end using the web.
“We decided to look at how to get this rolling,” Lachapelle said. “We asked: Do we hire a team of processors? Go to the open source community? In the end we decided to acquire a company in mid-2010, and it was a great buy. We now had the components to bridge the gap.”
To formerly launch webRTC, Google set out to engage the community, beginning with an informal lunch with key executives from Cisco, Ericsson, Skype and others from the ITF community. In 2011, webRTC was formally introduced.
Since then, webRTC has enjoyed considerable milestones, like when Chrome and Firefox were able to talk to each other for the first time, when Google added the first webRTC peer connections and began iterating on that and when the company added components like screen casting and SCTP data channels. But there are more advances to come and Lachapelle teased participants with what’s on the horizon for webRTC including:
- The new version of Chrome, No. 37 which will be released in a few weeks
- Better HD resolution for longer periods of time
- Faster connect times (think milliseconds)
- The ability to connect devices from JavaScript
- Tighter integration with Chrome
In short, webRTC is growing up, explained Lachapelle.
“As parents know, once a child starts to walk, the game changes forever,” he said. “You lose any sense of security or confidence in life. Once a child starts to walk, the next thing he or she does is start to run, and webRTC is now running.”
Edited by
Maurice Nagle
