Startup Takes WebRTC Through the Matrix

WebRTC is a web protocol that promises to turn any browser into a voice and video calling platform without the need for a plugin, but at least one group thinks it will never be more than a series of lonely islands without a standardized signaling plane.

The need for standardized signaling has been debated since Google (Nasdaq: GOOG) first introduced WebRTC in 2011. It defined the media plane, but purposefully left the signaling plane, or how the call connection is made, up to the application. That means that some apps use SIP; some use REST; some use proprietary signaling protocols. (See WebRTC & the Rise of the WebCo and Microsoft's Skype Embraces WebRTC on IE.)

To be clear, many don't see this a problem because they prefer existing protocols or actually want to build proprietary apps. But, with disparate signaling protocols in use by all, interoperability between WebRTC apps will remain a challenge. That's an issue that Matrix.org is hoping to fix, so it's formed a non-profit, open source startup to connect any and all WebRTC apps. (See Decoding WebRTC's Promise & Challenges .)

Matrix wants to create a new standard signaling plane for WebRTC that is open and free for all, based on RESTful, HTTP/JSON application programming interfaces (APIs). According to Amandine Le Pape, the group's business lead, its goal is to break down the silos within the VoIP and IP messaging market to ensure any service can talk to another "over the matrix." Le Pape co-founded Matrix.org in December with Matthew Hodgson, both of whom work at Amdocs Ltd. (NYSE: DOX) building unified communications apps by day. Amdocs is providing funding for the non-profit.

"Matrix can be a solution to link the silos to give the user a specific user experience using the app but still be able to talk to someone else on a different app," Le Pape says. "The idea is that, for the end user, I don't want to have to use the app someone else is using. I use the one I prefer."

But, she stresses, Matrix is not just another aggregation app. It's a free and open source platform for developers to easily create and host their own RTC functionality or add these features to an existing app -- consumer, enterprise or Internet of Things -- using a common language for basic chat and voice communications. (See WebRTC in the Wild.)

For more on this developing protocol, check out the dedicated WebRTC content channel here on Light Reading.

Is there room for telcos in the Matrix?
The challenge for Matrix.org will be to get the support of companies building WebRTC that might see this standardization effort as delaying the technology from really taking off. The group is primarily targeting over-the-top providers, but it also has its eyes on telco members, who will be a tougher sell. For one thing, they have a long history with SIP and want to tie WebRTC to IMS in many cases. What's more, some are participating in their own initiative unifying the signaling plane for WebRTC, Alliance for Telecommunications Industry Solutions (ATIS) Open Project ORCA for Open Real-Time Communications APIs, although that has been slow to progress. (See Ericsson Offers Bowser, OpenWebRTC as Open Source, First ORCA Sighting Expected in Q4 and Killer Apps Meet Killer Whale: ORCA Opens Up.)

Hodgson says Matrix is in the early stages of talking to telcos about joining the ecosystem, and he thinks they are smart to play nice with OTTs, something they haven't done willingly in the past. He explains:

    The way we see it, there is the telco side where it's a certain network and its IMS and PBX interchanges and the PSTN. It has guaranteed quality of service, billing relationships, SIM cards -- all these great things that make the PSTN as robust and ubiquitous as it is today. On the other side you have the Internet, which is complementary. You still have a paywall that is down the middle with PSTN and IMS on one side and the Internet on the other side. Who is best positioned to build the paywall and charge Matrix users that want to talk to the PSTN for sending and receiving phone calls? It will be the telcos that can run those gateways, provide phone number services and provide voice termination.

Some telcos are already wading in with WebRTC, in general. AT&T Inc. (NYSE: T), for one, became the first telco to offer a WebRTC API for developers to build related apps with at its developers conference in January, using existing signaling infrastructure -- most likely SIP -- and offering the ability to extend voice and video calls to landlines and mobile phones, not just browser-to-browser, as well as to have caller ID over the browser. (See AT&T Opens up WebRTC API.)

But Hodgson says it's proprietary to AT&T's network, so it doesn’t solve the federation problem. Adding Matrix as a subset of its API would make it interoperable, he says, and AT&T could still differentiate through offering payments, which is something Matrix doesn't yet cover. (See WebRTC: A Double-Edged Sword for Telcos and What WebRTC Means for Telcos.)

It may be a long road, but Matrix is hitting the ground running. Only a couple of weeks into its beta, the group plans to announce its first partnerships at Mobile World Congress next month, and Hodgson says it has a couple thousand users and 60 servers participating already, including big, well-known companies and "smaller, snappier" startups that he says are all looking to "reclaim communications on the Internet for themselves."

"We are fairly unique in what we're trying to do," he says, "and it's as fundamental as the web or email to create an open source ecosystem on the Internet."

— Sarah Thomas, Circle me on Google+ Follow me on TwitterVisit my LinkedIn profile, Editorial Operations Director, Light Reading

nasimson 2/15/2015 | 11:09:27 PM
So internet like It'll be so internet-like to answer a Skype call on Appear.in and to make a call from fring not worrying about if the recipient has fringe or not. Mathew I'm relieved to know that someone is working on this more important but less featured work.
sarahthomas1011 2/15/2015 | 8:18:30 PM
Re: telcos in the Matrix? Thanks so much for your feedback and additional context, Matthew. That's good to know about ORCA as well. I am looking forward to hearing more about both initiatives and watching Matrix.org progress this year.
arathorn 2/15/2015 | 8:02:57 PM
Re: telcos in the Matrix? Daniel: it's true that for some WebRTC applications it makes no sense to federate. If the conversation doesn't make any sense outside of the context of the rest of that site or app's functionality, then why bother federating it?

But in practice we find it's quite unusual for communication to be meaningless in isolation.  The voip and IM conversations folks have on WebRTC and chat-enabled websites tend to stand alone and be just as meaningful as anything you might have in your email "Sent Items" or similar.  The difference is that the data and metadata gets entrusted and locked away in the random service you're using, rather than being stored somewhere that you control (like your mail history).  For instance, this conversation right now is effectively chat - but it's siloed in LightReading's server.

This is not ideal because i'd like to keep a copy of everything i say online for future reference - and i want to be notified of responses using whatever my preferred mechanism is (e.g. a push notification - in practice, I didn't get any notification from the LR site that there'd been a response).  And if this were a video comment rather than a text comment, I'd much rather be able to record the comment direct from my iPhone using my favourite video recording+filtering app (the camera on my laptop is terrible).  And I'd like to keep track of the video comment and its contents on my own server just as much as publish it on Light Reading - it's my data after all.  Federation (and Matrix) gives you, as the user, this level of control over your online communication.

Obviously you can scrape by without federation, as we all do today.  But it's one of those things where you don't realise how much you're missing until you try it.  Imagine how bad it would be if we were forced to answer emails using the same system as the other person uses (i.e. forced to use Yahoo mail just because they mailed you from Yahoo mail).  Imagine how bad it would be if our phone UIs spontaneously changed when you receive a call to match the UI of the caller rather than showing your phone's actual UI?  Why should we accept this in WebRTC? :)
danielcawrey 2/14/2015 | 9:54:10 PM
Re: telcos in the Matrix? I really see WebRTC as a platforn for different protocols anyway; I don't believe that this is a big issue. We're not going to see interoperability from competing products anyways. That's just not how platforms work. 
arathorn 2/13/2015 | 9:39:37 PM
Re: telcos in the Matrix? Hi Sarah - huge thanks for writing about us!

Whilst Matrix.org itself is unashamedly concentrating on defragmenting the OTT space, we really do hope that the more forward-looking telcos will get involved and take the opportunity to run Matrix services and monetise them.  If they play their cards right and get involved early, they can even be the equivalents of GMail and Hotmail for the Matrix world - providing a carrier-grade SLA for Matrix users as well as monetising VASes and PSTN connectivity on top.  And even if Telcos don't get involved immediately, it's hopefully only a matter of time... assuming Matrix is successful on the Internet.

ORCA is an interesting initiative - we met with Jim McEachern from ATIS/ORCA at WebRTC Expo in San Jose last year, and actually ended up distributing ORCA's marketing materials from our stand :)  ORCA is very complementary to Matrix; it basically provides a standard clientside signalling API and SDK which aggregates different backend APIs (SIP, custom REST, Matrix, etc).  We haven't got to writing a Matrix connector for ORCA yet, but hopefully it's only a matter of time before we or someone else does.

- matthew at matrix.org
sarahthomas1011 2/13/2015 | 12:03:54 PM
telcos in the Matrix? I don't see a lot of telcos getting involved with this, which I think would be okay from Matrix's point of view, because they are more focused on OTTs, who are really leading the charge with WebRTC. However, with Amdocs funding it and wanting to sell to the telcos, I imagine that has to be a priority for them. 

It did remind me to revist ATIS' ORCA intiative though. Again, haven't heard much from it, but maybe that will change at MWC next month...
Sign In