What’s up with WebRTC Video as a Service in 2017

13/03/2017

Is it finally the year of video? Who knows.

Up until recently, there wasn’t much out there for anyone who wanted to really use video in his use case. In the past few months we’ve seen so many announcements and moves that makes video as a service seem almost commonplace.

Who are these new and old players and what is it that they are doing in 2017?

TokBox

TokBox has been the main player when it comes to multiparty video with WebRTC. In 2012, they got acquired by Telefonica, but left as an independent entity. This gives them stability of sorts.

Their main competitor – AddLive – got acquired in 2014 and taken off market. Other vendors tried to enter that same niche, but no one caught on in the same way.

What’s interesting is the path that TokBox selected for itself. It is resisting the path of connecting to legacy directly. Their offering doesn’t include PSTN or phone numbers. On the other hand, they have been progressing and expanding their video support into broadcast.

Their recent announcement tells a story of large scale live broadcast:

  1. 3,000 real-time interactive users. TokBox is probably achieving that by cascading their SFU infrastructure – not an easy feat by all means – especially if you consider the large variety of customers and the dynamics of sessions that they need to endure
  2. Adding RTMP support. TokBox had HLS support already. HLS makes it easy to stream video, but doesn’t work well for live broadcasts. RTMP does a better job at that – and enables connecting to YouTube Live, Facebook Live and others

It seems that any type of workload that relates to real time and video is where TokBox is.

Twilio

Twilio is getting ready to become a serious video player.

In 2015, almost two years ago, Twilio came out with video chat capabilities. Since then, it stayed mostly in beta, with 1:1 video chat support only.

Last year, Twilio acquired Kurento (or at least some important parts of it along with the its developers). This acquisition of Kurento was about “programmable video” – the ability to do multiparty, but also much more. Last month? Twilio introduced a new Rooms API – a first step in offering multiparty video.

For Twilio, this is a catch-up game, trying to fit the many requirements of video. It starts with multiparty video, moves on to recording, then hybrid support of video/voice/telephony and from there to live broadcast and whoever knows what comes next.

In 2017, Twilio probably won’t be the best of breed solution for video, buy hey – if you’re already using them for voice and/or messaging – it makes sense to source video from the same vendor.

Vidyo.io

Vidyo.io is new to CPaaS but they are not new to video.

Vidyo have been working over a decade on video conferencing. Longer than other CPaaS vendors. They did that in the enterprise, offering multiparty video using SVC technology. This gave them an edge in media quality competing with enterprise video conferencing giants Cisco and Polycom at the time.

Now they are bringing this technology and their know how to the cloud and developers via Vidyo.io. And it doesn’t hurt that they are collaborating with Google on getting SVC into VP9

They are also an enterprise player, which is where the action seems to be today for CPaaS.

One has to wonder how this is going to affect the other players in this space.

Others?

There are other players out there in CPaaS. Some offering video from the start while others adding it onto their voice offering. Notable names include Agora, TrueVoice and VoxImplant. Each with his own flavor. Each with his own story.

How big is this pond and is there room for so many players?

Can’t decide which  CPaaS vendor to use? Check out my CPaaS vendor selection matrix:

Is this only about CPaaS?

While we’re on this topic, what about enablers? Vendors or frameworks who offer the ability to use WebRTC video in your own installation – cloud or on prem? There we had only Jitsi in the past, but now?

  • We have Jitsi and Kurento as the leading open source frameworks
  • There’s mediasoup and medooze, building SFUs
  • There’s the Intel platform
  • There’s SwitchRTC
  • And FrozenMountain, who is working on their own multiparty offering

More choice than ever before?

Here’s the thing though… There is still no one-size-fits-all.

Responses

Lawrence Byrd says:
March 13, 2017

Thanks for your comments, Tsahi. To integrate with legacy environments, TokBox provides SIP Interconnect from WebRTC-based video sessions in the TokBox cloud to traditional telephony and unified communications systems. Read more on our website. But you are right that TokBox is completely focused on Internet video (backed up by good audio, messaging, screen sharing, broadcasting, distributed SFU, cloud-level control etc etc) and not on selling PTSN functionality that can be found elsewhere. Focus is good :).

Reply
    Tsahi Levent-Levi says:
    March 13, 2017

    Lawrence – thanks for the clarification. SIP interconnect is probably the next-best-thing to real PSTN connectivity with all of its “richness”. Most vendors I’ve seen who end up using TokBox, do so for its video capabilities and not for its PSTN connectivity.

    Oftentimes, I’ve noticed vendors who would look elsewhere because “they need to connect a phone into a video conference”. That being the case, I suggest beefing up the documentation in that specific area.

    Reply
Anton Venema says:
March 13, 2017

Thanks for the mention, Tsahi. LiveSwitch is currently slated for release in June, so there will be a lot more information coming in the next few months. It is a bit unique in that it operates as a licensed SFU/MCU hybrid product, where clients can connect in as either forwarded or mixed participants. The server manages the bridging/scaling internally, so you can simply connect as whatever makes sense for the endpoint capabilities. The connections are all client-driven, so a participant can switch their mode at any time. True to WebRTC spirit, it is also signalling agnostic, so the included gateway can be extended to support third-party signalling systems as well as VoIP, PSTN, and CDN. Anyone looking for more information as it becomes available can simply sign up to our newsletter at http://www.frozenmountain.com.

Reply
A says:
March 14, 2017
Reply
    Tsahi Levent-Levi says:
    March 14, 2017

    Thanks for the shoutout.

    Do notice that this article is about video as a service. The mention in the end is of frameworks you can use to build and deploy your own and it is by no means exhaustive – even the list of CPaaS vendors supporting video isn’t exhaustive.

    Reply

Comment