Paris WebRTC Conference: Thoughts and Impressions

Consider this my trip report for Upperside's WebRTC Conference 2013 in Paris, December.

This week I attended the WebRTC Conference in Paris by Upperside. There were over 100 people in this event, with the room packed full throughout the days of the conference.

Generally speaking, this is what I noticed in this event which was rather new:

• Telecom companies sent their labs to talk. The people who are allowed to play with technologies. It is a progress from last year's standardization people. Would be nice to see next those that launch services or drive projects and policies
• Telecom provider's vendors were less progressive than their customers. It is as if they don't even want them to go there. While the vendors talked about IMS and ubiquity, the telcos were talking about the post-IMS world. I wish they'd do more than they talk, but that just might be me
• Monitoring and deployment related issues were starting to creep into the show. This is a good thing as it is an indication that there's some real stuff going on and we're moving from evangelizing into implementing

In this event I had a long session in the preconference about the business aspects of WebRTC and another one about WebRTC Islands. You can find my own slides at the end of this post.

If you are into grueling details, then feel free to read the rest. This goes through the sessions of the event, sorted by telecom, their vendors and the rest...

WebRTC Award

The award went to MashMe.TV.

If you are not aware of this service, it is essentially a consumer/SMB type video conferencing service fit for the web. that said, the best way would be to explain it via a video:

http://www.youtube.com/watch?v=lY7WKWOp27s

Overall WebRTC Market Status & Forecast – Dean Bubley / Disruptive Analysis

The first day of the event started by Dean Bubley. He gave a high level view of what makes WebRTC different and with a good chance to succeed.

It boils down to two main reasons:

• Telecom is obsessed about their networks and less about their customers
• People call for various reasons, and it is time to develop as well as monetize these conversations based on the purpose of the calls instead of treating them in the same way

He stated he is failing to find reasons for WebRTC to fail, which says a lot – coming from someone like Dean.

WebRTC: The Telco Approach – Stephane Cazeaux / Orange Labs

Stephane started by saying that services are limited by what their protocol can do, and the only way to differentiate is in the user interface.

He sees a trend of communication migrating from an application and a protocol into the web – something that changes the basis of differentiation. It happened in email and instant messaging, and now that we have WebRTC it is happening with VoIP.

On the telecom side, he sees a migration happening from interconnection to multi-actor – a realm where platform, service and identity are provided by different entities.

All this coming from someone in the Telecom space was refreshing. The best part for me were the two slides where he detailed where SIP and IMS architectures are challenged today – and these challenges are actually easier to solve with WebRTC.

Telecom Italia WebRTC Approach – Fabrizio Caffaratti / Telecom Italia

Fabrizio started off with dreams and nightmares from WebRTC for telco vendors. It was a candid approach to what lays ahead. He questioned the current approach to Telco WebRTC – the one of having WebRTC as an access point into RCS and IMS, but in my view, it didn't go far enough with what can be done with it.

His main suggestion, which I totally agree with is the need to play with it now – don't wait until it gets standardized, approved or gains market share. You need to be ahead of the curve and the only way to do that is by trying it out.

WebRTC Services and Open Issues like Inter-domain Interoperability and Identity Management – Paulo Chainho / Portugal Telecom

As with the other Telecom presentations, Paulo also started in the theme of "leaving the chains of IMS for the agility of WebRTC".

Portugal Telecom are playing with a pure WebRTC solution, named WONDER. For that, they have made the following decisions:

• Triangle and not the Trapezoid architecture (more on that in my presentation below)
• Proprietary signaling instead of SIP or IMS
• Trying to crack the identity piece in WebRTC

The highlight was the use of "post-IMS" to describe our era of WebRTC.

Prototyping WebRTC – Jimmy Ehrbar / COLT

Colt is a pan European service provider, focusing on the B2B market (no consumers) and wireline.

They started looking at WebRTC a year ago, and Jimmy took the audience through the journey Colt has gone through.

Essentially, they didn't use their internal standard processes for development (saw them as too slow for the task), and instead went for prototyping. They selected their use cases. In general, in the areas of:

• Voice & unified communications
• Contact Centers
• SIP gateways

The advantages they have found out with WebRTC where: no client, no dial-back, context capture capabilities and no call charges. The disadvantages were: limited mobile support, PC hardware requirements and security concerns.

In the end, the decision they made was to not move forward because of their business need to support all browsers.

I think this is the way to go – play, prototype and then decide what to do (or not do) with WebRTC.

Building a One-Number Solution using WebRTC – Bodil Josefsson / Ericsson

Bodil started by explaining what One Number is (essentially being able to use your carrier's phone number of a laptop or any other IP-enabled device). She pointed Ericsson's view of One Number being the aggregation point of all communication related features – be it messaging, shifting between devices, call history or others.

Being Ericsson, she did the obligatory RCS-e/Joyn spiel of innovation and service creation over IMS and how WebRTC needs to extend beyond the web using interconnect into IMS.

How to Monetize an IMS Network with WebRTC – Bob Ding / Huawei

Huawei were funny.

They started with an interesting spiel about WebRTC bringing openness into IMS and how that is mandatory, but then I have to question how can that fit into the rest of the presentation where IMS was viewed as the center of the world.

The fact that I still don't have the presentation anywhere in digital format says volumes about how much openness plays a role here beyond lip service.

The interesting tidbit from the presentation was the fact that Bob did state that the best opportunity for WebRTC for telecom operators is most probably in verticals.

Building an Ecosystem to Support WebRTC Growth – Chris King / Oracle

Not much to say here.

The presentation started with the premise of WebRTC missing critical components. It had the classic FUD about ubiquity. This is something I've seen in the past. The same arguments said against VoIP were now used against WebRTC.

I was left unconvinced. The web is well prepared for handling scale and ubiquity. Time to get over it and join the crowd instead of standing in the sidelines pompously.

What Chris wanted to say was just "buy Oracle's SBC". In the end, the first "question" to Chris was a complaint about this being a long commercial for Oracle. Couldn't agree more. Vendors need to be able to talk about subject matter within their product's domain without pushing their product with such a hard sell.

Jitsi Videobridge and WebRTC – Emil Ivov / Jitsi

For those who don't know, Jitsi is an open source software communication client – of the kind threatened by WebRTC.

Emil gave an overview of how multi party video is achieved – with MCUs or with video relay. He also demoed the new Jitsi video bridge, which is based on the paradigm of Google Hangouts. The best part – it is open sourced.

Call Center Applications – Miguel Ponce De Leon / Waterford Institute of Technology

Miguel had a very interesting session about the challenges of WebRTC in the contact center (things done using  openRMC) – one that went from theory into actual details.

He started from what you want with WebRTC on a webpage and what issues you may have technically in achieving it - things like moving across pages during a call session.

The interesting part for me was how you can add the web notion of doing A/B testing to decide how to serve the WebRTC calls on the web – something unheard of in VoIP.

He raised some real deployment challenges:

• The need for a new backend portal for the agent
• Training for agents on such a system
• IT needing to switch browsers internal (usually IE to Chrome)
• The fact that now you have too much data coming from the browser and there needs to be some thought put into how and what to show to an agent – this is the opposite problem to having no data, which is what happens today

Can WebRTC Accelerate the E-Commerce Market Segment? – Chris Koehncke / GENBAND

Chris receives the most entertaining session award for this show from me. With his usual Kranky style, he made the audience laugh and… wake up.

He built up a story of the challenges of shopping and ecommerce for the SMB. It started from trying to shop for an engine for his vacuum cleaner – a true story.

The interesting point he made is the need for a reversal in view – it isn't about connecting the web to the telephone but vice versa – how you can connect the phone to the web (hint: it isn't about IMS).

Performance Monitoring of Media Flows in WebRTC – Varun Singh / Aalto University

Varuun gave the pitch about the need for monitoring the media flows in WebRTC. He did it from the angle of his newly found company – callstats.io.

The basis of this all is the fact that WebRTC brings with it complexities in how the networks are laid out, how the architecture of the media is being designed and operated and lastly – how it changes dynamically through time.

He listed the options vendors have in monitoring their WebRTC services and also provided some good insights on the need (and ability) to have a strategy in place to setting the initial constraints on the media of the session (as opposed to statically decide to go to the maximum possible bandwidth and resolution).

Challenges and Opportunities in WebRTC QoS – Justin Hart / Sonus

Justin Hart discussed the shift we see now from engineered VoIP platforms to general purpose software ones. You can view this as part of the general trend of virtualization and cloud computing.

Up until today, most VoIP platform were put in a managed "box" – the vendor built the hardware, knew what's in there, how it worked, etc. Now, vendors are expected to provide a CD – or more likely an image for VMWare or an AWS AMI file – one that will immediately work in a virtualized environment of a cloud deployment. This isn't that easy to achieve (but possible) due to the real time nature of VoIP.

He also stated the need to know the details of VoIP and how it works in order to be able to build a globally workable architecture.

Coming from this field, it is easy for me to agree.

Quality over WebRTC – Amir Zmora / Audio Codes

Amir provided in his session split the quality issues into two components: tools and network.

In the area of tools, there were some good suggestions on where to look and what to do in order to improve quality. In the area of network, he made a point of trying to use the VoIP backbone of service providers instead of the "open internet" – something "like IMS". While I can see the value, I don't think carriers offer such a thing globally in a way that makes it easy to achieve.

He also pitched yet-another-box which he named "Session Experience Management Solution" which is essentially an SBC with smarts as to the context of the call and the state of the network to decide if calls should be routed over the open internet or that "like IMS".

WebRTC Best Practices – Chad Hart / Dialogic

Chad gave a view of the possible options for building a gateway for WebRTC. He split it into 3 parts:

1. H2S – HTTP to SIP – the signaling side of the story
2. Media – where the transport of the media takes place – think SRTP to RTP and the other nuances of WebRTC/SIP interoperability issues
3. Transcoding – dealing with matching the codecs, and I'd also say that any media processing such as scaling and layout goes here as well

He showed the different packaging options to these 3 components – from having them all bunched together into a "gateway" to having each one as a distinct component.

The end of his session had an interesting case study of an airline's airport kiosk used to reduce the number of agents at the airport in the slow hours of the day.

Communications 3.0: A Feature not a Destination – Kavan Seggie / AddLive

Kavan Seggie made a pitch for the long tail of communications. He called it in-context feature versus a destination. Where he sees that? In various verticals: Enterprise communications, telehealth, education, gaming/social, contact center.

He went through some of the use cases of the customers of AddLive, who are using the AddLive API platform to build solutions - something I want to see more at conferences.

xWiki: A Prototype Implementation – Ludovic Dubost / XWiki

XWiki is a kind of a Wikipedia platform. It is open source, and built by 35 people in France and Romania.

Ludovic explained how they added WebRTC and the challenges they had – mainly the fact that they don't have or want to use SPA (Single Page Applications) and the fact that they use Java – which isn't that easy to integrate with WebRTC (at least there weren't any good solutions available).

They ended up integrating XMPP and having WebRTC use that as signaling.

The end? You can now have collaboration and screen sharing on top of a Wiki platform directly from the page using the same service and website. Neat.

WebRTC Islands – me

My session was about federation and islands and how they make up our future with WebRTC.

[slideshare id=29150181&doc=webrtc-islands-131212102458-phpapp02]

In the end there were interesting questions about the desire to federate between OTT services. The issue there isn't a technical one but rather a business one – most OTT vendors have no incentive to federate to other services.

Quickies

• Victor Pascual Avila of Quobis replaced a speaker on the topic From Standards to Scalable Deployments. He gave a good overview of the standardization around WebRTC – where it is today and where it is headed
• Victor also gave a session about Services over Data Channels which served as a good premier to the technology
• Frank Schultze of Fraunhofer Institute Fokus detailed their FOKUS WebRTC Telco Stack in his session on Enriching Enterprise Services through WebRTC
• Michel L'Hostis of Apizee gave a good introduction to how WebRTC fits well with trends like BYOD and the need for more context in his session WebRTC Opportunities for Enterprise Applications World
• Richard Ejzak fo Alcatel Lucent explained the 3GPP Standardization of IMS for WebRTC. I am unable to understand why we need to push WebRTC into the heart of IMS instead of just using a "simple" gateway for that
• Daniel Pockoc of Debian Project talked about Successful WebRTC Implementation with Open Source Software. To me it was too much open source in nature and less realistic in how people should tackle communications
• Christian Hoene of Symonics gave a Status Update on WebRTC Codecs and IPR Issues. He ended up showing his company's spatial audio solution of WebRTC
• Badri Rajasekar of TokBox talked of The Journey from Standard to Real World Application: Bridging the Technical Divide. He focused on the requirements of services from WebRTC and how the TokBox platform solves them. He did mention noticing a trend of mobile only use cases that their customers are now developing
  
• Luis Lopez Fernndez talked about Kurento, a Media Server Architecture and API. Essentially, a Java based communication framework that uses WebRTC

Panel: WebRTC for the Operators: Opportunity or a new Threat?

We had a lively panel that ended up as a heated discussion about IMS and WebRTC.

My main gripe is the inability of service providers and their vendors to think of WebRTC without limiting it to the world of IMS. There are so many opportunities that can be realized by service providers without the need for IMS, that I find it hard to believe they aren't pursuing them.

Go out. Do something with WebRTC. only later see if you need to connect it to IMS and how.

I don't want to be in the event a year from now, seeing carriers still talking about IMS and having 5 times the amount of use cases already implemented and launched by others. Seems like a wasted opportunity to me.

Closing Panel: Meet the Experts

The closing panel was interesting. It went through the survey questions that were handed to the audience in the first day.

You can see the results here:

[slideshare id=29149683&doc=webrtc2013openingsurveyresults-131212100818-phpapp01]

Would be nice to hear your thoughts about it.

WebRTC Paris Meetup

The first day of the conference was closed by the Paris Meetup group. There were a few intersting demos there - some that I haven't seen before. For that event, I did this special welcom presentation:

[slideshare id=29150303&doc=webrtc-paris-lovestory-131212102914-phpapp01]

-

My preconference session was about the WebRTC ecosystem. It is based out of the data I collect. Here it is:

[slideshare id=29150871&doc=webrtc-workshop-ecosystem-131212104905-phpapp02]

-

All in all it had been a great event. Would be interesting to see how this would turn into next year.