My thoughts from WebRTC Conference & Expo III
It is going to be a long one, so bear with me, skip to what interests you or just move on.
This has been quite a full and hectic week for me at the WebRTC Conference & Expo event in Santa Clara. So many sessions to attend. So many people to meet. If you didn’t get my full attention – I am deeply sorry – it isn’t because of you, but rather because of me. Blame it on me trying to multi-task while being severely jetlagged. Feel free to contact me anytime – I promise to try and make a better job of assisting you.
The event was a big one and well organized. I even had a chance to get in shape since the exhibits hall was on the other side of the convention center. For me, this was a chance to meet old and new friends – even see some virtual ones in person.
I heard some people saying that this was a repeat of the previous event in Atlanta, and while I somewhat agree, there were some very different things in this one:
- There were different vendors. A bit more than last time, but also a different audience set. While most weren’t customers, you did see some other folks walking the show floor. For me that was refreshing, as I got to meet new WebRTC vendors
- Mobile is ready. Last time, most had demos, prototypes or plans to go mobile. In this show, many of the vendors came with mobile apps ready for deployment
The disrupted went public. There were two types of vendors that actively participated as speakers and exhibitors that you didn’t see last time. I’ll place them in these comic categories:
- The vultures – those who have nothing in WebRTC, and probably never will – but have complementary solution that work well with WebRTC
- The disrupted – those that have nothing in WebRTC, but are going to be disrupted by it – astonishingly, they came trying to “sell” their products at the show
Those that weren’t there are the web guys. The ones we are all putting our hopes in with taking this technology and shaking the world up.
There were a few major announcements during the event. While I can’t and won’t cover them all, here are a few I found interesting:
- TokBox now supports recording and archiving. This is… awesome. It was one of the major things I heard people requesting and asking abou – there’s now a solution out there. They had more announcements, but this is the big one
- Temasys made a splash with several customer
wins, potentially getting their platform to support 50 million users during 2014. A clear indication for me to follow them closer
Things I learned
Here are a few tidbits I learned this week:
- AddLive is a 7 people outfit. And they are now providing a service for 5,000 businesses and 250,000 users. That’s large numbers in such a small operation of this kind
- There are vendors that do nothing with WebRTC, but offer a lot of capabilities that work well with WebRTC. More about them on a separate future post
- Weemo added support for iOS, Android, IE, Safari and PhoneGap. This got covered by Tech Crunch, so made some waves over there. For me the story is the fact that vendors are finding ways to support all those unsupported platforms, and Weemo is just the latest one to do so. What interested more was their integration with Salesforce, which should be a big red, flashing light to UC vendors – I’ll let them figure out why
- Platronics used their demo sessions to show how they are now packing more and more sensors into their headsets and making the data of these sensors available to developers. While not directly related to WebRTC, that information can be used programmatically in the browser alongside WebRTC or sent remotely via the Data Channel.
- If you plan to do a demo at the next event, make sure to prepare. Make it sleek. Less text on slides. Less words explaining what you and your company do. Get fast to the actual demo and then maybe explain how it is related to you. Make sure to bring something new from the last demo you show. Make it interesting. Make it entertaining. It was hard to listen and score 15-20 companies an evening – so be different and make me enjoy my time while there
- There were more women at this event than the previous ones; though still far from enough
- Vendors didn’t invest in giveaways. I went back home without any new T-shirts, which is a shame. I did get a toy car for my boy from Vidyo, which was cool
- When demoing connectivity to SIP, everyone did it with Counterpath Bria. Somehow, this is the darling of the desktop, which makes me doubt their future
- Ian Small, CEO of TokBox, is probably the best presenter at WebRTC conferences. He was born for that…
Here are some of my takeaways from some of the sessions I attended:
Business Intro to WebRTC
- Phil Edholm introduced the topic and covered some overreaching aspect. For me, the thing that hit home the most, was the way WebRTC is usually depicted as either a triangle or a trapezoid. While I have seen it many times, Phil made the distinction that WebRTC is all about the triangle, which is new to VoIP people who love looking at the trapezoid architecture for their networks. This is world changing
- Lawrence Byrd gave use cases for WebRTC, with a few that I now need to add to my vendors list. He made a point about SoCoMo (Social/Cloud/Mobile), a topic I’ve been meaning to write about but never got a chance. From him, I took the way WebRTC disrupts the enterprise vendors – it isn’t directly – it is just by being another tool at the hands of the SoCoMo hordes (=web vendors)
- Chris Vitek discussed contact centers and interaction experience. He explained well how building an Amazon Mayday-like experience can be achieved using WebRTC in existing contact centers
- Brent Kelly gave an analyst’s view. He made a point about the video codec indecision being an issue for enterprise adoption, suggesting enterprises should embrace voice and data channel capabilities of WebRTC while waiting on video
- John Burke connected WebRTC with Big Data and security. For me, the Big Data part hit home: vendors are looking for more data from more sources, trying to gain insights from them. By using WebRTC, they get access to more data easily – data that is relevant in order to optimize and personalize their services. This is a way of re-centralizing the channels of communication from customers back to the enterprise
Here are my slides from this session:
More details on this can be found in my WebRTC for Business People paper.
Google’s keynote was interesting. Justin Uberti and Serge Lachapelle took the stage and were sought after the keynote by many in the audience.
Interesting tidbits from this session included:
- Justin Uberti indicating that Opera is now supporting WebRTC again, after their migration to the Chromium codebase
- An estimation of 1.2 billion enabled browsers for WebRTC
Chrome’s roadmap when it comes to WebRTC includes:
- Official and stable support of screen sharing with the intent of enabling also the ability to select a specific app to share
- Video alpha channel support, enabling having parts of the video image transparent – and getting video out of its customary rectangle
- Better integration into mobile – supporting hardware acceleration, optimizing APIs in Android, etc.
- Native apps for mobile, making it easier to integrate into apps instead of HTML5
- Later on, VP9 and SVC support in WebRTC
- Justin placed the next challenge for WebRTC in fuzing Web, Mobile and the Server side of things – making sure they all have a common WebRTC API to use to make it easier to deal with. This is a considerable challenge and a very interesting one. For Google, this is “the next chapter of WebRTC”
- Gave an interesting demo of how WebRTC is used by a small startup called SoundTrap – check it out for yourself
The slides are available online.
I moderated the signaling workshop. Here’s what I gleaned from the sessions:
- Peter Dunkley from Crocodile RCS indicated places where SIP makes a good signaling solution. The main issue when using SIP in the examples he gave was matching the media profiles that are different between the VoIP/SIP world and WebRTC. His examples were based on the Kamailio SIP Server
- Kevin Wiseman from CafeX Communications asked the crowd where they come from. Unsurprisingly (and sadly to some extent), there were more telecoms and VoIP people than web people in the room. Kevin mentioned that signaling is easy, but the hard part is dealing with the media descriptors (that’s SDP). He went on to show different types of architectures of where to place interoperability between WebRTC and SIP
- Rod Apeldorn from Priologic gave what to me seems like the best suggestion: Where you don’t need SIP – don’t use it. He went on to list the types of messages you’d need to develop for proprietary signaling and gave good reasons why signaling and application are better off together than decoupled. He ended up by mentioning a possible future where WebRTC can be used to remove the control from the server side by way of the data channel
All in all, it was a good and informative workshop. What I missed in this session was a clear indication of what should be the compelling event for someone to choose to use SIP inside the web browser at the edge instead of taking care of that interconnection on the server side.
Gaming and Beyond
I moderate the Gaming and Beyond panel discussion, which was interesting and engaging.
The thing that struck me the most was how the panelists thought about the suitability of WebRTC to M2M and the internet of things. My gut feeling always indicated that there is a connection, but I couldn’t put my finger on what that really means.
While I don’t think we got to a good enough answer during the panel on this topic, I did find like-minded panelists who thought that WebRTC, and web technologies in devices that have no display or a browser is something that makes sense – and the reason for that, is simply because of the huge ecosystem of tools and developers available for the web stack of technologies.
Communications Enabling Business Processes
This track focused on how WebRTC can be embedded in existing flows of businesses – making voice/video communications into a feature.
The interesting comments I was about to capture from this session:
- Oded Gal from Blue Jeans stated that when they added WebRTC to the mix of their hosted video conferencing service, they’ve noticed a change in the usage pattern: 50% of the users now come through the browser using WebRTC, and only 5% now come from Skype. This shows how WebRTC is adopted naturally by the end users
- Ben Weekes from Requestec had an interesting use case story of a healthcare customer that uses WebRTC embedded into processes that enable FDA regulation compliancy. There it was used for training patients on the use of a medical tool remotely instead of physically going to a training site
- Rich Cerami from CenturyLink stated that from a contact center perspective, using WebRTC voice calling into the contact center doesn’t change much for them in terms of the operational aspects of it, while other mediums of asynchronous interactions (=chat, email) do change things a lot
Large Scale WebRTC
There was an interesting track about large scale WebRTC deployment. It was interesting, because most vendors on it missed the whole point…
For me, scaling WebRTC needs to be done in web-scale. Anything below 10 million or even 100 million isn’t going to be any interesting a year from now; so when vendors on the panel talk about 100K users as a large deployment, that’s a #fail in my book. Especially when they try to put FUD against open source tools.
Arnaud Budkiewicz from Bistri, on the other hand, had many interesting points to share with the crowd:
- Bistri does toy 1 million video minutes a month. I guess that’s more than most (or all?) enterprise video conferencing deployments out there. And the team is smaller than all video conferencing vendors
- Bistri stores 200M contacts in its platform
- Bistri uses fallback to Flash, and their distribution is 35% WebRTC calls and the rest is Flash. Who said Chrome dominates the market?
Arnaud talked about scalability in other aspects – of how today you can get to this scale. The way he sees it:
- You need to connect to existing social networks instead of creating new ones
- With WebRTC, you can offer guest mode, where people not registered to your service can attend sessions with people who are registered to it
- To enhance reachability, you need to think of mobile apps, and even desktop ones
- Cross platform support is important
- Bistri plans on using geo location for things like TURN clustering and database sharding – things you must do in large scale, global deployments
- They run on AWS, which offers them automation and elasticity – things you don’t see much yet in the traditional VoIP world
Another miss in this panel was the question of how do you federate between these networks – something I feel didn’t get a good answer. I’ll try to answer it in another post next week – this one will be long enough as it is.
Service Providers Track
The service providers track was the most disappointing for me. Most vendors presented a very limited world view in my opinion – one that focused solely on how to connect WebRTC back into the IMS core network. To me, IMS is a failure when it comes to innovation and launching new services. The most ridiculous part, was phrases thrown into the air like “when you connect WebRTC to IMS, it can enjoy all the great benefits that IMS provides which WebRTC lacks” – show me one such benefit please.
This focus on putting the WebRTC sticker on IMS equipment isn’t going to help telcos realize the value they can really get out of WebRTC. Different approaches are required, with different thought processes – and I have seen very little of that during this track.
The two interesting notes from this track are:
- Galeal Zino from Tata Communications stated in a panel that carriers need to use this opportunity of embracing WebRTC and using it to open up their network for external innovation
- Rich Cerami from CenturyLink quite surprised me when he mentioned in another panel that they are working with WebRTC already, without any vendor and using it to roll out services – not specifically WebRTC ones, but just as part of a larger service – just the way WebRTC needs to be used
Chrome Dev Summit
Here’s the thing… during the event, I took a “morning off” and headed to Google’s Chrome Dev Summit to hear the latest and greatest of what Google is planning for us. Here are some of my main takeaways from the couple of sessions I was able to attend:
Chrome and Google were all about closing the gap between the OS and the web. Well… now it is changing to close the gap between native (=mobile) and the web. The real distinction here comes from the things that needs to get done now:
- Rock solid 60fps UI rendering
- Caching application code and resources in advance to run online or offline
- Access to device peripherals and internals
- Enable the use of a larger toolset
The Chrome team is working to get the above, and generally speaking they are looking into these broad groups of challenges:
- Improving rendering performance
- Giving hardware and OS access – something where the Data Channel in WebRTC can be later used to transfer that information
- Adding WebGL, WebAudio and WebRTC
- Web notification – something that can be used to get incoming calls into WebRTC within the browser properly
- Now that there’s a Chrome webview inside Android KitKat version, next in line is having that updated at the speed of Chromium (i.e every other week) instead of at the speed of Android (=twice a year). This is a huge thing
- A lot of work is done for offline HTML5 apps. The way it was put, which I liked very much, was that developers of HTML5 apps need to think “offline first” for their online experience and that provides a performance improvement
- Last, but not least: “if the thing has a screen, then it most probably has a web browser these days” – and for me that means WebRTC and interactions everywhere
All in all, it has been a great event. One more to go this year…