I've been thinking a lot about WebRTC recently. How and where it will become important, and what it might do to our concepts of voice/video communications and the existing telecom value chain.
It's still very early days, but the momentum and details suggest that it will be of incredibly high importance. There are certainly complexities - not least of which is Apple not yet revealing its intentions - but overall the general premise "feels" right. There are no obvious irreversible "gotchas", and there are plenty of interesting use-cases, and a whole plethora of innovators from both small and large companies alike.
See more recent posts on WebRTC here and here , and watch out for the forthcoming Disruptive Analysis research report here.
This is diametrically opposite to things like NFC payments or RCS, for which there are plenty of hard, easily-described and unfixable flaws in the basic concept, and where support and innovation are thin.
WebRTC fits well with the idea that much of what we consider as communications "services" are in fact just "applications", and increasingly drifting further down to become "features" and eventually "functions". Messaging is already a long way down that curve - IM chat inside apps such as Facebook or Yammer or Bloomberg are not "services", any more than the bold type button is a service. They just send words from A to B, rather than highlight them on the page.
WebRTC extends that metaphor to spoken words or visual images. They will just be sent via a browser or web widget (obviously needing access to camera, microphone, codecs & acoustic processing). It is already possible to have direct browser-to-browser conversations without plug-in or downloaded applications on the desktop. Massmarket versions of Chrome, Firefox and IE are all likely to support WebRTC during 2013, with a steady move onto mobile over the next couple of years.
This will mean that voice communications (and in some cases video, although I think that will be minor) will become much more pervasive, cropping up in all sorts of interesting contexts. I've long talked about "non-telephony" forms of voice, such as Siri, in-game voice chat, push-to-talk, business-process integrated voice and so on. WebRTC is likely to be the single biggest catalyst enabling "voice as a feature" to be used by web developers in the same fashion as any other aspect of HTML.
Maybe in two years time, you'll be on the Amazon website and you'll suddenly hear a voice saying "Hey, congratulations to all of you browsing right now - there's a 10% discount on everything for the next 5 minutes!". It could be the web equivalent of a tannoy in a supermarket "Special on Aisle 3!". That's not a phone call. It's not a service, either. But it is voice communications. Other possibilities are too numerous to mention, but many have observed that this means that "the website becomes the call centre". Not "click to call" or even "Skype me", but just having an in-browser real-time voice interaction in the same fashion we already see with IM chat. Adding WebRTC voice to LinkedIn, Facebook and numerous other sites is obvious, and so are things like web-karaoke without plugins, or voiceprint-based authentication instead of passwords.
This is disruptive to both traditional phone calls, and also to "legacy" standalone VoIP clients such as Skype's. It is doubly disruptive to new VoIP platforms such as telcos' IMS-based VoLTE, which is mostly just a recreation of the old telephone mindset, and is having enough problems even doing that.
At the core of this is a central problem for the telecoms industry. It is addicted - perhaps even enslaved by - the idea of the "subscription". All operators report subscriber numbers, the word SIM means Subscriber Identification Module, and much of the technology elements such as HSS's and most billing systems assume subscription-type relationships. Regulation is also heavily subscriber-centric.
Now subscriptions are a very valuable business model. Ongoing payments are attractive for companies, and predictable for users. Many businesses - include a lot of technology analyst firms - are heavily dependent on subscription revenue streams.
But they're certainly not the only business model, and neither are they without flaws. They mandate an ongoing customer relationship. They assume that the capability being provided is an identifiable and separable service.
While that has been fine for the past 100 years of telephony, it is clear that the landscape is changing. Voice or video communications is going to appear in lots of contexts - service, application, feature, function. Sometimes it will be based on the need for enduring relationships and "reachability", for example with a phone number and subscription. Sometimes it will be transient and in-app.
Some communications capabilities will continue with ongoing identities and billing relationships. Others will be sponsored, free, ad-hoc, one-offs, occasional use, ambient, ad-supported and so on. I'll get my spoken words delivered - and paid for - in as many ways as I get my italic words. Sometimes I'll get italics in my subscribed and paid-for magazines. Sometimes they'll be on a website or billboard for free.
If I want to speak to an Ikea customer service agent with a query on how to put my cupboard together, I don't need their number, and they won't need mind. I'll just click the "help!" button in the Ikea app which has already tried to show me where I'm going wrong, perhaps with a one-off fee associated with it.
Now it's possible that could be done with Telco APIs, hooking into an IMS core and telephony app server. But I might be using a WiFi-only tablet with no associated phone number or operator relationship. And Ikea, in this example, is not going to want to deal with either 100 telcos or the constraints of some collaboration like OneAPI, when it could just add the function simply and easily into the browser or app, at no cost or hassle.
My view is that WebRTC will ultimately be the "ubiquitous" voice and video communications service. There will be more browsers, and voice-embedded websites and apps than mobile and fixed phones. The telco/IMS world will be a subset of this, constrained by the narrow formula of subscription-style relationships and defined identities.
Yes, there will be security issues around the perceived dangers of anonymised voice communications. Yes, in some cases network quality will be too poor to support good-enough voice using best-efforts connections. But those will be (fixable) corner cases, and not things to derail the wider trend.
We already see service providers looking at opportunities around WebRTC - addressing services, legacy interoperability, premium billing, perhaps quality enhancement or emergency-calling-as-a-service. AT&T, China Mobile, Telefonica and others have spoken publicly about WebRTC, and I know many more that are watching or involved in the standards work. Vendors like Ericsson are looking too - this is not just a Google / Microsoft / Apple (??) fight, with traditional telecoms getting squashed in the middle.
There are still plenty of questions, and this won't all happen overnight. But one thing is, to my mind, utterly inevitable. Those companies who refuse to see beyond the "subscription" - and those technologies which cannot flex enough for non-subscription relationships - are facing decline into niches or outright irrelevance.
(Footnote: WiFi doesn't need a subscription either. LTE does)
(Footnote #2: One good way for Telcos to get around the legacy subscription mindset & infrastructure base is to pursue Telco-OTT services and business models. Buy the report!)
It's still very early days, but the momentum and details suggest that it will be of incredibly high importance. There are certainly complexities - not least of which is Apple not yet revealing its intentions - but overall the general premise "feels" right. There are no obvious irreversible "gotchas", and there are plenty of interesting use-cases, and a whole plethora of innovators from both small and large companies alike.
See more recent posts on WebRTC here and here , and watch out for the forthcoming Disruptive Analysis research report here.
This is diametrically opposite to things like NFC payments or RCS, for which there are plenty of hard, easily-described and unfixable flaws in the basic concept, and where support and innovation are thin.
WebRTC fits well with the idea that much of what we consider as communications "services" are in fact just "applications", and increasingly drifting further down to become "features" and eventually "functions". Messaging is already a long way down that curve - IM chat inside apps such as Facebook or Yammer or Bloomberg are not "services", any more than the bold type button is a service. They just send words from A to B, rather than highlight them on the page.
WebRTC extends that metaphor to spoken words or visual images. They will just be sent via a browser or web widget (obviously needing access to camera, microphone, codecs & acoustic processing). It is already possible to have direct browser-to-browser conversations without plug-in or downloaded applications on the desktop. Massmarket versions of Chrome, Firefox and IE are all likely to support WebRTC during 2013, with a steady move onto mobile over the next couple of years.
This will mean that voice communications (and in some cases video, although I think that will be minor) will become much more pervasive, cropping up in all sorts of interesting contexts. I've long talked about "non-telephony" forms of voice, such as Siri, in-game voice chat, push-to-talk, business-process integrated voice and so on. WebRTC is likely to be the single biggest catalyst enabling "voice as a feature" to be used by web developers in the same fashion as any other aspect of HTML.
Maybe in two years time, you'll be on the Amazon website and you'll suddenly hear a voice saying "Hey, congratulations to all of you browsing right now - there's a 10% discount on everything for the next 5 minutes!". It could be the web equivalent of a tannoy in a supermarket "Special on Aisle 3!". That's not a phone call. It's not a service, either. But it is voice communications. Other possibilities are too numerous to mention, but many have observed that this means that "the website becomes the call centre". Not "click to call" or even "Skype me", but just having an in-browser real-time voice interaction in the same fashion we already see with IM chat. Adding WebRTC voice to LinkedIn, Facebook and numerous other sites is obvious, and so are things like web-karaoke without plugins, or voiceprint-based authentication instead of passwords.
This is disruptive to both traditional phone calls, and also to "legacy" standalone VoIP clients such as Skype's. It is doubly disruptive to new VoIP platforms such as telcos' IMS-based VoLTE, which is mostly just a recreation of the old telephone mindset, and is having enough problems even doing that.
At the core of this is a central problem for the telecoms industry. It is addicted - perhaps even enslaved by - the idea of the "subscription". All operators report subscriber numbers, the word SIM means Subscriber Identification Module, and much of the technology elements such as HSS's and most billing systems assume subscription-type relationships. Regulation is also heavily subscriber-centric.
Now subscriptions are a very valuable business model. Ongoing payments are attractive for companies, and predictable for users. Many businesses - include a lot of technology analyst firms - are heavily dependent on subscription revenue streams.
But they're certainly not the only business model, and neither are they without flaws. They mandate an ongoing customer relationship. They assume that the capability being provided is an identifiable and separable service.
While that has been fine for the past 100 years of telephony, it is clear that the landscape is changing. Voice or video communications is going to appear in lots of contexts - service, application, feature, function. Sometimes it will be based on the need for enduring relationships and "reachability", for example with a phone number and subscription. Sometimes it will be transient and in-app.
Some communications capabilities will continue with ongoing identities and billing relationships. Others will be sponsored, free, ad-hoc, one-offs, occasional use, ambient, ad-supported and so on. I'll get my spoken words delivered - and paid for - in as many ways as I get my italic words. Sometimes I'll get italics in my subscribed and paid-for magazines. Sometimes they'll be on a website or billboard for free.
If I want to speak to an Ikea customer service agent with a query on how to put my cupboard together, I don't need their number, and they won't need mind. I'll just click the "help!" button in the Ikea app which has already tried to show me where I'm going wrong, perhaps with a one-off fee associated with it.
Now it's possible that could be done with Telco APIs, hooking into an IMS core and telephony app server. But I might be using a WiFi-only tablet with no associated phone number or operator relationship. And Ikea, in this example, is not going to want to deal with either 100 telcos or the constraints of some collaboration like OneAPI, when it could just add the function simply and easily into the browser or app, at no cost or hassle.
My view is that WebRTC will ultimately be the "ubiquitous" voice and video communications service. There will be more browsers, and voice-embedded websites and apps than mobile and fixed phones. The telco/IMS world will be a subset of this, constrained by the narrow formula of subscription-style relationships and defined identities.
Yes, there will be security issues around the perceived dangers of anonymised voice communications. Yes, in some cases network quality will be too poor to support good-enough voice using best-efforts connections. But those will be (fixable) corner cases, and not things to derail the wider trend.
We already see service providers looking at opportunities around WebRTC - addressing services, legacy interoperability, premium billing, perhaps quality enhancement or emergency-calling-as-a-service. AT&T, China Mobile, Telefonica and others have spoken publicly about WebRTC, and I know many more that are watching or involved in the standards work. Vendors like Ericsson are looking too - this is not just a Google / Microsoft / Apple (??) fight, with traditional telecoms getting squashed in the middle.
There are still plenty of questions, and this won't all happen overnight. But one thing is, to my mind, utterly inevitable. Those companies who refuse to see beyond the "subscription" - and those technologies which cannot flex enough for non-subscription relationships - are facing decline into niches or outright irrelevance.
(Footnote: WiFi doesn't need a subscription either. LTE does)
(Footnote #2: One good way for Telcos to get around the legacy subscription mindset & infrastructure base is to pursue Telco-OTT services and business models. Buy the report!)
Funny thing about Apple is that, since all flavours of Safari are WebKit based, it is basically getting the work done by Google. Not entirely all maybe, but enough to put them in much a better position than Microsoft/Skype -- that is indeed fighting back with all its heavy weights in the IETF and W3C with the clear intent to delay the specification process.
ReplyDeleteOther interesting fact is the composition of the developers team that is working on WebRTC at Mozilla. Guess who's the big guy helping the small (?) opensource company ;-) https://wiki.mozilla.org/Platform/Features/WebRTC
Interesting comment about WebKit.
ReplyDeleteBut I would have thought that Apple is more interested in getting H264 involved somehow?
Microsoft - is it really about delay, or getting more favourable specs (eg again on codecs etc) and squeezing the timing a little bit to prep WebRTC-flavour Skype, while Apple is still procrastinating anyway?
I'm not intimately involved in the politics of all of this, but my general read (and expectation) is about a bit of a 3-way tactical tussle between Google, Apple & Microsoft, with the traditional telecoms world being caught in the crossfire between them.
Microsoft is most definitely about delay. Common understanding is that, implementation-wise, they are at least one year behind Google and Mozilla. Among many, worst implication of current specs becoming standard soon consists in Windows Phone being the only mobile platform lacking WebRTC support at the time of the big explosion. Android of course will have it, IOS too, either natively or through Firefox/Chrome apps.
DeleteCodecs are also an issue, but for the browser guys a marginal one. And in any case orthogonal to the actual solution. Current consensus is to have g711 and Opus mandatory for audio, no mandatory codec for video. For as much weird as it seems, that is not going to change, unless something really disruptive happens (like MPEG-LA licencing h264 as royalty free for web use, this time forever).
Great way for Microsoft to further marginalize their browser and mobile OS. Consumers fled IE for better speed and reliability; they'll do the same when Chrome and Firefox and Opera and Safari and Mobile Safari and Android support WebRTC, and IE lags again. Oh, people might keep it around, the same way as they'll keep Skype installed and running alongside WLM, or GTalk. But it'll be a purpose-fit use-case, rather than a general one.
DeleteHaven't been following WebRTC closely, so this is very helpful. From a Public Wi-Fi perspective, my comments are:
ReplyDelete-I absolutely agree that service/app/feature should be abstracted from the access network. WebRTC will make Voice-over-WiFi even more viable than it is today with VoIP-based implementations. I'd be curious to understand better whether this enables a simpler call handoff between various access networks.
-When you talk about relationships and "reachability", I think of identity and directory services. From an architectural standpoint it makes now sense whatsoever for those services to reside at the access layer, they are core services. I've been of the opinion for some time that eventually Apple and Google (possibly Amazon & Facebook) will provide those services (think AppleID and gmail account). In return, they'll get all manner of additional information about your network usage, visited venues, and even CDR-type info with WebRTC. You'll still have to subscribe or "paygo" for licensed wireless data, but it'll be tied to your Apple or Google account. {This is also why I suspect Apple and Google already have plans to use these accounts as Passpoint credentials when it begins to be widely deployed)
should read: "it makes no sense whatsoever"
DeleteBefore we killed competition over here we had a very healthy and vibrant centralized procurement model developing for a significant portion of communications. That is, if the carrier prices communication sessions/bandwidth to reflect marginal cost, the high-volume corporate buyer will subsidize access for the low-volume user. The reason for this is the economic value add of the commercial event/transaction is far greater than the underlying communications event/session.
ReplyDeleteAre any carriers or vendors thinking this way? Because Google may well be developing a hammer (Droid) and anvil (fiber) strategy to do just that over the next decade.
Fascinating perspective, but (a) unless connectivity is going to be free, people are going to need some sort of subscription to pay for it, (b) from the perspective of a web site, I'm not sure there's a difference between a Skype plug-in that makes a voice connection between the browser and the site's call center and another bit of HTML that creates the same, and (c) an awful lot of calls we make are still person-to-person, which means that there needs to be a number, name, or other ID that lets me say who to call and lets calls to me get to me.
ReplyDeleteKrulwich
ReplyDelete- Subscriptions (enduring relationships) are only one form of business relationship for connectivity. You also have one-off payments (eg WiFi hotspot or even just buying a prepay SIM for a day when travelling & then throwing it away), free, venue-sponsored, shared etc. You could say "allow my Facebook friends to use up to 20% of my data via tethering" or 100 other options. Subscriptions are just one approach
- Skype is the outlier here, because it's the only 3rd-party VoIP client with sufficient penetration for that to work, although it's far from ubiquitous especially on mobile. Ultimately, part of the appeal of WebRTC is not needing a browser plugin or separate app. It also allows different user-interaction models, rather than being constrained to Skype's
- c) Sure, that's what the PSTN will live on for. But I think an increasing proportion of "calls" will move from being standalone A-dials-B to something triggered by context or concurrent app, ie when you're inside SAP or Facebook. We probably need to keep some sort of universal identifier - or relatively easy way to federate/interconnect multiple directories - for lowest-common denominator comms. Should definitely be decoupled from your access provision though, as it will need to work on multiple devices, multiple network accesses, run by multiple network operators.
I don't get people waiting for webrtc. Similar software is already available as flash or java applets such as webphone: www.mizu-voip.com/Software/WebPhone.aspx.
ReplyDeleteWebRTC will be a plus, but I don't think that it will change so much.