Speaking Engagements & Private Workshops - Get Dean Bubley to present or chair your event

Looking for a provocative & influential keynote speaker, an experienced moderator/chair, or an effective workshop facilitator?
To discuss Dean Bubley's appearance at a specific event, contact information AT disruptive-analysis DOT com

Monday, October 27, 2014

Microsoft now definitely supporting ORTC / WebRTC

Big news today - Microsoft has officially confirmed that it will be implementing ORTC (Object Realtime Communications) in future versions of IE, and that it will be integrating it with Skype. It's not planning to support WebRTC 1.0 it appears, but that has been pretty clear for a long time. (At least, it's what I've assumed in my own forecasts anyway). ORTC has been labelled as WebRTC 1.1, although it's still not an "official" standard but a "community" enhancement.

Details are still only in outline. The most controversial aspect will be the support of H.264 video codec rather than VP8, although that's hardly surprising given that (a) there has been no agreement on mandatory codecs yet anyway, and (b) VP8 is obviously a very Google-centric technology. Given that Firefox is supporting both, and Apple is definitely an H.264 fan, some will find this annoying but hardly surprising.

My take has long been that a certain level of fragmentation is OK for most WebRTC use-cases and developers. It's a bit of a pain, but it's not the end of the world. Yes, there will be questionmarks about licensing / royalty fees, but if that's picked up by the browser supplier (or third parties like Cisco) that should be tractable. 

I'd expect most of the WebRTC cloud platform providers to support ORTC endpoints - which may mean that the cloud/gateway model of WebRTC gains even more traction vs. pure peer-to-peer use cases. (The cloud approach was growing anyway, as it's the main way to support WebRTC-type voice/video inside iOS and Android apps).

Frankly, the details are important but manageable. We can safely ignore the usual anti-Microsoft brigade's whining about this; Google has worked on the ORTC spec alongside Microsoft and Hookflash, and it is widely (albeit not universally) seen as a good approach, especially as WebRTC 1.0 uses an arcane approach called SDP (a legacy of SIP) to set up connections. 

There are still unanswered questions, which should hopefully get clarified in due course:
  • Will ORTC pop up in IE12 or IE13 first? Will it initially be launched in beta, and under what conditions?
  • Will ORTC support DataChannel properly?
  • What happens with Lync? How does that fit into the ORTC/IE landscape?
  • Might Microsoft support VP8 in a future release, anyway? Or does that depend on how the market evolves?
  • When will ORTC make its way into Windows Phone, and how?
  • Will Microsoft use its Azure cloud platform to offer WebRTC as some form of managed service to developers?
  • What toolkits and support will Microsoft offer to developers for ORTC?
  • Will we see WebRTC/ORTC support in Dynamics CRM, X-Box or other Microsoft properties?
  • Does this limit the scope for 3rd-party IE WebRTC plug-ins (eg Temasys') or does it legitimise them, as a way of getting RTC capability into IE11 and below?

The bottom line is that this is a good day for the "democratisation of voice and video communications" - one way or another, most browsers from mid-2015 onward will be able to support embedded communications directly into websites. Given the various cloud platforms and plug-ins available to smooth the path for developers, this is yet another sign that WebRTC (including ORTC) is proceeding along the right path.

I'm maintaining my view that there will be upwards of 2 billion active individual users of WebRTC by 2019 - a large proportion of the overall Internet and smartphone app user-base. For more details about use-cases, strategic analysis of Microsoft's & Apple's roles, see my full report published recently. Purchasing details are here.

Telecoms service innovation & differentiation - The car model-year approach

Over the next few weeks, I'm speaking at various conferences and events that cover telecom service innovation, including TADSummit and The Great Telco Debate. There will be considerable discussion on whether (and how) telecom operators can compete in the services domain, both amongst themselves and with Internet peers.

One of the things which baffles me about the telecom industry is the constant focus on standardised applications and services. There is no other industry in the world which has hundreds of companies manufacturing identical, commodity products which have zero shipping costs. Where the cost and time of developing, or deploying, those products is measured in years - and where the process involves immense bureaucracy and compromise.

For most telcos, differentiation manifests in network (eg coverage, speed), or billing/charging of standardised components, particularly access to the Internet. Fixed and cable operators often have a greater range of unique options, because they are able to negotiate content deals, and exploit set-top boxes as a delivery mechanism and anchor-point for innovation. 

Mobile operators, however, still mostly compete on minutes/messages/data allowances, perhaps with some bundled-in content, and partner WiFi access. Some plans might "innovate" with pricing for Internet access, using zero-rating. 

But few operators have interesting and unique "own brand" services in their bundles. Telefonica has TUGo, its WiFi-access OTT VoIP app. Orange has its Deezer music service and Libon messaging app. Verizon has cloud storage and exclusive sports content.

While these types of things are important, they are generally the product of huge development efforts - large teams of people, big budgets, and development cycles measured in years.

What is missing is any way of creating smaller, unique "surprise and delight" services for bundles, which could capture customers' imagination, and gain extra market share or customer loyalty. 

I don't see operators thinking along the lines of "What could we do with 5 people, 3 months and less than $1m budget?" or even "What could we do with 100 teams of 5 people, 3 months each and $100m budget?".

My view is that the telecoms industry needs to start acting more like the car industry. 

It's not a direct analogy, but consider that most cars have a "model year" approach. The 2015 version of the Ford Whatever is very similar to the 2014 one. But it's got new headlights, an extra airbag, redesigned mirrors, coloured stitching on the seats and 10 other tweaks. It's essentially the same price, but it's had various changes made. This differentiates it from older models (encouraging upgrades) and from its rivals. Hopefully, it will gain the manufacturer a few % points of extra market share, and a few % points of extra margin.

In that scenario, the product manager has to work out how to add in the maximum number of new features that make a difference to customers, for the minimum extra design and tooling cost. They probably have a long-list of possible options, which gets whittled down to those which offer the best "bang for the buck". Some will be big changes - a new engine, perhaps - while others will be minor cosmetic evolutions. There's also a longer 5-6 year cycle of doing complete re-designs. The next "G", to use a telecom analogy.

A similar approach should occur in telecoms. By all means work on big strategic service innovations like cloud storage offers. But telcos should also be looking at creating and deploying new and cool extra services rapidly with small teams, or with partners. They need good ways of "on-boarding" such services - and perhaps deploying them to existing customers in a similar fashion to a handset/browser OS update.

This should happen at the level of customer segments, in a similar fashion to the car industry. Telcos could have plans or sub-brands for specific groups: music-lovers, teenagers, sports fans, home-workers, seniors and so on. Or the segments could align with regions - a Londoner plan, perhaps - or techno-psychological, like an Early Adopter Plan.

Within each plan would be standardised elements like phone calls and SMS, plus an appropriately-dimensioned mobile data offering. Then would be bundled content or partner apps. But then there could be unique, cheaply-developed and unique services, that could evolve and be added-to over time.

So for example, the Music-Lovers' plan might include Spotify in year one, but then add in a karaoke app, plus a free introductory online music or DJing lesson & discounted extra tuition. Perhaps  there could be a competition among subscribers, so once a month one lucky winner got an hour-long videoconference with a pop star. 

The alert among my readers will have noted that all these involve two-way realtime communications of some sort - precisely the sorts of apps that WebRTC can help bring to reality in weeks, given a suitable team. But that's just one of the rapid service-development tools available - internal and external APIs from the telco platform or external providers, web and app development frameworks, cloud-based hosting and so forth will all help. 

The key thing here is the management and process. There needs to be acceptance that the 5-person team works on the basis of optimising the time/price/performance variables independently. They're not forced to use internal APIs, infrastructure and service elements, but can if they choose. There's no forced mandate to use QoS or interoperate with IMS or the phone system. They can use the telco infrastructure, or they can host on Amazon if it makes more sense and saves money and time. They are free to use open-source or rebrand 3rd-party white-label products, take a DevOps approach and launch fast with minimal corporate scrutiny. They can add in OTT-style messaging or voice/video. They can use carrier billing, credit card or even Bitcoin - whatever makes most sense. The management can force the internal legal team to have 24-hour turnarounds. There needs to be an acceptance of glitches, and a good way to update / bug-fix if necessary.

It becomes the brand manager or product manager's job to do the overall bundle "design", in a similar fashion to the 2015 model-year car. It has to be cohesive as a proposition, but it also has to launch on time and to cost. Things which are too complex, costly or delayed may get canned - but because there are a dozen other new things, that's no huge loss.

Here's a random selection of potential differentiating service offers which telcos could/should offer in some of their bundles. The challenge should be "How can we get one or more of these in the bundle in 3-6 months, from a standing start?"

  • Mobile Karaoke app (between subscribers, or mobile-to-desktop)
  • Realtime sentiment analysis of calls - is the person speaking really angry, or just pretending?
  • Crowdsourced WiFi help function - "who knows how to log onto the Internet at Starbucks on Baker Street?"
  • Swipe-to-screen, pushing a web-page or image from a phone to the user's TV
  • A fan app, allowing people liking the same band or movie to send messages to each other, or set up a call
  • A concierge app for hoteliers to suggest visitors download, including zero-rated or sponsored roaming calls to the hotel guest-services desk
None of the above are "scientific". They're all just things I've thought up in a few minutes in a cafe. But telcos ought to be able to come up with 100 of these, which map to their market and customer base (using, ahem, external consultants if needed), and filter them quickly down to 10 or 20 of the most attractive, relevant, and easy-to-implement.

It is this type of grass-roots service innovation process that will make a difference - not years-long industry industry initiatives to get to a clunky lowest-common denominator that nobody will use. But it will require a software mindset, a good platform for on-boarding (and updating/removing) new services, and a lot of "air cover" to make sure new offers don't get mired in corporate wrangling. If the internal API team's proposition is that good, the new app development team will choose it. If not, they'll look elsewhere - and should not face problems for doing so. Speed and pragatism - a startup mentality - need to overrule the conservative approach to standards and network controls.

The 2015 and 2016 service plans should be subtly different from the 2014 one, in lots of small ways, which enhance the overall telecom model-year design beyond the sum or the parts.

Wednesday, October 22, 2014

WebRTC Delamination: shifting vendor roles, and recent moves from Ericsson, Temasys, Cisco, Tokbox & Mozilla

One of the major themes I address in my current research on voice and video / WebRTC trends is one I call "delamination". It's the opposite of vertical integration - the carving out of thin, narrow layers of services or products, in areas where companies have particular skill, focus or market reach.

It might be provision of a particular bit of client-side software, an effective hosted TURN server, an "enabler" of some kind, a particular security or authentication element, an especially-scalable virtualised network function, testing or analytical product, and so forth.

In a design-led approach to communications, this means that developers have both "raw ingredients" to create their own "recipes", as well as broad array of partly-prepared and pre-processed garnishes and side-dishes.

To continue the food analogy, if you really like salads, you can:

  • Plant & grow your own lettuces & make your own salad
  • Plant your lettuces but pay someone else to cultivate them
  • Buy lettuces from a farmer at the farm
  • Buy lettuces transported to a local market
  • Buy lettuces from your supermarket
  • Buy lettuce leaves from your supermarket
  • Buy chopped & prepared salad from your supermarket
  • Buy salad in a restaurant
  • Share someone else's salad
  • Watch someone else eating salad on YouTube
  • Subscribe to the Salad Channel
A similar approach is now applying to WebRTC (and to voice/video capabilities more generally). Developers - and these include intermediate developers like in-house groups at vendors and telcos - have a whole array of choices. They can build applications and services from scratch, use commercial or open-source software elements, host software internally or in the cloud - or use a third-party platform, and so on.




(This is one of the reasons it's so hard to answer the question "what's the $ market size for WebRTC?" - you can define it at any level you like, and there's potential for rampant double-counting if you include the whole value chain).

Now all this is nice in theory... but is it actually happening, and at "industrial scale"?

A few recent developments suggest that it is, indeed happening in a very meaningful sense:

  • Platform provider Temasys has announced a commercial version of its WebRTC plug-ins for IE and Safari browsers, with extra "enterprise-grade" support, intended for major software developers wanting to do full-scale deployments to business users.
  • Mozilla is working with WebRTC platform provider Tokbox to embed communications capabilities directly into the browser as "Firefox Hello". Its platform has hundreds of millions of users.
  • Cisco has open-sourced its H.264 implementation for use in WebRTC applications, and it too has been adopted in Firefox.
  • Ericsson has open-sourced its Bowser browser and OpenWebRTC client framework. This is really important both in terms of "big company delamination", and because it involves a rare step outside the telco/IMS domain for Ericsson with WebRTC. (Differentiating it from rivals ALU, Huawei & Nokia Networks, from which I've only heard about IMS use-cases for WebRTC)
  • Apparently, Ericsson's OpenWebRTC also uses Cisco H.264 (thanks Victor!)
[Disclosures - many companies quoted here are Disruptive Analysis clients for research and/or advisory work]

There are plenty of other examples too, but I'm highlighting these to illustrate that the concept is now becoming mainstream, and not just by small developers but also accepted and perpetuated by major players. This is likely to further accelerate market development - and also highlights the need for telcos and enterprise vendors to occasionally swallow their pride and use 3rd-party software or services where it makes sense. Only the very largest and most aggressive companies will be able to develop everything in-house, without using at least one other provider's platform. And for telcos, a vendor IMS box is not, on its own, a complete platform.

Monday, October 20, 2014

Apple's SIM card - less of a gamechanger than many people think

Let's get one thing clear straight away: it's not a Soft-SIM that's in the new cellular version of the iPad Air. Apple isn't becoming an MVNO, either. And it's not about to put Apple in the position of supreme gatekeeper for the world's telcos, much less charge them for the privilege.

Although the precise details remain available only in outline and inference, it appears that the "Apple SIM" is just that - a physical SIM, in familiar card form. It just comes pre-installed at the factory, but (it seems) can be removed and replaced as normal, for example with a SIM from an operator that isn't currently providing its profiles to Apple for its menu system.

This isn't actually a new concept, either. It's a variant of the GSMA's programmable SIM / eUICC standard, intended to promote cellular use in M2M devices, by helping manufacturers and users avoid some of the "friction" involved today. Operators want more new devices on their networks. Manufacturers of non-telephony devices want to avoid the hassle of setting up local distribution channels to provide SIMs. And users don't want to have to hunt for SIM cards, especially for short-term service plans.

The idea is that a device ships with a SIM card that either has
  • A "master" profile from one operator, which can then switch in/out the profiles to a partner or affiliate MNO in another country, or... 
  • (As appears here) a basic "bootstrap" profile from the device vendor that can then be used to trigger operator-specific downloads & activations from a menu. 
NTT DoCoMo has already launched its programmable-SIM services, and in discussions recently at CTIA I found out about several more that are working on their own propositions. The names of the operator groups concerned won't come as much surprise, given those that have been publicly supporting the GSMA initiative. Where this differs from some of the proposals is that the device vendor (Apple) is issuing the cards, and presumably has to bear the costs and administrative work of getting them produced.

Although the main emphasis of the programable/eUICC model is on fully-embedded SIMs, for example in vehicles or ruggedised industrial machinery, the Apple SIM model makes perfect sense for tablets as well, in a situation where some operators have signed up, but others still demand a conventional SIM.

It is not yet clear whether the Apple SIM ships with the AT&T/Sprint/T-Mobile/EE profiles pre-installed, or whether these have to be downloaded once the user chooses from the menu. It's also not exactly clear what happens if the user switches - does the old profile get deleted and another downloaded? Or does the user just swap to another that's already pre-embedded? If the device does hold multiple profiles upfront (rather than one at a time), how many can it accommodate? There's a lot still to be answered. 

However, the idea of multiple profiles (and IMSIs) in SIM cards is not new - operators like Truphone have been doing something broadly similar for years, but with MVNO deals rather than the operators' own brands upfront. The innards of SIM cards, IMSI, security credentials, MNC/MCC codes and related topics is pretty arcane - there are numerous ways to create multi-operator SIMs, and that's before we even get to fully-soft SIMs in the phone's own OS and processor chips. (I first wrote about multi-IMSI cards in 2009).

It's also worth noting that Amazon's Kindle Fire HDX's ship with pre-installed SIMs, although provided by Vodafone in Europe and (I think) AT&T in the US. The Kindles and VF SIMs, shipped from central distribution centres, can then be configured to support whichever country the subscriber lives in, with the appropriate local network credentials for the relevant Vodafone OpCo.

The obvious question is whether a similar approach could be applied to the iPhone. My view is that it's very unlikely, especially in the next 2-3 years. There are several major reasons:

  • Tablets have a cellular "attach rate" of  less than 20%. Operators want to increase this. Phones have an attach rate of 100% already.
  • Tablets don't need a phone number. Data devices can (& do) switch between networks easily, so short-term contracts and "ephemeral" connections work fine. Phones do need a stable phone number, which would need to be ported if you switch providers - and which obviously cannot be ported internationally. (We may be able to go number-free eventually, but not yet)
  • There is an agreed standard & set of business processes for programmable SIMs for M2M-type applications (eUICC), which tablets fall into. There isn't an equivalent standard for phones - and judging from GSMA and operators' clear lack of enthusiam, there won't be one any time soon. Given this model requires operator consent, that is a major stumbling block. Apple might try to end-run this with a couple of renegade operators, but I have my doubts.
  • Tablets, along with many other M2M devices, don't fit well with the retail model for mobile phones and phone-oriented SIM cards. Customers buy tablets in computer shops, not phone shops. Or they order them online. While some might overlap (especially phablets), that's the exception. Getting SIM cards into the computer retail/online channel is hard (as is getting SIMs into car retailers, camera retailers, electric appliance channels and so forth). Staff in those stores or websites are not trained to sell cellular plans. So having "blank" SIM cards pre-ship from warehouses by the manufacturer reduces friction massively. Conversely, the mobile phone channel is pretty good at selling both phones and cards.
  • Tablet usage is very different to phone usage. People have much "burstier" patterns of needing mobile access - when travelling, when in places with poor or expensive WiFi, with children, with work and numerous other variables. Many are not conducive to traditional, monthly subscriptions, so mechanisms that allow short-term use is beneficial.
Aside from this, many observers have also suggested that Apple wants to be an MVNO, or try to get carriers to pay to be one of the menu choices, or take cuts of the various iPad-specific data plans. I have huge doubts about this. I don't think Apple wants to be an MNO or MVNO. The margins are lower, the customer support headaches are higher (billing systems & queries, anyone?), there's a ton of regulatory hassle, and - ultimately - mobile services represent a market that has already peaked. Plus, many operators subsidise iPhones hugely, which eases their path into the market. Potentially, Apple could offer its own finance/subsidy plans - but that then just turns it into a bank, with all the credit risk that involves.

And of course, this is not to mention the low likelihood that operators are willing to do this anyway, and that MVNOs are not legal or easy to set up in many countries, besides. And, finally, Apple isn't going to want to put itself on the hook for operators' network deficiencies, outages and so on. It might try out one country - or even set itself up with a Dutch-style newly-liberalised MNC code (ask me offline for details of why this is interesting) and try to do roaming deals instead of MVNOs - but that's some way off as well as not necessarily being an attractive commercial proposition.

What Apple does want, as always, is higher-priced, higher-margin hardware. It makes a big chunk of its profits from selling upgrades - more memory, external covers/batteries and so on. A lot of Apple's profit comes from - essentially - selling memory chips at retail prices far above what it pays for them wholesale. The cellular variants of the iPad are £100 or US$130 higher - and you can bet that the cellular modem probably costs Apple $50 or less in incremental price. Added to that, it is possible that Apple gets some sort of "bounty" payment from carriers when users activate a SIM - which is sort-of equivalent to a subsidy.   

So if Apple can sell 10m more cellular-enabled iPads, it can probably make a tidy extra billion or so for very little effort. To make an extra billion of profit out of taking a slim margin on lots of MNO data plans would be much harder work.

Is all this a precursor to a long-run approach to full Soft-SIMs? Maybe. But my guess is that Apple will try to get in front of the next curve rather than try and subvert the status quo. It recently joined the NGMN in the hope of helping define 5G. I wouldn't be surprised if it suggests SIM-optional or Soft-SIM models in that forum. A successful programmable SIM in the iPad would give a nice weight of evidence about the possible benefits.

Overall, I think the Apple SIM is definitely interesting. But it's not the game-changer that many people think. It's not obviously a thin-end-of-the-wedge, or slippery-slope, or any other similar metaphor. It's specifically a win-win-win approach to get more tablets to use 4G networks. It learns from the failed attempts at getting 3G-embedded laptops into the market (which I predicted in 2008) by getting the processes more aligned with consumer needs. It doesn't relate to phones, at least for a long time.

I've also realised one other absolute, killer argument that seriously limits the role of any future programmable or SoftSIM model. It's one I haven't seen discussed in any articles about the Apple SIM so far. I'm not going to be revealing it on this blog either - but it's big enough that many of the "doomsday scenarios" for operators become almost impossible to realise. Instead, I'm going to be thinking about the implications a bit more deeply, and writing it up for my mobile broadband research subscribers only, in the forthcoming update to my "non-neutral" mobile broadband report due out in November. The information and purchase page is here. (If you urgently need the "big reveal", I can let you know the outline by email in advance, if you buy the report).

Wednesday, October 08, 2014

Mobile IMS/VoLTE is just for cost-saving. WebRTC & Telco-OTT are for new revenue

A question to readers: what is the overall penetration of carrier VoIP lines into the installed base of broadband connections, both fixed and mobile?

The answer may surprise you. It's currently just 8% overall.

It's that low partly because of the huge volumes of 3G devices and connections that now exist, which all use circuit telephony and SMS. It's also because carrier VoIP use is still only creeping up slowly in consumer broadband, to around 30-35%,12 years after the first such service launched. Finally, although LTE is now cracking through the 250m user barrier, only a tiny fraction currently have VoLTE.

Another question: how do you think that number will change over the next 5 years?

The answer - it will grow pretty fast. In fact, it will double, to a huge 16% penetration of carrier VoIP into the global broadband base. That will be pretty variable by country - some markets like South Korea, Japan, France and the US will be disproportionately high.

Well actually, that's not huge. Only about sixth of the world's broadband lines will have undergone "IP transition" by the end of the decade.

Why so low? Well, again 3G is largely to blame. While advanced users will upgrade to 4G, even more previous 2G users will move up to 3G in their place. But also, many LTE users will not use VoLTE, but will stick with circuit fallback. Only operators with great coverage, high ARPU and predominantly post-paid, contract user bases will tend to invest in all the complexity of IMS in the medium term, given the lack of any obvious new revenues. We see this today - even established operators in markets in Europe are dragging their heels on VoLTE, or perhaps only talking about using it in rural areas.

Why is this all so slow?

The reason is that it's ultimately all about cost-savings, maybe with a tiny veneer of new services. 

Fixed-line carrier VoIP is about reducing OpEx on clunky old switches, and in some case releasing assets tied up in big, old, phone exchanges. Property refarming, if you like. There's a small amount of enterprise hosted UC and a handful of other products/features, but they don't move the needle enough to enhance telcos' overall fixed-line revenues.

In mobile, IMS is being driven by VoLTE - either because operators are being forced into it (end-of-life CDMA moving to LTE) or in order to chase cost-savings again, from OpEx with older core networks, or with the ultimate hope of refarming spectrum assets. There is very little obvious additional benefit - VoLTE will not increase revenues and there are no obvious major new revenue-earning IMS applications.

(Some are arguing that RCS might save money compared to old SMS infrastructure, but I'm unconvinced, especially as it brings the value-negativity of joyn and the pain of getting people to use RCS applications. I'm increasingly hearing of telcos finally coming to their senses about RCS, fed up with 7 years of wasting money, resources & opportunity to vendor wishful-thinking and GSMA cluelessness).

Will WebRTC help IMS thrive? My regular line is that "IMS needs WebRTC much more than WebRTC needs IMS". It can help extend IMS-based VoIP onto a greater number of endpoints, more cheaply and quickly - for example, onto PCs and tablets that don't have native IMS capabilities and likely never will. This is indeed valuable, as it amortises the costs over a wider base, and helps accelerate beyond normal upgrade cycles. 

I am less convinced that long-tail developers will be interested in IMS capabilities exposed via WebRTC APIs - I have not seen (m)any use-cases that make sense and can't be replicated more quickly and easily via other platforms, especially given low device-side penetration and fragmentation of operators. More interesting is the potential for the reverse - accessing IMS-based services using a 3rd-party login rather than being a "subscriber" to a given operator. This allows (for instance) log-on with Facebook to an IMS-based conferencing system.

Some other innovations will also help tip the balance slightly - Metaswitch's open-source cloud IMS Clearwater is a cheaper, more flexible way for operators or developers to experiment with IMS as a platform, or create niche or prototype services and mashups. But I would argue that reducing the costs and vendor-led rigidity of IMS is necessary but not sufficient here - it may enable some new telco-cloud offers, but it still confines developers to some rigid raw ingredients like "calls" and E164 numbering. A lot of growth and value in IP communications will come from fragmentation and "de-lamination" (a theme I'll pick up on in another post).

Overall, IMS is slowly increasing its presence in mobile. But the keyword is slow. Despite the few recent launches of VoLTE, it is important to remember that devices running mobile IMS services are still incredibly rare. Apart from Korea & the US, few countries will quickly migrate 3G user bases to very high % of LTE anyway. 

Like fixed IMS, we will see mobile IMS used either for cost-savings, or to replace obsolete end-of-life core networks. There are still no clear additional sources of revenue or differentiation. WebRTC will help extend it and implicitly reduce per-user costs, and enable a few enhanced use-cases, but it is not a magic bullet.

IMS will not disappear - it will likely still exist as legacy even in the future post-IMS cloud/web telco world. But it will certainly not dominate the mobile landscape - despite its creeping growth, it will never account for more than a small fraction of total mobile-user communication events, whether measured in messages/minutes, or more modern metrics of perceived value, engagement and relevance.

Telcos need to maintain a strong and growing emphasis on non-IMS forms of communication if they are going to find new revenues from voice, video or messaging, rather than merely slowing the decline. These can be sales of major vendors' UC systems to businesses, Telco-OTT applications like Telenor appear.in, WebRTC platform plays like Telefonica Tokbox & NTT Skyway, or numerous other forms of innovative "embedded communications". While WebRTC will be a critical component here, it is not the only approach. 

Organisationally, this all remains a challenge. It is very hard for telcos to reconcile the conservative, traditional IMS mentality with the more software-centric world of modern communications capabilities. Yet that is the challenge which needs to be managed by telco CEOs and CFOs. If they allow future "communications services strategy & architecture" to be dominated by IMS, they will stifle true innovation and potential sources of value and competitiveness. 

It is right that most mobile operators continue to work on IMS, but that must be seen as just one of the paths to future services and capability exposure. Protected fiefdoms and entrenched groups must not be allowed to delay or distract other units working on non-IMS options. I hear too many examples of the "white blood cells" mistaking a beneficial (but fragile) initiative as a pathogen and trying to kill it. 




Mobile IMS is not a "special flower". C-level operator execs must make sure that it is matched by investment in other forms of communication platform & capability as well. 

IMS is for cost-saving. WebRTC and Telco-OTT is for new revenue. By the end of 2019, only 11% of global mobile broadband users will be reached by IMS/VoLTE.

If you are interested in the future role of IMS and its links to WebRTC and service innovation, it is covered in depth in the new 2014 WebRTC Strategy Report & Forecasts from Disruptive Analysis. Details are here



Sunday, October 05, 2014

Net Neutrality solution: Force telco execs & network engineers to swear an "IPocratic Oath"

I've been having one of my periodic Twitter debates about Net Neutrality with colleague & co-disruptioneer Martin Geddes. [Rough timeline is here although it blends a couple of threads & I've edited a bit to make more sense]

As per normal, it's involved bit of back-and-forth about the nature of the Internet, broadband - and how to manage networks operationally and how this relates to policy and the law. Given the current FCC and EU debates around neutrality, so-called "specialised services", fast lanes and related topics like zero-rate charging and app-based business models, it is a timely discussion.

Martin often takes the view that laws should be based "bottom up" starting with how networks actually work, taking into account the realities of TCP/IP, queuing and scheduling in routers and their integral buffers and so forth. He (and others) argue that the Internet isn't really "neutral" anyway, because at a microscopic level there are always algorithms that decide how and when packets are forwarded and in which order. 

In other words, all our cute analogies with pipes or aircraft seat-classes are wrong. All those cartoons of multi-coloured packets flowing through networks? Nonsense. That's not the way it works. And at that level I agree, absolutely.

Where I disagree is how laws and regulations should be framed, and from what starting point.

My view is that policymakers should think about "big pictures". And when the word "outcome" is used, it should be at a macroscopic level of "What do we want our citizens, businesses & society to get from the Internet and/or broadband?" and working down from there. Obviously there is a point where this political "policy" has to map onto network "policy" and operational realities. 

But stating upfront what objectives you would like to achieve - and perhaps more importantly, what you want to tell voters in elections - is a critical starting point. And obviously, there's a host of other things here around promoting innovation, enabling productivity, encouraging investment in networks & Internet infrastructure, promoting freedom of speech, reducing red-tape on businesses, encouraging healthy telecoms competition & companies, fostering employment, social inclusion and so on. Some of these will mutually conflict, as will the technical or economic realities of running the networks. That's life. Compromises, grey areas and legal terms like "reasonable" are inevitable. We can't run the world based on mathematics, even if that's how we'd like to engineer it.

The standard response for that type of argument is "right, so what's the technical standard for neutrality and reasonableness, which engineers can read the specs for?"

And it's true, it's damned hard to codify "political" policies in the form of strict technical specs and metrics. I tend to use terms such as "not differentiating Internet packets based on stated or inferred application" although Twitter is hardly the platform for detailed nuance.

But then Tim Panton dropped in a great suggestion "Doctors have a 'do no harm' rule which then needs practical interpretation in specific circumstances".

In other words, medical practitioners have an over-riding statement of "ethical principle". As it happens, the precise phrase "do no harm" is a relatively modern interpretation of a much older promise called the Hippocratic Oath. It's an interesting read, and the parallels in terms of Internet are quite pertinent:
  • "I will take care that they suffer no hurt or damage"
  • "Whatsoever house I may enter, my visit shall be for the convenience and advantage of the patient"
  • "Whatever, in the course of my practice, I may see or hear (even when not invited), whatever I may happen to obtain knowledge of, if it be not proper to repeat it, I will keep sacred and secret"
In response - and, amazingly, according to Google I appear to be the first to use the term - I suggested that maybe ISPs' execs and network engineers should be forced to swear an "IPocratic Oath" if they want to practice offering Internet services.

Initially, I just thought that was an amusing snappy comeback. But actually, I'm wondering if maybe we're at a similar point to medical practice, where the Internet has become so complex that framing laws in a way that relates to the intricate operations involved is impossible. The impasse at the FCC, and the weird wording in the revise EU Draft legislation seems to suggest this.

Instead, perhaps having an affirmation of "good intentions" - with serious sanctions available if an individual or company is found guilty of transgressing them - is an interesting possible compromise. Truly believe your "specialised services" are really "special" and not merely an anti-competitive ruse? So swear a sincere IPocratic Oath first, and put your career or telco licence on the line...

It could be a neat way to allow genuine innovation and clever use of managed network services - say, realtime monitoring of heart pacemakers needing QoS - while outlawing negative practices like blocking/degrading services ISPs don't like, or trying to subvert existing Internet competition & meritocracy by allowing "fast lane" access in discriminatory fashion. In other words, it introduces the concept of "Ethical QoS" (or "Ethical Zero-Rating & App-based charging").

Yes, I realise this is completely fanciful, and probably impractical & unenforceable. But at the very least it's a useful thought experiment, and, I reckon, a really good pun. I'd love people to contribute possible drafts in the comments, though - who knows, maybe I've stumbled on something?

(And if you haven't already noticed, I have a much more hard-nosed analysis of net neutrality/non-neutrality in this report on mobile broadband business models. Please inquire for details on the forecasts & analysis, or if you're interested in any other form of engagement around this field).

Thursday, October 02, 2014

New WebRTC Analyst Report & Forecasts published: Towards 6bn devices & 2bn users


I've now completed & published the updated 2014 version of the Disruptive Analysis WebRTC Industry Status & Forecasts Report. I believe that it is the most detailed research-company study on WebRTC available, based on ongoing primary research, extensive quantitative market modelling and a thorough investigation of as many use-cases as possible.
  • WebRTC “Democratises” realtime comms from specialists to a broader developer base
  • Use cases in many sectors: telecoms, consumer web, enterprise comms, M2M etc
  • One of the most disruptive & important web/telecoms innovations for 10 years
  • 6+ billion capable devices, and c2 billion individual users of WebRTC by end-2019
  • Evolving from original browser-based standard to richer “comms philosophy”
  • Smartphone & tablet capability for WebRTC starting to ramp up strongly
  • Pivotal role for new tier of cloud platform / API players. Also open-source critical
  • Microsoft involvement in new ORTC standard a good sign. Apple still MIA
  • “Headline acts” like Amazon, Amex, Snapchat, Google catalysing wider interest
  • Early use-cases for customer care, web calling, conferencing & some verticals
  • Enterprise will see WebRTC in both unified and “dis-unified” comms contexts
  • Telcos/SPs getting more serious. 10+ service launches, but IMS integration slow
  • Magnifies the “OTT” threat for telcos, but also eases Telco-OTT innovation
  • Numerous “gateway” sub-types for vendors to target, but moving to XaaS models
  • Monetisation of WebRTC will be heavily use-case dependent
  • Still some impatience with pace of development – but a lot happening below surface
  • As in 2013, Disruptive Analysis believes the hype is still justified
To order, see base of this page

To give some background - I've been watching WebRTC since its inception in mid-2011. I've advised numerous market participants on WebRTC specifically, and the broader Future of Voice/Video for a long time prior to its arrival; I'm also a regular figure at industry conferences & webinars covering the technology in Europe, North America and Asia. I was recently annointed as a "WebRTC Pioneer" by the guys at TMC (thanks!)

The first edition of the report, published in February 2013, set the standard for forecasts, with data ("4bn devices by 2016") used almost ubiquitously across the industry. 

Since then, much has changed. 2013 and 2014 have seen huge evolution. In particular, it has evolved from a “purist” web- and browser-centric standard, to one that has broadened to integrate with established communications systems, as well as being embedded directly into mobile apps and M2M devices. While its deployment varies in pace by sector, it is now applicable to so many different areas and use-cases that delay in any one context cannot derail the overall proposition.
Looking towards 2015, Disruptive Analysis sees the picture evolving again. WebRTC-based capabilities will crop up in day-to-day websites, as well as dedicated platforms for businesses or telecom verticals. New formats will also emerge, including integration with cable set-top boxes for interactive TV-based usage, or various other M2M and data-connectivity variants.  
The all-new 190-page report extends it the original analysis considerably. It covers all the key WebRTC use-cases & companies within:
  • Enterprise communications
  • Telecom service providers & cable
  • Consumer web & mobile apps
  • M2M &IoT
  • Cloud platforms & APIs

The report contains the most detailed & comprehensive forecasts on WebRTC adoption available from any analyst firm:
  • WebRTC-capable device numbers, by PC/phone/tablet/M2M & geographic region
  • Consumer WebRTC users for standalone & embedded voice/video
  • Business WebRTC users for contact centres, UC and app-embedded communications
  • Telco WebRTC users for VoIP/VoLTE extension, Telco-OTT services & cable/IPTV
  • M2M/IoT estimates for WebRTC integration


The report has detailed qualitative analysis & many examples of use-cases, live deployed services and lists of vendors in all the key areas. It covers the critical strategic issues facing WebRTC (standards, browser support, business models, regulation etc) and provides actionable recommendations for the main stakeholder groups. It also has full background information on market drivers such as telcos vs. OTT, enterprise BYOx trends, and the emergence of WebRTC cloud and platform providers.


The report is 190 pages in length and contains over 100 charts, tables and graphics.



It is available either as a one-off report, or as a subscription with two interim updates of c25 pages length, scheduled for publication in Dec/Jan and Apr/May timeframes. 

Pricing of the report is as follows:
  • 1-3 user licence: US$1700 for main report, $2500 with 2x interim updates
  • Corporate licence: US$2500 for main report, $3700 with 2x interim updates
Payment can be made by credit card, Paypal or bank transfer - see the end of this page for online purchasing, or email information AT disruptive-analysis DOT com to order via PO/invoice, or request further information. 

Please also get in touch if you are interested in internal workshops, custom advisory projects or speaking engagements about WebRTC.

CONTENTS


Executive Summary and Recommendations
  Strategic issues
  What has changed since 2013?
  Use cases & market segment
    Platforms & cloud APIs
    Enterprise
    Telecoms operators
    Consumer web & apps
    M2M & IoT
  Device support forecasts
  WebRTC adoption forecasts
     Enterprise/Telecom/Consumer

  WebRTC industry timeline 2014-2019
  Companies & vendors
  Recommendations
    For web & app developers
    For telecom operators
    For enterprise IT & telecom end-users & suppliers
    For network vendors & cloud platforms
    For VoIP, video & messaging providers
    For industry bodies & regulators
    For investors & consultants 

Strategic Issues for WebRTC in 2014/15

  What is WebRTC & why is it important?
  WebRTC standards evolution
    Roles of IETF, W3C & others
    ORTC
  Mainstream developer awareness & adoption
  The battle for the soul of WebRTC: Beyond the Browser
  Value-chain Richness & “The WebRTC Ecosystem”
  The Emergence of Mobile WebRTC
  Participation of key players
     Apple & WebRTC
     Microsoft, WebRTC and ORTC
     Where are the OTTs?
  Usability and interaction design
  The mix of WebRTC voice, video and data
  Regulatory considerations
  Evolving attitudes to interoperability
  Regional differences within WebRTC
  New monetisation and business models          


WebRTC Device Support Forecasts
  Introduction
  PC support of WebRTC
  Smartphone support of WebRTC
  Tablet support of WebRTC
  TV & M2M/IoT support of WebRTC
  Summary            


WebRTC Platforms, Gateways, Tools & APIs
  Introduction
  What is needed to develop and launch a WebRTC service?
  Market categorisation & vendor/PaaS fragmentation
  Platform use-cases & differentiating dimensions
  What to develop in-house, what to outsource?
  Platform/API providers: industry dynamics
  WebRTC Platforms/APIs conclusions & predictions
    Overview
    Solutions-oriented platforms
    Distribution channels for WebRTC platforms & APIs
    WebRTC platform interoperability?

 WebRTC & Enterprise Communications
  WebRTC for business: The Big Picture
    The strategic “Great Game”
    Introduction to WebRTC in enterprise
    Beyond UC: The emergence of enterprise WebRTC frameworks
    Background: General trends in enterprise communications
    CEBP revisited
  Contact centres & customer support / interaction
    Contact centres & customer interaction: Key background issues and trends
    Customer interaction: WebRTC use-cases, roles and adoption drivers
    Amazon Mayday as a Catalyst
    Growth of multi-channel & web live-chat
    Customer interaction: WebRTC deployment
    Contact centre outsourcing
    Customer interaction: WebRTC vendor strategies
  B2B Conferencing
    Conferencing: Key background issues and trends
    Conferencing: WebRTC use-cases, roles and adoption drivers
    Conferencing: WebRTC challenges and limitations
    Conferencing: WebRTC vendor strategies
  Unified communications, IP-PBXs & collaboration
    Overview 
    UC defined: Key background issues and trends
    UC: WebRTC use-cases, roles and adoption drivers
    UC: WebRTC challenges and limitations
    UC: WebRTC vendor strategies
  Application-embedded WebRTC
  Enterprise WebRTC adoption & forecasts
    Overview
    Business demographics
    Adoption of WebRTC in UC, collaboration & business apps
    Contact centre adoption of WebRTC
    Sensitivities

WebRTC & Telecom Service Providers
  Introduction & overview
  Key background trends for telcos & SPs
    Ubiquitous broadband driving all-IP communications: 4G, 5G, fibre, cable
    Service innovation, strategic direction & telco organisation
    Network-based services innovation vs. Non-network
    The “OTT model”: Decoupling access from service
    Network & OSS evolution: NFV & SDN
    “Peak Telephony” and the importance of “intent” for communications
    Fragmentation of communications – is “disunified comms” the real trend?
  Telcos’ current involvement with WebRTC
    Overview of WebRTC service segments & business models
    Telco network & IMS/VoLTE/RCS extension
    The extension strategy
    WebRTC & VoIP/VoLTE extension
    Developer IMS/WebRTC APIs
    WebRTC standards evolution for telcos
    WebRTC to extend RCS / joyn
    TV & cable integration of WebRTC
    Non-mainstream communications service providers
  Non-integrated Telco WebRTC
    Build vs. buy vs. resell vs. acquire etc.
    Massmarket telco-OTT VoIP & video-call services
    WebRTC + telco developer / API initiatives
    Enterprise & cloud
    Digital Home services
    Internal uses for WebRTC at telcos
    Investment, incubation etc
 Carrier-focused vendor landscape
    Telecom WebRTC adoption & forecasts
    VoIP/VoLTE+WebRTC forecasts
    Telco-OTT VoIP/video WebRTC forecasts
    Cable/IPTV WebRTC forecasts
 Summary

 WebRTC & Consumer Web & Apps
  General drivers and inhibitors for consumer WebRTC
  Mobile WebRTC Apps
  Video Chat & General Communications
  Social Networking & Messaging
  Games, Entertainment, Music & Media
  Personal Productivity, Tuition & Online Life
  Dating & Adult
  Advertising
  Consumer WebRTC adoption & forecasts 
    Standalone consumer WebRTC VoIP/video forecasts
    Embedded consumer WebRTC forecasts

 WebRTC, Data, IoT & Emerging Use-Cases
  DataChannel evolution
  M2M / IoT
  Emergency Communications & Public Safety
  Forecasts for M2M/IOT & WebRTC 

About Disruptive Analysis            


ORDERING

Payment can be made by credit card or Paypal below - report will typically be delivered via email within 24 hours. Please select:
  • 1-3 users vs. corporate licence
  • One-off report vs. inclusion of 2x interim updates
  • VAT details / payment for UK/EU purchases
Please email information AT disruptive-analysis DOT com for PO/invoice, or request further information. 


WebRTC Report, 1-3 users



WebRTC Report, Corporate Licence