Sonus In Session

Some thoughts on ths from my perspective: 
 
- Very different issues with adoption of IMS in mobile and fixed telecoms space 
- Key issues in mobile IMS relate to implementation in devices, integration with radio, desirability of services to consumers, fragmented routes to market for mobile devices & services, integration with both OSS/BSS IT and also the typical subscription pricing models. 
- Fixed IMS could at least use PSTN-replacement VoIP as the “anchor tenant”, especially as consumers were buying broadband anyway, and could disaggregate the access device (gateway / modem / router) from the user device (phone) by means of an analog terminal adapter. Plus, operators had an incentive to use IMS for fixed VoIP as they had old, creaking opex-intense PSTN infrastructure that was reaching end of life. Also, VoIP on fixed actually works quite well (now) especially on the fast DSL & fibre seen in places like Japan and France. 
- Mobile IMS doesn’t have an anchor-tenant service. The original candidates suggested were PTT (niche and suffering from latency in original SIP versions), image/video-sharing (better done online as realtime isn’t that important), mobile IM (competing standards such as IMPS, presence (kills battery, questionable use case). There was also some vague talk of “combinational services” but nobody really knew what they were.  
- The handset issue has been consistently overlooked by the IMS industry. I wrote a report in 2006 pointing this out (and had been researching it from 2004-5). Nobody knew or bothered to define what an IMS handset actually was, beyond one technical interface. The handset was just a little box marked "UE" down in the botton left-hand corner of the architecture. Nobody thought that actually pre-defining services might be important in terms of developing and integrating the UI and quite a lot of the firmware. Nobody thought about the impact of smartphones, multitasking, interaction between apps on the devices, giving IMS apps access to underlying device APIs, issues with dual-mode WiFi, the role of the SIM card, how to get IMS clients onto devices for those not supplied through carrier channels etc. This is still not properly fixed, 5 years later – and worse, it now has to compete with flexible smartphone app / web / widget architectures that get access to native APIs, can spread virally, can be updated regularly etc. 
- More recently the Mobile IMS emphasis has been on VoIP (complex, needs a lot of work, very radio-dependent, likely a lot of work on acoustics and interconnect as well) and RCS (late, bureaucratic, unclear value proposition to the user, massive competition from apps like Facebook, unknown risks of cannibalising SMS revenues) 
- No clear developer model for IMS. Where’s the SDK for Rovio to develop Angry Birds for IMS? 
- No clear web integration model. How do you integrate “Like” or “Retweet” or “Vote Up/Down”? 
- No upfront thought about business model evolution. Why should IMS services have “subscribers” rather than “users” or visitors? How does a freemium version work? These things can’t just be left to the billing system – they are fundamental to the design of the architecture. (eg if 95% of all events are non-billable) 
- No upfront thought about two-sided business models (eg advertiser-funded) and collection of user data in such a way that it can be monetised by the operators & third parties. No built-in analytics capabilities etc.  
 
Yes, we will see IMS slowly permeate into the mobile arena. But I'm much more pessimistic than Dan about timelines for VoLTE - I'd say that it will be around 2016-17 before there's wide use globally with all the problems fixed. There will also be plenty of LTE operators that don't use IMS at all, so any future applications will either need to work standalone within a restricted community, or interoperate with Skype, Facebook & others as direct peers rather than treating them as second-class citizens. 
 
Dean Bubley 
Disruptive Analysis

I'm very much in agreement with Mr. Collins' statements. I think the issue has mainly been economic. I guess I would challenge the general premise that IMS has been particularly slow to be adopted. We're talking about an architecture, not a technology or a product. I'm not sure that the expectations for adoption embraced by IMS vendors was ever realistic. You don't replace your infrastructure overnight. The rewards of migrating to IMS, I think, are substantial, but never really strong enough for operators to overcome their inherent tendency to drive as much life out of their investments as possible -- or to insist on a killer app before investing in equipment. I'm more confident about my belief that vendors serious mangled the marketing of IMS. It was all about destination and reaching nirvana, with very little practical guidance to assist with the journey. All of that is changed now -- I think. Most vendors are not really talking about IMS these days, at least they shouldn't be. Operators don't want to hear about technology and features and standards compliance. They want to know how vendors can help them transition their networks and keep subscribers happy. It just so happens that demand is catching up with supply -- the particular benefits of IMS have finally been matched with a need. Operators still aren't going to overlay IMS across their networks overnight but at least there's a bit of urgency connected to adopting the architecture. IMS has had its issues and there's no doubt that the architecture is completely superfluous from an Internet purist perspective. And for some reason, that's the perspective that has dominated the IMS debate. From the operator perspective, was there really any chance they were not going to follow what was prescribed by the 3GPP? Sure, IMS is expensive and complex but what's the alternative for a next-generation architecture? With IMS, it was always about how and when -- not if.

Greg, Dean and Joe- Thank you all for your great comments. It is nice to hear so many perspectives. I have been a supporter of IMS from its early days, but these days with the proliferation of smartphones and over-the-top (OTT) multimedia applications running on an IP cloud, it makes me wonder if IMS will truly takeoff. While many have been talking about all the possible (IMS) multimedia services to come, a bunch of these apps are now available on smartphones over 3G! The IT and Telecom worlds are converging. Let's think about how best to provide new innovative services to consumers first, regardless of whether it is a traditional IT or Telecom technology (or architecture)- then the telcos can leverage their installed base to enjoy a two-sided revenue model - that is the Telco 2.0 model that service providers must adapt.

Great article – and I’m looking forward to part two. 
 
My perspective on this brief history is very slightly different and maybe sheds some light on the poor “take up” of IMS that you point out. 
 
When Intelligent Networks (IN) hit the scene I was working at British Telecom, rolling out the latest upgrade to digital switching systems. The concept of IN came along and suddenly seemed so intuitively “correct”. In fact it was hard to imagine that operators were previously installing new services (every few years) in hundreds of local switches. Added to that, the switches were highly proprietary with unique operating system and specialist languages.  
 
IN truly brought general computing systems into the telecom’s world. So was this successful? Honestly I believe that IN was hugely successful… but perhaps not quite in the way that people imagined.  
 
Over a number of years, the IN systems were deployed very extensively in both fixed and mobile networks, and these have allowed the rollout of both financially lucrative and “customer experience changing” services including highly flexible number routing, portability, prepaid and account based services etc.  
 
What it perhaps did not deliver, as you pointed out in the blog, was a plethora of advanced value added services, changing and adapting monthly or weekly and being tailored to individuals and small groups of users, as well as for the masses.  
 
There was a time when everyone seemed to be talking about trialling services in hours or days and launching new services every week. We can all agree that did not happen. 
 
So how about IMS? Firstly, I don’t think there has been quite the same hype about the promises of IMS (a good thing, but maybe that’s also a problem?). Secondly IMS is perhaps not quite as revolutionary as IN (and maybe that’s another problem?). Thirdly, the plethora of services has arrived, with more appearing daily, and this had nothing to do with either IN or IMS (smart phone apps – just in case anyone’s been sleeping). 
 
So, accepting that IMS has not quite “happened” yet, what could change this? Is there something that could suddenly make IMS the must have network architecture?  
 
The question I asked myself once I had read your Blog is…. 
Are there services or other demands on network operators that would ensure the rollout of IMS across the globe? 
 
The principle concept behind the IMS architecture is that it should provide an infrastructure capable of delivering IP based multimedia services with security, quality of service and of course such services must be chargeable. 
 
IMS has not been promising (as IN did) a huge number of new services arriving at your phone on a weekly basis, but instead it is promising that significant new multimedia services can be delivered in a way that the alternative infrastructures (public internet, public accessible cloud etc) cannot.  
 
The key word here is “significant” new multimedia services. IMS is not about small services taking advantage of VoIP or Presence or FMC. We all know that these can be delivered without IMS. It’s time to think big and consider some more revolutionary services that can change business and how it is conducted. 
 
Even if I thought I had identified the killer services that can help IMS grow into a revolutionary “must-have” architecture, I would probably not post them here. 
 
However, I will venture that these killer services (if they happen at all) are going to be corporate services rather than consumer ones. One view I have heard often is that all new consumer services can now be provided by the web, Internet and cloud, combined with smart phones and quality is nice to have – but not critical.  
 
High-end Corporate services demand quality, demand integration and demand to be managed properly over a lifecycle. Such services require something slightly more sophisticated for delivery and perhaps this is where IMS can shine? 
 
Perhaps it is time to start thinking about what businesses could achieve through integrated multimedia services offered over IMS.  
 
Most businesses today have a diverse set of networks and services on which it relies. The public Internet, for hosting its web pages and blog and occasional webinar. A public telephony network, for other forms of voice communication. As for video, if it is used at all, it is in a limited way and once again probably separate from the above.  
 
What if businesses started to think about an integrated set of multimedia services, information and content that other businesses and customers could access over an IMS network?  
 
Corporate services not consumer services are the best candidates for being the key to IMS success. But if time passes by without these services being identified then IMS will probably be condemned to a uneventful demise, replaced by the next new technology or architecture (no doubt with a new set of acronyms). 

Sorry to be a bit late to this party - very good topic and most great comments are already made.  
 
I was going to refrain from commenting on the IN angle (lest revealing I've been around that long) but cannot help myself (for the same reason). 
 
 
 
Considering there is always unrealistic hype/oversell in any fundamental architectural change, IN was rather successful in my opinion. IN managed to centralize most revenue generating (but boring) services (800, VPN, CNAM etc). It democratized SS7 and enabled new vendors to enter the market only on the "server" side but not touch the "client" side. IN/AIN fell short on the over-promise of "one exciting new service every month". 
 
 
 
Regarding IMS, I actually think there was even more oversell - with promise of TV-over-IMS, Gaming-over-IMS in the early days. At the end of the day, the acid test question: "what can I really do with this architecture that I cannot do today? (that I want to do and pay for)" did not have a good answer. Hence the slow uptake of IMS.  
 
 
 
Agree with Dean that fixed and mobile networks have very different perspectives on this. Fixed is still struggling, while mobile seems to have rode mobile data explosion all the way to VoLTE. Interestingly, some converged operators (Fixed and Mobile) are asking themselves why maintain 2 different infrastructures (one aging) if they are already deploying IMS for mobile. Also a great call by Dean on highlighting the oversimplified "UE". Industry is still suffering from that oversimplification (RCS - anyone?). 
 
 
 
I am not yet sold on Richard's point on corporate services (which I interpret as Enterprise / Business services). Primarily because Enterprise and Consumer service needs, deployment environments and economics are fundamentally different; and most of those fancy blended services exist today. Besides, penetrating enterprise markets has always been challenging for the service providers - that I hope will change with IMS. 
 
 
 
I guess my short answer as to why IMS has not taken off as planned is that - architectures (with unclear value) that are defined by big committees and in need of adoption by large organizations almost never take off as envisioned. The things that take off like a rocket are almost always end-to-end closed/proprietary services/systems (designed by for-profit entities) that have a crystal-clear high value prop for the consumer on the street who can vote instantly with their wallets.