Person to Person IMS applications – will they take off or will there be only Web applications

Note: I changed the subject from ‘P2P’ to ‘Person to Person’. P2P was not the right phrase in this context

The Telecoms industry continues to promote IMS applications

Typically the list goes something like this.



Push to talk



List management

(From the book IP multimedia subsystem (IMS) by Gonzalo and others – A book I recommend)

Before we address the problem of P2P IMS, let me reiterate that .. IMS itself is a bit like 3G. It is needed by Operators since an IP core reduces cost. In that sense, it is futile to talk of IMS

Success/failure – just like it is futile to talk of 3G success or failure.

The question is: what do we need IMS for beyond the IP core? What applications are possible? – and furthur – why cannot they be done by the Web? – i.e. where is the value proposition end to end for IMS if the Web can do most of what IMS can do – but for free?

Agreed that IMS provides QOS, but look how many people use Skype? They dont worry about QOS as long as we have global connectivity at an affordable cost?

The problems with P2P IMS applications are

a) Lack of a value proposition: Many of these applications can be done by non IMS means on the Web. So, why would I need IMS for them(say for IM). This is the main reason for the interplay between Web 2.0 and IMS i.e. IMS is the ‘Telco’ way of doing the same functions that are done on the Web. The difference is the Web is free and it is global. Telecoms is not free and it is mainly local(sub national i.e. many operators within a country not all of which can interconnect). This is a killer for P2P applications.

b) No End to End: Many of the IMS applications promoted are P2P i.e. need IMS on both ends of the network AND need IMS devices at both ends i.e. customers need IMS devices.

This is not visible at the moment and far from any roadmap I can see. So, why will P2P IMS applications take off?. The IMS devices problem is well known .. As in .. there are not any on the horizon.

c) Merged services – same problem – bigger scale: Some IMS applications are based on merged services(phone call at the same time as we view a video clip). Again the P2P, device problems apply. So this is also a non starter unless other issues are resolved (for example partnerships with other vendors).

So, questions are:

a) Are these problems valid? I think they are. I cant seem to find anyone who can reliably answer these questions – and

b) What can be done about them?

We have a precedent here in MMS .. i.e. P2P MMS did not really take off at all because it has exactly the same problem for IMS applications(in a simpler form) – a.k.a you need network and device support at both ends, the user experience needs to be seamless, the charging needs to be transparent ..

Here are some approaches who claim to address this problem (end to end IMS / P2P applications)

Much of the industry seems to be taking an ostrich like view to this i.e. somehow all operators will simultaneously upgrade + all devices will be quickly capable of supporting IMS (and will be Operator locked down i.e. other means like Wifi connectivity are not possible – :) ) and then anyone can call anyone else using a video call by IMS

For that matter – to use video calling, we strictly don’t need IMS at all .. that’s a different problem which we are not addressing for now

I don’t endorse any solution and I have no commercial relationships with them – but to create a value proposition for IMS beyond reducing OPEX – we need to address the problem of P2P IMS.

a) IPX network – you can see more using the link tab on Technology -> IP Interworking -> IPX (for general descriptions) Also, tab on Technology -> IP Interworking -> IPI Documents

IPX is undergoing pre-commercial implementation (PCI) trials. It still has many questions unanswered in my view .. I welcome any comments on IPX if you know more(pros and cons)

b) By adding a new network element like Application Session Controller or ASC. This approach is explained HERE.

c) SDP as the glue that unites IMS i.e. SDP on top of IMS. See this approach HERE(and its the most common of the approaches so far )

d) SDP as a replacement for IMS! Relatively new .. Who needs IMS when you have SDP 2.0?(pdf)

In any case, the problem needs to be addressed .. and there is little point of speaking of End to End IMS applications unless we do

I am also interested in knowing more about the experiences in Japan and Korea

While we do have interconnect and interoperability in Japan and Korea – I do not see the same picture being replicated in the west

So, my question is: are there any P2P IMS applications in Japan and Korea – specifically ones with differential charging(which is the goal of IMS from an Operator standpoint). These are topics of separate blogs – differential charging with IMS and its viability and the experience of IMS in Japan and Korea ..

However, the questions I raised before are valid .. and I don’t see any answers for them in a western scenario i.e. Why do we think that P2P/end to end IMS applications will take off in the West?


  1. Antoine says:

    “These are topics of separate blogs – differential charging with IMS and its viability and the experience of IMS in Japan and Korea ..” –> What are these blogs ? I am eally interested by such sites…

  2. Ajit Jaokar says:

    Antoine – I mean I have yet to write them! rgds ajit

  3. Todd Spraggins says:

    [Cross-posting from ForumOxford since this got picked up by Telco2.0]
    The use of P2P here is a bit troubling as most of the example applications can not function entirely without some level of central server. IM, for instance relies on ‘something’ to manage the publishing and subscription of presence and the store-and-forward management of being online/offline. I would agree, however, that there is a very valid concern around reaching critical mass for end-to-end delivery of an IMS service. But this is nothing new to the telecom industry as it has made several transitions of this scale before (3G being a very good example).
    Furthermore, the industry addresses this issue often by creating intermediaries e.g. gateways, so that multiple generations of technology can interoperate. For IMS this is done through re-use of OMA services, something already in use with 3G networks, and the creation of interworking specifications like SMS-to-IM, and VCC for CS-to-PS voice continuity. These become the intergenerational stepping stones.
    With that said, the industry is being awfully cautious (polite for slow) in the deployment of IMS services, and this could be heavily attributed to the point of no visible value proposition as you point out. But I’m not sure that IMS as a body of specifications, is necessarily in a position to solve that. For the most part, the standard is about creating a session control plane that is separate from the application. This allows for the industry to be innovative in how applications are created/delivered; unfortunately, innovation isn’t the foremost attribute of this industry, so we just have not seen much action here. For now, all of the methods you describe, IN-SCS (with AppTrigger’s proprietary marketing term of ASC) and SDP are very valid service delivery models which do not belittle the IMS purist in any way (in fact I would argue that most in telecom embrace them today).
    Finally, overreacting and pushing something like SDP2.0, where all you need is a softswitch++ or a SIP enhanced IT application infastructure just exasperates the end-to-end problem, ignoring the need for standardized infrastructure to build upon as a foundation for interoperability to reach critical mass. So, yes SIP is an internet standard and can work ‘naked’ in monolithically integrated environments, but it takes something like IMS to give it AAA, mobility and specific operator centric profiles and architectures so that it can be successfully deployed in a repetitive and interoperable manner.
    (Author’s postscript: For my given argument, I will say that IMS is way to heavy for what it principally brings to the table, that is a standard SIP session control plane. This can be attributed to the standard’s embedded hooks to business models which are rapidly becoming dated and other beholden special interests. Nevertheless, at the moment its the only game in town to provide the required stable NGN infrastructure. An analogy would be envision making the transition from analog to digital without the DS1.)

  4. Ajit Jaokar says:

    I agree the phrase ‘P2P’ is misleading in this context. I have changed it to ‘Person to Person’ whic h is more in line with what I wanted to say. Thanks for your insightful comments rgds Ajit

  5. Jerry says: