Tuesday, December 28, 2010

The Rich Communication Suite

The first time I heard about RCS, this year, when a company asked if Mobicents SIP Presence Service (MSPS) was compliant, I was curious about what it was, but I promised myself, after 6 months struggling without a MSPS release, we would not dive into another layer of over complicated specs, no way we would consider one more possible dead end. In short, we ignored it… till another company asked about the same thing, and another one, and it kept going… In the end of last summer we were involved in a big RCS network plan, and it was perfectly clear to me, the SIP Presence industry was asking for RCS, and RCS only. So I went through an evaluation of what was RCS about, to finally decide if MSPS should embrace it.

RCS stands for Rich Communication Suite, and it is a suite of IP telco standards, which works out other standards. Confusing, the least… RCS is done by the GSMA workgroup, and does not defines any new technology… What? This is getting weird!  My first thoughts, also the fact that it was done in IMS fashion, through numbered "Release"s, made me fear the worst, after all the GSMA members possibly overlap with 3GPP and OMA, chances it could be the same people, using same methodologies, drawing the RCS specs.

So I spent some time reading the specs and let me tell you, one must praise GSMA work, quite simply, RCS takes on IMS and OMA specs, rearrange them, simplify as much as possible, so it can be applied to today's networks and services, as fast as possible, an using agile methodologies! Sounds great, sign me in!

Wait a minute, I know, some would say it is all wrong, we should instead had first the basic layer, the RCS simplicity, and only then we should have it extended to provide the zillion features IMS has, and which nobody cares or needs. Not a surprise, I fully agree with such view. Also some say it is too late, that the consumer already decided for services which are not provided by operators, maybe not, hopefully not, in my humble opinion, RCS may be the very last chance, at least for SIP Presence.

Let me try to give you an insight of what we are talking about, RCS has 3 releases, each upgrades the previous one. I will focus on SIP Presence only, but RCS touches more than SIP Presence, it also works other services such as IM.

RCS Release 1 evolves around the concept of the Enhanced Address Book (EAB), an evolution of the usual address book. In short the address book is decorated with enriched information, coming from different services. This plays nicely with today's wishes for cloud stored information, unified social networks status updates, contact content such as portrait icons. I'm not going into technical details, but I for sure am someone who is aware of the design issues around SIP Presence, its hard time scaling due to huge traffic, the dozens of ugly workarounds to make it work, and RCS is a nice step forward into the right direction, there are simple decisions that deeply simplify the network design, making it more like "old" presence networks, which simply work. One remark, it takes quite an effort to define this endorsing IMS and OMA, 27 pages of functional description, plus 39 of technical realization, it should be a lesson for everyone in these standard bodies when defining more extensions or new versions.

The RCS Release 2 effort focuses on enabling access to rich communication services from a wider range of devices. In short it tells that the user has multiple devices, for instance a mobile phone and a PC, possibly concurring for services, and adapts Release 1 for that. It also introduces the Network Address Book, which is just the realization that the EAB needs to be in the network and sync the multiple user devices.

The RCS Release 3 mostly consolidates Release 2 features, and adds some minor enhancements, such as preparing the network for different usages of it, for instance users with devices, which are not connected to mobile network, instead only have broadband connections. In my humble opinion a very important and positive decision, it's about time to consider these scenarios and find out new opportunities. It is weird to say this, but the fact that the industry finally acknowledges that content sharing between two users may happen off the voice/video session is a victory, welcome to the world not session centric. Can you imagine what would be the outcome if we have specs that release the session protocols from all these extra services almost nobody uses, how much simpler, cheaper and efficient the session networks, services and clients would be?

As I said, RCS is a big step in the right direction, a revolution without new technology. For MSPS you can expect to target RCS compliancy as soon as possible, as a matter of fact the developing tasks for such work area already in the Issue Tracker, with a total estimation of about 200h of work, at this point we just need to understand what 3rd parties are interested to collaborate, to come up with a release date, Mobicents does not have the resources to walk this path alone, or perhaps I should say, not before it may be too late. Please get in touch with me if you are interested in contributing. We know that RCS needs Mobicents too, an open source implementation with a strong community behind it.

Stay tuned.

References and Additional Resources:

RCS Homepage
RCS Release Documents
RCS Market Survey


  1. Great post!
    Come test our open source RCS Client: http://code.google.com/p/boghe/

  2. I concur, very good post Eduardo. Please also have a look at OMA Presence Access Layer (PAL). This attempts to provide an aspect-oriented overlay for watcher-clients on top of OMA SIP/SIMPLE Presence 2.0.

  3. Nice overview. Would be good to see feedback from operators who are already deploying RCS for their users.

    Of the big networks that manage Address Books for hundreds of millions of users - Google, FaceBook, Yahoo, Microsoft... are there any that announced RCS interoperability?

  4. Ivelin, that is the ultimate challenge, not only for RCS, even SIP too. It seems that RCS is a global bet by operators and vendors, and it's fairly easy to put RCS clients on mobiles, but unless these big global IM services (GTalk, MSN, Facebook, Skype, etc) jump onboard it is very difficult to converge.

    I think the ball is now on vendors, if RCS implementations prove to be more efficient or cheaper to run, there is a chance. Perhaps the major problem is that overcomplicated and over featured network below, generating lots of SIP traffic to do very simple things, in my humble opinion it would be much easier to separate IM and Presence...