Friday, March 10, 2006

Why do we need IMS? - 3G IMS explained 

I finally wanted to know how the IMS really works, so I purchased the excellent book "The 3G IP Multimedia Systems (IMS), Second Edition" from the two "Finns" Gonzalo Camarillo and Miguel A. Garcia-Martin.

The book gives an excellent overview on most aspects of the IMS, such as general principles and architecture, session control, AAA and accounting, security, policy, QoS, media encoding and transport, presence and instant messaging. The book is well structured and I like that most topics are coming in two sections: Internet (plain IETF) and the IMS, so one also sees the difference. There is one exception: the charging architecture ;-) This is BTW one of the most complex topics.

Remark: maybe in the next edition a third section will be available: Internet, 3GPP and ETSI TISPAN for fixed access and what the differnces are - this would be nice ;-)

The information is quite detailed, e.g. in session control you are really walked though the call flows via all functional "boxes", and every message has a full header message example. BTW, all figures (including header examples) are available here.

I have only one serious criticism The usage of Tel URIs, phone numbers and the mapping of E.164 numbers to SIP URIs is not covered very well, adequately and in detail. It seems that the authors are not quite firm in this area, as one can derive from the fact that they still reference RFC 2916 regarding ENUM!

Of course there are also some interesting statements on IMS, e.g. I was especially interested to read in the Introductory Section 1.3 "Why do we need IMS?"

The authors raise the question by themselves:
... In fact, any cellular user can access the Internet using a data connection and in this way access any services may provide ...
and further down:
... This means that any given user can install a VoIP client in their 3G terminal and establish VoIP calls over the packet-switched domain. Such a user can take advantage of all the services that service providers on the Internet offer, such as voice mail or conferencing services...

... Why do we need the IMS, if all the power of the Internet is already available for 3G users through the packet-switched domain? ...
My remark: and in addition the same services are also available via any packet-switched domain, not only 3G.
The answer is threefold: QoS, charging and integration of different services.
Ok, QoS is the usual stuff I will only believe when I see it. What I am missing here interestingly is one topic always raised by the IMS-freaks such as mobile operators, TISPAN vendors and the GSMA is security and reliability., one of the main arguments for the IPX(GRX/CRX).

At least the authors spelled out the main reason: charging.

The arguments for this are really cute - my comments in [..]
... Another [the main] reason for creating the IMS was to be able to charge multimedia sessions appropriately [keep the existing business model]. A user involved in a videoconference over the packet-switched domain usually transfers a large amount of information. Depending on the 3G operator the transfer of such amount of data may generate large expenses [one may say prohibitive expenses if you are roaming in a visited network], since operators typically charge based on the number of bytes transferred [typically 10$/MB]. The user's operator cannot follow a different business model to charge the user because the operator is not aware of the context of those bytes [poor guys].
This is all well known, it is called discriminative pricing. But now it comes:
... On the other hand, if the operator is aware of the actual service that the user is using, the operator can provide an alternative charging scheme that may be more beneficial for the user.
What? This is really cute. Touching. I am moved to tears. Or are you guys pulling my leg? Yes:
... for instance, the operator might be able to charge a fixed amount for every instant message, regardless of its size. Additionally, the operator may charge for a multimedia session based on duration.
Aha, here we are: this is what the operators want: if I use e.g. Skype for IM, I do pay for some bytes, now I will pay 0,20 cents per message like for an SMS. The reason is that somebody has to pay for the expensive billing systems.

I also have some doubts about the 3rd topic: providing integrated services (this reminds me of the term ISDN).
... operators want to be able to use services developed by third parties, combine them, integrate them with services thy already have, and provide the user with a completely new service ...
Also here, seeing is believing. I am really waiting for the first real implementation here.

I found another nice statement a bit later in the IMS Architecture section 3.4:
Before exploring the general architecture in the IMS we should keep in mind that 3GPP does not standardize nodes, but functions. This means that the IMS architecture is a collection of functions linked by standardized interfaces. Implementers are free to combine two functions into a single node (e.g. into a single physical box).

In general, most vendors follow the IMS architecture closely and implement each function into a single node.
Of course they do, because they like to sell single boxes, because you still get more for a box then for a (SW)-function.

Anyway, to summarize: a book worth reading.

The reason why RFC 2916 is quoted for ENUM, is because RFC 3761 was not available when IMS was first conceived, in 3GPP Rel-5. If you look in 3GPP Rel-6, you'll find a reference to RFC 3761.
Post a Comment

This page is powered by Blogger. Isn't yours?