.comment-link {margin-left:.6em;}

Wednesday, August 02, 2006

IMS Services Forum - Hypothesis

We're right in the middle of preparing our strawman framework for the IMS Services industry brainstorm on 4-5 Oct. (Update on process and stimulus speakers to follow...summary: everyone's getting involved now!).

All brainstorms need a good 'hypothesis' to work against. Below is our's re IMS, using the headlines 'Situation, Complication, Question, Answer, What Next' (see also complementary Telco 2.0 hypotheses). It describes the context for the event:

1.) SITUATION – Lack of clarity of how to thrive in an IP-based world

- Decline in traditional sources of revenue due to market saturation, new technology, regulatory changes, and new (non-traditional) competitors.
- Telcos are looking to dramatically reduce costs: to become leaner and more efficient, and at the same time find new sources of growth.
- IMS, as part of NGN development, aims to enable telcos to compete in an IP-based world – reduce costs, increase operational efficiency and time to market and stimulate the growth of new revenue streams.
- IMS Insider/STL research shows that Telcos are not clear about what applications and services IMS should support and how precisely they will make money out of them.

2.) COMPLICATION – NGN/IMS development is still technology-led not business-led

- This lack of clarity is symptomatic of the more fundamental commercial uncertainty about how telcos will make money in an IP-based world. (See
www.telco2.net/manifesto and www.telco2.net/report)
- Fundamentally: the industry is not clear how money will flow between devices, connectivity and content/application services in 3-5 years time (mature markets at least): cf. Plethora of ‘Free’ TV, ‘Free’ broadband, ‘Free’ telephony, ‘Free’ handsets offers in the market today.
- Telco management is increasingly unclear about where to invest its resources, so ‘Next Generation Networks’ are being built to support vertically-integrated telco business models which are increasingly threatened by new market entrants.
- Commercial and technical functions do not have a common vision of the required applications and business model(s)
- Investors are not convinced by telcos’ corporate strategies (commercial and technical).

3.) QUESTIONS – in the minds of forward thinking CTOs…

- Where should we focus our network investment and development efforts? Invest in access network, core network, content delivery network, SDP, FMC, CPE (home hub)?
- What NGN architectures do we need to build for different business model scenarios
- What is the (real) role of IMS? What does IMS enable the operator to do that can’t be done by Skype/Google/Yahoo/MS, or P2P/in the handset
- How can we (network architects/IMS leaders) work more effectively with our commercial colleagues?


1.) Fully understand the strategic context of ‘Telco 2.0’, ie. how to make money in an IP-based world (by playing to operator strengths).
2.) Fully understand your emerging competitors (internet, media and commercial transaction players)
3.) Explore different business model scenarios and their impact on NGN architectures.
4.) Adopt a more adaptable and loosely-coupled technology platform approach – around your (genuinely) unique telco assets
5.) Focus IMS on what it’s best at and do the rest in other ways.


a.) Read the ‘Telco 2.0’ report (we would say that, wouldn’t we!) and discuss in-depth with colleagues probably in a workshop environment.

b.) Be proactive in defining an integrated (commercial and technical) action plan to make your (commercial and technical) strategy ‘Telco 2.0 compliant’. ‘Telco 2.0 Compliant’ = a business model that maximises the leverage of telco assets, extracts maximum value from non-network assets (billing, logistics, physical retail distribution), takes advantage of the unique position of the operator in the network. The innate advantages of telco operators over network edges (internet players) are: costs, scale, security, flexibility, trust, integration, standardisation.

c.) In the meantime, challenge yourself, your colleagues and your vendors about IMS using these prompts:

The Trouble with IMS...

Here is a summary of the current theory behind the revenue growth potential from IMS-based services:

“We (telco operator) and our partners will create compelling applications for customers. Our partners will be especially keen to work with us (telco) in preference to others because a.) we (telcos) provide the best channel for reaching customers (consumers/businesses) and b.) because IMS gives us (telcos) so much better functionality vs alternative channels (internet giants, direct, other service providers). As a result we will generate more revenue, and make up for the decline in traditional telephony revenue sources. We, and the rest of the telco sector, will therefore attract more investment and our share prices will rise.”

Clearly, some kind of realtime-aware service delivery platform is required for realtime services. But there are some important questions to ask:

- Who actually gets to deliver those services?

- And, is it wise to tie up so much of the AAA, billing, identity, profile, preference, and QoS infrastructure in the same pot?

If the telco isn’t delivering the service - which increasingly is the case with IM, multiplayer gaming and video sharing - is integrating so much in a giant ‘SIP proxy-cum-application server’ really the right way forward?

We analysed the IMS white papers of the major NEP vendors. Most don’t start with the user. Most start with the technology. Here’s a summary of our analysis of the general approach to IMS today:

User Experience is not fully considered

- By closely tying network functions to access, you're limiting users' choice of access networks and/or having to invest in complex roaming schemes which don't have a good historical precedent of success (eg. MMS).

The IMS services vision is not compelling

- The IMS service examples of vendors and operators today (Presence, VOIP, IP Centrex, Multimedia, Unified Messaging, etc) tend to either be:
o Rather trivial (and implementable on existing switching systems), or
o Assume total service control by one operator (single number - how likely am I to take all my services from one source?), or
o Assume an implausible degree of technical and business co-operation across multiple operators (e.g. call forwarding from one number at reasonable rates and without creating any forwarding loops).
o In the enterprise market the focus is on "selling the disease and then the cure": ie. expensive firewalls that require expensive and inflexible holes like SBCs (Session Border Controls).

Telco competitive advantage/sources of value are not being leveraged

- The normal IMS service examples replicate "tablestakes" features that IM providers offer for free, and don't leverage any of the unique assets or advantages of the operator (billing, enterprise relationships, distribution, logistics, brand, etc).

- There is a misleading assumption in IMS that a central service provider is required for each technical function, and that the edge nodes are incapable of co-ordinating themselves: a simple historical example is flow control & retry in TCP/IP which eliminates the need for a "circuit aware" network; a more complex example today is the adaptive jitter buffering in VoIP codecs which mitigate the need for QoS.

- An excessive emphasis on ‘QoS’, when the real crisis on the Internet is ‘Security’ (and lack of underlying identity infrastructure).

- IMS strategies today fail to justify how value-add is being created from the IMS infrastructure when people can already talk on Skype for free -- and most of the QoS problems on Skype are really either internal to the PC (e.g. contention for CPU or network access) or local to the customer premises (first-hop to WiFi access point).

- Many of the IMS functions when decomposed are replicated for free on today's Internet (e.g. basic dyndns.org dynamic DNS routing -- eg. I can make a SIP call that will reach me wherever I am -- the "home agent" function is very cheap/free).

Poor understanding of internet design philosophy

- ‘Seamless mobility’ is already a reality: it's called the Internet, and offers perfect functioning interconnect of all nodes from all places (more or less – NAT, firewalls and censorship notwithstanding).

- The major limitation of IMS is the focus on SIP; too far up the stack. Most software apps are *not* SIP-based, even if they can be shoe-horned into SIP; or they don't speak the IMS dialect of SIP. There's nothing wrong with SIP per se, just that the lessons of 1981 and the end-to-end principle (
http://en.wikipedia.org/wiki/End-to-end_principle) were ignored -- the foundational design principles of the Internet, which made it so successful.

- An example of where the "SIP: everything's-a real-time-circuit-because-we-can-charge-for-that" approach breaks down is: voicemail. I want to be able to record VM locally on my device at maximum audio quality and then upload it directly into another person's VM box -- and also not make their phone ring. This is a natural task for web services, and unnatural for SIP.

Cross-Carrier Implementation

- Interoperable inter-carrier IMS has the potential to be a hideous nightmare to deploy -- as soon as new features start coming out, we've got a nasty feature interaction problem.

- IMS has technical risks, e.g. forcing traffic to follow ridiculous routes for roaming customers.

Device vision is blurred

- What we really want is devices that are "bearer aware" (rather than "bearer agnostic") and understand the non-functional constraints of the access network and adapt accordingly. Example: don't do my scheduled laptop back-up over expensive 3G network. This is generally ignored today.

Not focusing enough on the real telco opportunities from IMS

- Email spam has shown the limits of (internet) edge nodes to be able to coordinate to resist "negative-value traffic". Some great progress, but nothing close to what's normal on the voice/SMS networks (due to the identity and billing infrastructure of telcos and the legal/regulatory environment).

- HSS is the crux of IMS – a valuable identity store. That part will survive, because the IP/Internet model doesn't have a lot to say about provisioning to networks. (Technical standards exist like 802.1x, but business ones don't.)

Lack of appreciation that operator market control is disappearing

- Mobile IP centrex and other enterprise solutions face an adoption barrier with employees, who have a diversity of handsets from many operators.

- Integration with other services that the users already have (e.g. Yahoo photos) doesn't seem to have been considered - living inside a telco bubble (excuse the pun).

So, maybe we should review our IMS plans and focus on how we can support the things that are really broken in both POTS telephony and SIP+Internet VoIP telephony ...from the user’s point of view. Why? Because these are the things that people will be willing to pay for.

Below is an initial check-list to run through with your IMS vendor:

1.) POTS Telephony – what’s (really) broken from a user's point of view:

- Integration of call handling and messaging across fixed and mobile domains.
- Control over voice messaging: easily create, forward, browse and search messages.
- Interactions between businesses and consumers: badly timed and targeted today, layers of redirection at each end to get right person, laboriously slow and insecure exchanges of personal and transaction data via voice dictation.
- User mobility between home, travel and work domains: inconsistency and incompatibility, difficulty with pricing and apportioning usage, ultimately leading to things like carrying 2 mobiles.
- B2B calling an issue: no exchange of presence, or directory data; unencrypted, and exposes multinational companies in particular to espionage.
- Opaque charging for users, particularly when interacting with premium services and/or across national boundaries.

2.) Pure SIP+Internet VoIP telephony – what’s broken from a user's point of view:

- Absence of reach and adoption (who to call?)
- Device and service compatibility
- Spit (Spam over Internet Telephony) and security problems
- No packaged solutions via mass retail and distribution channels

[A more detailed version of this piece appears in this month's IMS Insider report - subscribe at www.ims-insider.com]

The Editor, IMS Insider - editor@ims-insider.com

Comments: Post a Comment

Links to this post:

Create a Link

<< Home

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