Wednesday, February 08, 2006

ETP on NGN (Detailed Comments) 

Some time ago I posted that the European Telecommunications Platform (ETP) published a paper On the technology, business models and regulatory aspects of NGN.

The ETP is a consultative body which was formed in Brussels on 4th February 1998, following the opening of the European telecommunications market on 1st January of that year and consists basically of the European operators and their manufacturers.

The goals of the ETP are to promote self-regulation by the industry, encourage the ongoing development of competition and clarify operational and strategic business issues.

The published document on NGN analyses the Next Generations Networks along these lines, and although I do not completely agree with every detail, I consider this paper as a very valuable input to the current discussion within Europe regarding an update of the New Regulatory Framework and the current market definitions.

Considering the composition of the group, the document is generally pointing in the right direction, but sometimes one gets the impression that no consensus was achieved to take the last obvious step in deriving the final conclusions. So the group seemingly could not jump completely over their own shadow.

My major points of critique are:
  1. There is the intrinsic assumption that the NGN is NOT the Internet.
  2. Although all over the document the openness for a variety of innovative services is mentioned, there is also all over the document the idea about service control.
In more detail:

Since there is no common definition of the term Next Generation Network, ETP has adopted the ITU-T definition as a starting point:

ITU-T defines NGN as a packet-based architecture fostering the provisioning of new/emerging services through a loosely coupled, open and converged communications infrastructure.

As I have already pointed out extensivly in my posts reporting from the joint ITU-T/IETF workshop in May 2005 and WTF is NGN, this ITU-T definition is nice, as long as you do not look closely what ITU-T SGs are making of this in detail, and the IETF view is completely different.

The basis for the bright future of a plethora of converged services is the emerging next generation network. This will consist of:

• various next generation access networks, which will allow true platform competition,
• global next generation core networks with nearly unlimited bandwidth in the backbone and
• next generation service control, which will provide the framework for intelligent and convergent service creation

I have no problem with the first two points, but a serious problem with the last within the NGN environment.

The demands for real world implementations are clearly and correctly recognized:

Operators worldwide are striving to deploy new solutions that can adequately address the demands being placed on them by market and by technological developments:

Demands from businesses for:
  • Integrated data, voice, and video
  • Ubiquitously available electronic communications services with high speeds and media flexibility
  • Flexible Virtual Private Network (VPN) solutions
  • More innovative services and network intelligence—security, storage, application layer routing, and adaptability—to support the trend in enterprise markets toward better integration of their networking and information systems
Demands from consumers and governments for:
  • Broadband access to every home
  • Business-class telecommuting: High-performance, highly available, and secure to support the growing population of telecommuters who work from home or abroad
  • Attractive pricing bundles for data, voice, and video services, with both wireline and wireless access methods
  • User-friendly network and information security
  • On-demand multi-media services
Demands from investors include:
• Increased revenue
• Improved profitability
• Broader service areas
• Greater productivity

It is also recognized that the most important challenge for service providers in addressing these demands by using their existing solutions is:
  • Offering data, voice, and video services over complex collections of service-specific networks that are costly to operate and unable to deliver the flexible, integrated services demanded by the market
  • Deploying a broad range of access methods and protocols that are better suited to providing connectivity and differentiated, profitable services than today’s access methods
and that there exists the vision of one network for all services that delivers:
  • The mobility of wireless networks with ubiquitous accessibility to services
  • The reliability of the public switched telephone network (PSTN)
  • The speed of innovation of IP technology and applications
  • The security and service isolation of private lines
  • The operational efficiency and cost reduction of a common, consistent infrastructure
  • Greater speed to market for a much greater variety of services
Note the emphasis on speed, cost reduction and of course security and reliability.

And the we have the famous NGN funcional entities Access, Core and Service Provision/Control

Now I do not have any problems with Access and Core, but I have a problem with Service Control and Provision. And in addition, what is this little Application on top?

I am just fine with Access, Core and Application. BTW, nobody has ever clearly defined what a Service is. And what about Do-It-Yourself Service Provisioning?

... and it is immediately mentioned in the document:

... Finally, the management and provision of services needs to evolve to a next generation service control capable in providing features such as Identity Management, Policy Management, Mobility Management, Dynamic Session Management – allowing the operators to provide personalized services on a per user basis and develop innovative services.

I agree that this "services" are required, but I may not need an operator to provide these services (and I also doubt that operators are capable of developing these and other innovative services, there are others already well on their way).

The document now discusses the Core:

The migration to IP networks provides operators with sufficient flexibility in their cost base to reduce both Operational Expenditure (OPEX) and Capital Expenditure (CAPEX)

... and the Access quite correctly:

Enduring competition, the main goal of EU Information Society policy, can be fostered only by developing access competition and can thus only be achieved by establishing multiple access infrastructures. ...

The next generation access infrastructure and the edge technology will be digital and it should support the whole range of services including speech, video and internet access.

Fixed, cable and cellular operators today have the technological means to upgrade or build access networks with a large digital bitpipe allowing the delivery of triple play services.

Provided regulators can establish a favourable environment this will lead to true infrastructure competition and vanishing of access bottlenecks.

and then we come back again to NGN Service Control:

It starts quite well:

In a next generation environment services will no longer be restricted to voice and internet access. Convergence of media, computing and communications will lead to a wide variety of services, which add value for the user and the community. Although the technological basis for this evolution is here, the business case still needs to be explored. In this situation it is paramount to deliver services quickly and efficiently. This is not a new prerequisite. In the telecommunications industry the concept of an intelligent network was elaborated in the 1990s in order to centralize the service logic of the PSTN/ISDN. Although number translation services for freephone and value-add services rely on intelligent networks the further evolution is rather limited.

Ok, self critique is always good, but what is the conclusion? Now comes a paragraph where I have some problems in agreeing:

The Internet has achieved an independent "service logic" through the hourglass model, which decouples the infrastructure from applications. [not Service Control] The Internet, however, has grown from a scientific background and relies on trust, which cannot be assumed throughout today's environment. The public Internet allows easy deployment of services but lacks basic security features and digital rights management for access control which should be expected from a global telecommunications network as a basis for sound business cases.

Hold on:
  1. The Internet is End-to-End, and so is security on the Internet. And this can be achieved and is achieved now.
  2. I would like to have all the money earned currently on the Internet with sound business cases. The Internet is currently used by many businesses as a global communications network already.
Next comes a list of items needed in the service environment, which also contains QoS mechanisms.


In service control? QoS is basically a function of Core (although some say there is no QoS problem at all) and primarily in the Access.

The next statement is interesting in this context:

In order to fulfil these requirements the next generation service control will be functionally independent from the network layer and from the access. The converged next generation core network needs to become a reliable basis for the quickly evolving services, which should be accessible from all kinds of access.

Agreed, but ff it should be functionally independent, how is the QoS controlled?

The next statement is provocative (for me to give a nasty answer):

The service layer itself should provide a framework for service creation. Today the IP multimedia subsystem (IMS) is seen as the future concept of a service creation environment. IMS has been developed by the mobile industry. ITU and ETSI-TISPAN have adopted the IMS specification and adapted them to a fixed network environment. Standardisation of service environment is necessary to allow for interoperable services and for the benefits of a mass market.

Sitting currently in a hotel room in Sophia Antipolis exhausted after the TISPAN mid-plenary, I am unable to give a fair statement to this one. No comment, but my position regarding IMS is well known by the readers of this blog ;-)

A bit later these Service Control issue is somehow relativated (is it?):

Next generation service control might be developed functionally separated or fully integrated with the infrastructure (next generation access and next generation core).

Next generation service control needs to be accessed through the core network by the customer premises equipment.

So direct access by the user transparently through Access and Core:

Although the access and the core are seen as technologically transparent for services certain restrictions from a business point of view might apply.

Aha, one step forward, and immediately rowing back? I can imagine the discussions on this paragraph and the adding of the last sentence :-)

The next part of the document deals with the regulatory implications:

The vision of "one network - many services" underlines the necessity and explicitly forces to arrive at a technology neutral approach. Although vertical integrated business models might remain among others, the boundary between network and services needs to be reflected in the regulatory approach.

A clear statement to re-consider the current EU framework and the market definitions

The paper does not try to educate and give food for thought to the regulators like GSMA, but only hints on challenges ;-)

e.g. on Legal Intercept:

Lawful interception becomes more problematical if, in addition to delivering a service (e.g. VoIP), also pure connectivity is delivered by an operator as the traffic moves from the core to the access. Packets may take different routes and the only feasible way to intercept communications may be in the access network.

Access to emergency services and Universal service.

About economic regulation:

The technologies for a next generation network are available and mature. The business and investment models in the telecommunications industry are not.

I even doubt the first statement. Definitely the second ist correct.

The regulatory framework has moved from four markets to a more fragmented approach by defining 18 service markets which apply to all 25 Member States. The process of refining the markets, establishing if working competition exists and defining regulatory measures is still not completed. It however has turned out that the potential for deregulation has hardly been used.

The markets according to Commission Recommendation 2003/311/EC as well as the remedies are very much connected to legacy infrastructure and services and do not sufficiently take into account the picture of a next generation network painted in chapter 3. To what extent, NGN services fit into existing markets or will generate new unregulated markets is not evident today. Some players might provide integrated solutions including access, core and services. Others might provide services only. These services will increasingly resemble network enabled software applications. NGN itself is not a market. NGN services will lead to new markets and changing market structures. The framework should not restrict this process. However there is a tradeoff with the need for regulatory certainty and the need to avoid foreclosing competition in the future.

This clearly indicates that the EU framework and the market definitions need extensive rework for NGN.

And finally also this paper contains 10 points for a new regulatory environment (why always 10? ;-)

The new regulatory environment ...

• should be based on technology- and provider-neutral regulatory criteria:
Regulatory asymmetry leads to competitive asymmetry and creates the incentive and the opportunity for artificial competition, which is not based on market demands. It is of utmost importance to find a proper and workable definition of technological neutrality.

must be based on legal principles drawn largely from competition law:
Current and likely marketplace developments demand a statutory structure, which makes regulation dependent upon a finding of lack of effective competition. Although reliance on competition law is a long term goal, ex-ante regulation is likely to be needed for some time if some foreclosure of markets occur, which can not be solved with other instruments.

• non-economic regulation should be hands-off regulation:
The societal regulatory regime ought to pursue a regulatory hands-off approach with as light a touch as possible. This hands-off approach should not only apply to the European level but also to the national level. However, harmonisation on a European level (e.g. regarding EMC or terminal equipment) remains necessary to gain the benefits from a European market and to achieve user protection.

• economic regulation (if required) should focus on enduring bottlenecks:
Although it would be desirable eventually to base economic regulation entirely on competition law, it might be necessary to deal with network bottlenecks while they still exist. These bottlenecks are likely but not exclusively to be found in the access network.

• should allow for innovation and investment:
There needs to be a favourable regulatory environment for investment and innovation in Next Generation Networks. This is a necessary precondition for the achievement of the goals set in i2010, which was announced by the Commission on 1 June, 2005. It promotes an open and competitive digital economy and emphasises ICT as a driver of inclusion and quality of life. NGN can become a means to achieve the policy goals of the Commission.

• should provide legal certainty:
For investment in telecommunications legal certainty is a necessary prerequisite. The current
regulatory framework does not provide this certainty (e.g. the definitions of PATS and ECS do not reflect some technological evolutions).

• should focus on services not on technology:
Every asymmetry in technology triggered by regulation can potentially distort the market. Therefore technology decisions should be left to the market. Regulators should not interfere by one or the other technology. It is therefore necessary to define the boundary between technology and service. The Next Generation Core itself falls out of the regulatory scope. It is the services that are subject to possible regulation.

should avoid fragmentation of markets:
A common market allows the competitive provision of services and products. Fragmentation could endanger the position of the European industry and operators and has therefore to be avoided. NGN interconnection will be based on standards voluntarily agreed upon by industry in standards bodies. That does not imply that standards need to be imposed.

• should balance harmonisation and innovation:
Harmonisation is certainly necessary to avoid fragmentation and to allow for interoperability of services. It has however to be recognized that harmonisation in an early stage could stifle innovation by favouring a possibly suboptimal solution. Priority should be given to market forceswhich will lead to harmonised solutions at the right time. This is probably different for different layers. Innovation at service level is much more important.

• should address the question of cross-border services:
In a next generation network the possibility for cross-border services exists. It has to be assured that national operators are no worse off than competitors from abroad. Therefore the EU should try to accede to agreements with other parts of the world. NGN has a global focus. That's the reason for global standards.

The new regulatory environment should rely on market forces and competition. Non-economic regulatory goals should be pursued with as light a regulatory touch as possible. In that way next generation networks will significantly contribute to the political goals set out in i2010.

To summarize: a document worth reading.

Comments: Post a Comment

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