Are NGN and 3G really going to help people by providing useful, effective and attractive new services? What are the premises that will guide us in exploring next generation telecommunication services based on SIP?

Introduction

Second Generation (2G) mobile networks and traditional fixed networks provide a relatively complete set of voice services. However, even when enhanced by the use of Intelligent Network (IN) technology, services evolve slowly and the paradigm remains essentially the same. The roadblocks they encounter relate to the definition of the service itself, the network-centric architecture and the technology, which is mainly centered around voice.

Next Generation Networks (NGN) and Third Generation (3G) mobile networks are designed to overcome these limitations by reorganizing the network architecture in order to separate the provision of services from the network (see Figure 1), to merge information and telephone technologies, and to introduce open protocols, like the Session Initiation Protocol (SIP). At the same time, they are expected to provide new and successful added-value services that will meet the needs of users and operators alike. These networks include all the ingredients necessary to meet a wide range of user expectations with regard to services.

The article considers the applications that next generation and 3G networks offer. Each is based on a set of capabilities which can be used to define new services. While SIP is a major protocol, it is not the only one. Bringing together all these capabilities and interfaces results in a complex space within which existing services can be enriched and new services defined.
Finally, the traditional architecture based on extensive call control, in many cases using the IN, must evolve into one with light basic connectivity control and powerful service application platforms with open service development capabilities.

NGN and 3G Capabilities

NGN and 3G have three main strengths as regards their service capabilities.
First, they bring together the information and telephone worlds (see Figure 2).
Second, they are a move away from intelligence in the network to intelligence in edge servers complemented by active networks (see Figure 3). And third, they no longer need a state machine with triggers for the call service control; transport in the Internet Protocol (IP) domain is more stateless (see Figure 4).

Consistent Information and Telephone Worlds
Third generation networks bring subscribers the best of both the voice and data worlds. Although 2G and 2.5G networks are already offering web services alongside traditional call services, there is still little coupling between the information and connectivity domains:

• A service can be a pure voice connectivity service: The call does not involve any interaction with the information domain.
• A service can be a pure information/web service.
• A service can comprise triggered processes in both domains. As an example, the “click to dial” service, which encompasses a family of generic services when it is considered as a process within the information domain (e.g. white pages, dynamic yellow pages, personal or shared address book, diary event, call by
pseudonym, call by topic rather than addressee, etc), triggers a process in the connectivity domain (e.g. person-to-person session setup, conference or any other SIP session setup). Microsoft, for example, is providing tools for these applications on the PC and for web application creation. The converse is also possible with, for example, the termination of a connectivity process leading to the sending of an e-mail. Similarly, SIP capabilities, such as pushing a web page and process, can be initiated on request during a session.
• A service can be a mix of both information and connectivity services, a typical example for 2G being the call center. Applying a similar concept to 3G requires further work on defining generic interfaces and protocol enhancements (see next paragraph). The service concept, comprising parts in both the information and connectivity domains, is a potential basis for many future applications.

Benefits of SIP
SIP is a signaling protocol for creating, modifying and terminating sessions. These sessions can be multimedia conferences, IP telephone calls and similar applications consisting of one or more media types, such as audio, video and whiteboard. SIP uses the Session Description Protocol (SDP) to create sessions and carry session descriptions, which allow participants to agree on a set of compatible media types. Participants can be people or gateways to other networks.
Communication can be via multicast, or a mesh of unicast relations, or a combination of the two. A major benefit of SIP is its simple yet powerful third-party call control.

SIP is simple and extendable
SIP, which is being developed by the Internet Engineering Task Force (IETF), is a textual protocol defined for client/server architectures. Its simplicity enables services to be developed easily and rapidly. Like the HyperText Transfer Protocol (HTTP), SIP is designed so that extensions can be added relatively easily. In common with other text-based protocols for the Internet world, it has end-to-end transparency. SIP-T is an extension for communication between softswitches. ISDN User Part (ISUP) messages with parameters can be transported by SIP as MIME-type (Multipurpose Internet Mail Extension) attachments.

SIP encompasses more than the strict call process. It not only includes the registration process, but also supports information domain capabilities such as the presence protocol (SIMPLE). As a consequence, the use of SIP is a key factor for achieving consistency between the information and telephone worlds

SIP enables next generation voice services
SIP can carry the Simple Object Access Protocol (SOAP), enabling one endpoint to access applications at another endpoint. For example, a server application can launch an application on a SIP handset (e.g. to change its ring tone for an incoming call). In the same way, account information relating to a prepaid card can be carried from a SIP application server to a proxy to control the duration of a call. Terminal-to-terminal applications, like network games, are also possible. Intelligence and status information will reside in the terminals, which will use whatever applications they need. Such applications are “end to end”; the network value is mainly in service acceptance, routing, network resource mastering and accounting.

Open Application Programming Interfaces (API) for HTTP – Common Gateway Interface (CGI), Hypertext Processor (PHP) and servlets – were a key reason for the success of the web. In addition, there is a large community of web developers. SIP is copying the web, with several open SIP APIs (SIP CGI, SIP servlet, SIP PHP) currently under development.
A key driver behind the introduction of SIP services will be their integration with web applications. SIP services for the web will be described in eXtensible Markup Language (XML) scripts. XML gives the operator or user the power to create and use an abstraction layer in order to easily define service scripts.

A variety of new services will be deployed in the near future. The roll-out of Microsoft Messenger means that each desktop is now a SIP endpoint, and each web developer is equipped with XML and a SIP Service Creation Environment (SCE). This is a major opportunity for the SIP industry, both in the corporate domain (SIP private branch exchanges) and in the carrier domain (SIP proxies and softswitches). Alcatel offers both product lines, which go beyond telephony to new multimedia applications. These are strong grounds for using SIP services in the enterprise and public Internet domains today, and in the IP mobile domains in the near future.

The value of these new services for the network operator is that they increase either the subscriber base (new voice over IP lines, new corporate virtual private networks) or the network traffic (connectivity), or the network value (e.g. quality of service management, location, billing or presence management). It is not found in cannibalizing voice and leased line services at a lower price, even though there is a market for best effort voice in most countries.
Table 1 shows some typical applications.

SIP’s extensibility, interoperability and presence capabilities make it possible to provide new and differentiated services at minimal cost. Most services will combine more than one of the previous domains. The mix can be so diverse that personalization and the appropriate environment become as important as the library of services.

Benefits of HTTP
Second generation communication services are almost all driven by the control protocol. In 3G, control flows can originate from two sources, since both SIP and HTTP can be used to set up a connection.
The current methodology used to set up a simple connection (look and feel, procedure, behavior) on a mobile terminal is also likely to be used for multimedia calls, so SIP will be the main protocol. However, complex connectivity services, such as multiparty managed conferences or training applications, involve more complex control and many more semantic control elements. In this case, the obvious approach is to use web technology for the main service control flow, with SIP only handling the simple point-to-point connectivity control.

SIP and HTTP can be used separately to control a connectivity service. In addition, they can cooperate to control a service. Furthermore, HTTP is a content carrier protocol. A basic SIP-driven call can turn into an HTTP-enhanced SIP service. As an example, one can:

• Push a web page while the connection is being established, in a similar way to the 2G Calling Line Identification Presentation (CLIP) service. However, it can be much broader in concept, extending to all the phases in a call (answer, connection, release, assuming these phases are identified).
• Pushing this web page at invite time can lead to multiple HTTP-driven choices on what and how to proceed.

Benefits from Mixing Information and IN capabilities
A typical example is enhanced terminal location. In 2G, location information is used mainly for location-dependent call routing, such as emergency calls on a highway. Within a 3G environment, the capabilities of such services can be enhanced to provide the subscriber with data (simple position coordinates or a map or information on the nearest requested points of interest with instructions on how to get there from the user’s current location). The same capability can be enhanced to provide an e-commerce server with information on incoming customers and their profiles as they approach a related shop. Different servers can support these capabilities either for a fee or free of charge.

The information domain also provides capabilities that can be used for designing IN services. For example:
• Presence can be used as a capability to enhance telephone services: Click to call via contact list.
• Messaging and instant messaging: Deliver messages in the correct format; sort and choose to read a selection of messages from a list of messages from different sources and in a variety of formats.
• Information-enhanced calls: Journal of calls, unanswered calls, caching as you talk (personal diary, virtual video cassette recorder, profile), etc.
• Auto-registration, self-provisioning, self line-testing, service self-profiling / script creation, filtering, parental guidance, access rights and control, security, encryption, integrity management, virus screening, etc.
• Community services: Closed private communication groups, syndication, content subscription, Virtual IP private branch exchange services for enterprises with a number of sites.

Benefits from Subscriber Data Organization
Subscriber- and service-related information can be used to enhance connectivity services, which can benefit in two ways:

• Personalization: Setting rules for the call or the script for the service (e.g. “my own filtering”, “my own call pattern”) via a browser enables a subscriber to implement powerful personal services, such as filtering, log information on used services, and much more. For example, the subscriber can route a service according to multiple criteria and can set up facilities.
• Communities: Setting up communities, such as a VPN, can be used to handle the same kind of services for a closed group. Such services may range from a simple private addressing scheme to community-specific and organization-driven logic for applying connectivity services.

Architectural Support

The strengths and features described above can be used effectively in new services, either by end-to-end intelligence or by using some network value, assuming that the underlying architecture is modified accordingly. To provide the necessary network value versatility and achieve the required time to market (comparable to end-to-end service maturity) without impairing the robustness of the system and its ability to evolve, the architecture must comply with at least the following main rules.

Shift Complexity from Traditional Call Control to Connectivity Services
Traditional call control is a complex standard call state model which serves any trigger needed by IN services. One of its main benefits is that it is common to all users. A drawback is that it is slow to evolve and cannot be personalized. However, the functional split between call control and IN still supposes that the basic process is a call with a given control semantic.
In the case of 3G and SIP, the call architecture is minimized; the value comes from a variety of connectivity services which can be developed using a powerful SCE.

Strong Decoupling and Stable Interfaces Between Domains
The information and connectivity domains have their own evolutionary paths. Loose coupling between them and stable interfaces are basic requirements for providing versatile services that can evolve in line with user needs. The main relationships relate to:

• synchronization of information (e.g. personal information manager data);
• service triggering (e.g. click to dial);
• capabilities offering (e.g. messaging);
• control flow (e.g. web-driven connectivity service).

Powerful and Open SCE
Apart from basic connectivity control based on SIP, which is essential everywhere, all other services will be based on a service creation concept. This is a step forward compared with the one currently used for IN services. Given the number of semantic elements to be handled, the challenge for this SCE is to combine power with ease of use. Furthermore, not only has it to create new services, but also to wrap existing services into the consistent framework of the service environment and management system.

SIP is an extensible protocol. Applications residing in terminals must be consistent with the SIP semantic elements of services residing on the application server (see Figure 5). Consequently, the SCE relies on variable SIP semantic elements, or capabilities, depending on the terminal application. The SCE must be versatile enough to support evolution of the various capabilities, including SIP.

Open SIP service environment
SIP services exist in a distributed environment which is open for adding value and therefore presents an opportunity for third-party service providers.

The following reference service was developed within an Alcatel multimedia program. It shows how a SIP service and a web application can work together in a distributed environment. “Web advertisement with voice button” is a typical web-based “click to dial” application. This idea is not new, but it is a good example of voice and data convergence.

An advertiser can use the web interface to change its routing policy (a “follow me” service). A Call Processing Language (CPL) script is created for this routing policy and sent to the SIP application server. If a new click-to-dial call arrives, it is routed to the advertiser’s new destination, as shown in Figure 6.

The developer of this web application does not need to have any knowledge of SIP. He or she just needs to know how to use a high level API for third-party calls. In subsequent steps, this kind of API could be replaced by an open SIP API, such as JAIN SIP Lite or SIP servlet.

Conclusion

Services sometimes evolve by enhancing well known ones. Initially minor aspects may then become key success factors or even the core of the service. Rather than waiting for a successful service simply to turn up, the best approach is to analyze the environments in which the widest variety of services are likely to exist, then provide the most suitable architecture to support and stimulate this evolution. The article is a step towards this process.

References

1. http://www.nttdocomo.com/home.html; NTT DoCoMo is Japan’s largest mobile service provider, with more than 40.7 million subscribers (estimate as of March 2002).

2. D. Attal, S. Wolfe: “Warning: bumps in the road ahead!”, Alcatel Telecommunications Review, 2nd Quarter 2002, pp 97-101 (this issue).

3. http://www.orange.com; Orange is the Europe’s second largest mobile operator with operations in 20 countries across Europe and beyond.