Fixed Mobile Convergence
Chapter 5, The C in FMC
The ultimate goal of fixed mobile convergence (FMC) is to optimize transmission of all data, voice and video communications to and among end users, no matter what their locations or devices. In the more immediate future, FMC means that a single device can connect through and be switched between wired and wireless networks.
This chapter will explain how fixed and mobile networks come together in the delivery of converged telecommunication services, and will discuss VoIP-based converged solutions.
The C in FMC
Table of contents:
1. Convergence technology fundamentals
2. Levels of integration
3. VoIP-based convergence
4. The dual nature of dual mode
5. IMS and MMD fundamentals
The delivery of converged services implies the involvement of both the end-user device and the network in creating the converged communication experience, allowing the user, either transparently or with explicit actions (depending on each service definition), to access a service through a variety of access networks. Essentially service is always available wherever there is coverage via one of the converged access technologies. Such availability is achieved potentially via different levels of fixed and mobile access integration.
A converged service and applications also benefit from information stored in user profiles and some context information such as user location and real-time availability status (also known as presence). The combination of user context and user profile allows an operator (and the subscriber, if the user profile is user-customizable) to define a number of converged access policies and reachability profiles, depending on location, time of day, user preferences, and the access network type the user is camped on at a given moment.
Various technologies have been proposed in the industry to enable fixed-mobile convergence. Most of them are more oriented to converged services delivered by a mobile network operator (like UMA/GAN), since the level of integration there is higher and implies the use of the existing mobile network standard interfaces such as the GSM A interface and the GPRS Gb interface.
Other technologies -- such as the IMS-based approaches -- are more neutral to the type of operator delivering the service and can be adopted more widely across the industry, as the level of integration with the underlying fixed or mobile networks is looser, because the converged service experience is realized at the Service/Application layer.
Let's then delve into the analysis of service availability and details of specific levels of integration and their impact on the user experience.
The success of a given FMC solution is a matter not only of making sure subscribers can access a service via particular access methods, but also of making network selection and switchover seamless and unnoticeable unless explicit notification is desired as part of the service definition ("you are now in the home zone where calling is free"). The seamless mobile experience users are accustomed to in a cellular environment should therefore be preserved and replicated as closely as possible in other access networks and application domains.
Of course, this broad requirement presents significant technical and usability challenges. Solutions to address them have been found and keep being identified as the industry successfully deals with the intricacies of delivering the converged network service experience.
From the integration point of view, convergence solutions can be broadly classified as "application-level" solutions (where an application on the terminal and application servers in the network cooperate in delivering the converged service experience), or "vertically integrated," that is, based on the user terminal cooperating with the network below the Application layer. In this model applications are not really required to support the delivery of the converged service experience, as the underlying infrastructure creates the appearance of a seamless connectivity via heterogeneous access networks. The former case is supported by technologies such as Voice Call Continuity (VCC), and the latter case is supported by UMA/GAN and femtocell approaches described later in the chapter.
Vertically integrated approaches can be then seen as ways to interface to the existing core infrastructure without the need for fundamental changes in network operation. This leaves the service delivery paradigm virtually unchanged and keeps it independent of the method of access. In other words, the investment necessary to bring vertically integrated solutions to life is concentrated in the access, terminals, and business models.
The application-level solutions, conversely, tend to leave the access network invariant to the deployment of converged services (no need of special access controllers or devices other than classic wireless routers with DSL or LAN interfaces in the premises where service is accessed through noncellular transport). These solutions, however, require terminals to have special software clients, and the core network to cooperate with these devices at the application level. The core network also must make sure that service delivery is uniform and service settings remain synchronized in the converged networks' domains.
The transition of traditional PSTN telephony to VoIP is contagious. The rise of interest in Voice over Wi-Fi is the latest evidence of this trend. The cellular industry is not far behind with a slew of new radio interfaces focused on the support of multimedia services such as CDMA EV-DO revA (also known as DOrA, to be used by the operators relying on 3GPP2 standards) and HSPA (to be adopted by 3GPP-bound operators) being rolled out around the world. The practical deployments of the VoIP-based solutions in cellular environments, however, have to deal with many issues, both technical and nontechnical, mostly absent in the fixed environment.
For one thing, cellular operators tend to maintain more stringent control over the access to the radio resources necessary to provide cellular services than wireline operators do over local loop or cable. The reason for that is quite simple, in that expensive licensed spectrum is shared among subscribers, and operators seek to maximize the return on their investment.
Second, the currently deployed cellular networks cannot fully support the VoIP service. For instance, they are lacking mechanisms to support guaranteed QoS for packet data services. On the other hand, although the new technologies capable of supporting VoIP in cellular environments are already available, they are still months if not years away from widespread deployment, thus making cellular VoIP service support very limited, and only available in some vertical segments of the market, in the short term.
Finally, cellular VoIP requires new devices or new clients in the existing devices, which in turn calls for significant investment and long-term effort to be put in place by device manufacturers and their suppliers or independent software vendors.
For these reasons it can be argued that VoIP in a wireless environment, at least in the coming years, will be less likely to experience the level of freedom and growth we are witnessing today in wireline broadband. Despite that, the long-term future of cellular wireless VoIP is bright. We are convinced that competitive forces and economies of scale will eventually lead to parity between VoIP services over fixed and wireless networks and even the potential of uniform treatment of non-operator-controlled VoIP applications.
Unlike non-operator-controlled VoIP, the operator-controlled VoIP services4 require the deployment of specific telephony applications and the necessary interworking infrastructure needed to interface to the legacy circuit-switched voice subscribers both on cellular networks and in the PSTN. It is also necessary to ensure voice call continuity in areas with mixed VoIP and circuit coverage, or in other words, seamless handover of active calls between VoIP and legacy circuit TDM.
However, once IMS-based VoIP is uniformly deployed in the converged cellular and fixed networks, the resulting access independence of the converged core will ensure the right user experience. The mechanisms to ensure service continuity and active call handover between cellular and noncellular can at that point in time be implemented at the IP layer, using mature technologies such as Mobile IP (see RFC 3220 ) or other IP mobility support protocols.
Until that happens, the industry will have to deal with a difficult task of building FMC solutions combining dissimilar packet and circuit voice services implemented over heterogeneous access networks. The rest of this section is devoted to the discussion of such solutions.
The contemporary voice FMC solutions supporting operation over Wi-Fi and cellular access networks are most often called dual-mode solutions. Such dual-mode FMC solutions can be based on a variety of technologies. The most prominent are the two standards-based approaches: One is built around IMS; the other is based on UMA/GAN, defined for the GSM cellular systems. Both approaches effectively converge circuit cellular voice and VoIP over Wi-Fi/broadband by "hiding" the Wi-Fi access media and signaling from the cellular core network.
However, that's where the similarity ends. While UMA/GAN attempts to tackle the problem of delivering converged services from the lower layers of the protocol stack, the IMS enables convergence at the higher layers. The IMS VoIP approach places the handling of the voice call in the IMS core network and allows for a significant part of the voice traffic to be offloaded from the cellular network core. Unlike IMS, the UMA/GAN standard enables the support of a Wi-Fi infrastructure by making it look like a set of GSM base station controllers, thus making Wi-Fi appear as just another 3GPP radio interface and requiring all traffic to traverse the 3GPP core network.
To better understand these aspects and more, in the following sections we discuss the fundamentals of both IMS and UMA/GAN FMC methods.
The IP Multimedia Subsystem (IMS) and the Multimedia Domain (MMD) are the 3GPP and 3GPP2 versions of the same thing, that is, an IP- and SIP-based system defined to handle multimedia signaling in the wireless domain. This system supports communications among SIP user agents accessing an IP network and various SIP servers within the network, using the 3GPP and 3GPP2 systems' packet data networks' access services.
There are only minor differences between the 3GPP and 3GPP2 versions of IMS, and from almost every practical angle these versions can be considered similar, if not identical. There is in fact a cooperation agreement in place between 3GPP and 3GPP2 organizations, whereby 3GPP2 adopts the 3GPP IMS specifications, only with minor modifications to suit the unique 3GPP2 market needs.
Over time, the intrinsic access-independence of IMS (albeit IMS still has some access-specific details surfacing mostly at the Protocol layer) has prompted its adoption as a core enabler for other wireless access technologies such as WiMax and even for wireline networks.
It is important to point out that the proliferation and applicability of IMS to multiple access technologies has been its goal from the beginning. IMS was originally defined by 3GPP around the year 2000, following intense activity by the 3G.IP industry focus group, initially driven by AT&T, at that time interested in both wireless and cable networks. Therefore, the driver behind the standardization effort was to define a solution that, while delivering an IP-based platform for wireless multimedia services, as required by the wireless arm of AT&T, could also be used for other access technologies, especially cable.
It is no surprise, then, that as of late the wireline operators' community is increasingly considering IMS as the platform for their VoIP services. For instance, the BT 21st Century network, the first major-scale IP transformation project in the world carried out by an incumbent wireline operator, will migrate over time to IMS after an initial deployment based on an architecture closer to the server model. Cable operators are also embracing IMS (as it was meant to be from day one, now we can say!), through their CableLabs industry forum.
But why are all these different branches of the telecom industry embracing the IMS and not its alternatives? There are two main reasons:
- The IMS is not only a standardized architecture for delivery of VoIP service, but also a general platform supporting all kinds of IP-based multimedia applications (text, images, instant messaging, conferencing, presence, and video telephony).
- The IMS core network can be shared by multiple access technologies, so it can be deployed by operators intending to pursue the convergence path, seeking to enter triple or quadruple plays, or just looking for more flexibility in the "last mile."
Download and read The C in FMC in its entirety.