Analysis of Key Issues of Fixed Mobile Convergence

In recent years, a growing focus from the global communications industry has been put on Fixed Mobile Convergence (FMC), into which the big international telecom players have participated, especially from 2005. BT took the lead to launch Bluetoothbased Fusion service. Recently TeliaSonera has expressed that they will converge GSM mobile service with Unauthorized Mobile Access (UMA) based wireless broadband service, when fixed service converges with mobile service in Denmark. Deutsche Telekom, Telecom Italia, Spain Telecom and even Vodafone, and others have also actively paid attentions to FMC.

    The report presented by Informa, a professional investigation company, showed that FMC services will demonstrate a trend of faster development. Before 2011, the global FMC services will have 92 million users. By that time, the users’ annual expenditure on FMC services will reach 28 billion US Dollars, up to 30% of the annual telecom revenue.

    The IP Multimedia Subsystem (IMS) architecture has been adopted by the leading telecom standards bodies including ITUT, 3GPP, 3GPP2 and TISPAN, and is recognized in the industry as the ideal network architecture to realize FMC. Among all the standards organizations, 3GPP’s standardization develops fastest, as starting some work of Stage 1 of R8 version. On the other hand, IMS faces many difficulties in the actual commercialization, of which some crucial ones have seriously hindered the progress of the deployment of IMS and FMC[13].

    First of all, the main revenue of the global telecom operation is still coming from the basic voice services. The revenue of the Chinese telecom operators’ voice services even takes up over 75%. No matter how the network evolves, voiceservice users and their revenue are of the first consideration of operators. But the IMSbased voice services cannot guarantee QoS, and the provision of other similar realtime services is still to be demonstrated. Meanwhile, because of the inertia of the users’ utilization habit, the services provided by traditional network still have to maintain the same mode as of the traditional network in a long time so as to fulfill the services’ smooth migration, which is also a challenge for IMS’s
realtime services.

    Secondly, the security problem brought by IP network’s open nature cannot be ignored. The intelligent terminals take up the main part of the IMS network, thus the problems like virus spread, network attack and so on will certainly affect the network’s usefulness. The telecom operators have to invest more to solve the IP security problems.

    Thirdly, the global telecom operators can be generally classified into fixed operators, mobile operators and hybrid operators. Behind this classification is the government’s industry regulation policy. Those who can really benefit from the network convergence are those hybrid operators, but their internal organization structure always separates the fixed from the mobile for operations. Therefore, the converged operators must have internal reorganization. Therefore, these operational mode problems, including industry policy and internal operation and maintenance, also affect the final deployment of IMS.

1 Provision of Realtime Services—IMS’s Replacement of Traditional  Services
Realtime voice services are still the main source of revenue. When providing realtime services, IMS has to solve two problems—guarantee of QoS and utilization of Session Initiation Protocol (SIP)—to fulfill the support for the service flow of the traditional Circuit Switch (CS) domain. If these two problems cannot be solved well, it will be difficult for IMS to completely replace the CS network. Accordingly, it is impossible to reach the goal of network convergence, and not to mention, the benefits of Operating Expenditure (OPEX) saving and fast service deployment of the network after the convergence. IMS is more likely to fall into a marginal service network, or it can only provide either nonreal time services or realtime services not restricted by Service Level Agreement (SLA) among some Internet Service Providers (ISPs). The provision of realtime services is one precondition for the replacement of traditional services, and the other is the adaptation to the access of the traditional nonintelligent terminals.

1.1 Solution of QoS in IP Domain
The nature of the realtime services of IP network lies in its provision of sufficient bandwidth and appropriate time delay for each communication. If it is completely restricted to the endtoend IP application environment, two optional solutions are available.

    The first solution is to preserve endtoend resources, similar to the solution of QoS in Integrated Service (IntServ). Of course, its defects are also similar to those of IntServ, requiring a middle network router to store large amounts of status information, and with relatively higher cost of deployment. As an improved solution, special resource management system can be introduced to implement the management of the whole network’s bandwidth resources, and store the dynamic information of bandwidth usage of each node in the service route. The middle router need not store such large amounts of status, but carry out authorization request for each flow direction to the resource management system, and implement the actions of opening and closing the door according to the authorization result. In addition, the border router, instructed by the resource management system, will mark the flow of the service level, while the middle router will implement packet forwarding according to the labels of Differentiated Service Code Point / Type of Service (DSCP/TOS).

    The second solution is to perform capacity planning and traffic load prediction to guarantee that the network can always provide sufficient bandwidth at high data rate. Within this total bandwidth, the services share the same bandwidth. Based on this framework, the Differentiated Services (DiffServ) can be used to improve the fairness in bandwidth allocation. Because IMS adopts flat networking, the capacity planning has a higher difficulty. Despite of this, telecom operators can generally allocate independent core bearer network for IMS to reduce the difficulty of capacity planning. But IP access network lies on the outside of the IMS core network, and may be operated by those operators different from IMS operators. Thus the QoS problem at this section cannot be effectively solved through capacity planning.

    The first solution requires reforming of the router that IMS services go by. The architecture itself is the same as Network Attachment Subsystem/Resource and Admission Control Subsystem (NASS/RACS) of TISPAN, but introducing bandwidth management in the whole network is a complex project. If the IMS core network becomes an independent network, its deployment difficulty lies only on the side of the access network. For wireless networks, the bottleneck is not in the Border Router, but in the scarcity of wireless bandwidth. To perform effective bandwidth management, it is necessary to manage the actual available bandwidth and the SignaltoNoise Ratio (SNR) of wireless signals of each wireless cell before obtaining the actual available bandwidth at the access side. This enhances the difficulty of QoS solution deployment in wireless network.

1.2 CS Accesses IMS
As it is wellknown, the transmission bandwidth in wireless networks is restricted by the available wireless spectrum bandwidth. The bandwidth is a scarce resource. For wireless Voice over IP (VoIP), besides the basic source encoding, the Realtime Transport Protocol (RTP) message overhead has to be included before the channel encoding at the end. Accordingly, the efficiency of the wireless VoIP is lower than that of the adoption of direct channel encoding. In dense traffic areas, the disadvantage of VoIP is more apparent. It is generally believed that after the large scale deployment of High Speed Uplink Packet Access (HSUPA) in 2009, the efficiency of VoIP transmission in Universal Mobile Telecommunications System (UMTS) will be similar to its transmission in CS. Thus, it still takes some time for the research of 3GPP’s System Architecture Evolution (SAE) to achieve its goal of adopting IMS to completely replace CS, and it is worthy to explore if there is a necessity to adopt IMS to provide realtime services for traditional users in the near future.

    The solution to fixed network has been proposed in the IMS architecture of TISPAN, transforming C5 end office and Softswitch (SS) end office into Access Gateway Control Function (AGCF). The traditional fixed network still accesses through fixed line and Access Gateway (AG), and the core network adopts all IP transmission of IMS. Because the access bandwidth of the mobile network is more limited, it is more reasonable to access IMS through CS, transforming Mobile Switch Center/Mobile Switch Center Server (MSC/MSC Server) into Mobile AGCF (MAGCF), enabling mobile users to have transparent access to IMS network without changing the terminals and all the traffic logics to be executed in the IMS service layer.
When the CS control layer of the mobile network descends to MAGCF, the architectures of fixed and mobile accesses of the circuit domain tend to be unified, thus avoiding the QoS problem of IP access. Meanwhile, those over the access layer of the network completely tend to be unified so that the true convergence can be realized, and telecom operators can also get the real benefits from the deployment of IMS as early as possible.

    We should notice that AGCF of the fixed access mainly completes the conversion of the traditional fixed access signaling H.248 into SIP and their adaptation, which are relatively easier to realize. Under the mobile circuit domain access, the problem becomes more complex. It is necessary to consider the user’s mobility, as well as the compatibility and inheritance for the authorization and authentication method.
Currently after the achievement has been made for the standardization for IMSbased PSTN Emulation Subsystem (PES) and AGCF of TISPAN NGN R1, 3GPP R8 has started the feasibility study on the research of IMS centralized control (the supplementary services of IMS controlling CS). The research will be established by 2006, and by 2007, the requirements will be surveyed. The plan is predicted to be rather complex, and should be given close attention.

1.3 Mapping from CS Service Flow to IMS Flow
In the traditional PSTN, accessed terminals are nonintelligent terminals and are limited by the available keys and the direct signaling between Class 5 switches. Many service flows can only be described by these terminals as simply pressing the hook and then dialing again, and then translated by the switch according to the context into different service operations. The service layer in IMS is isolated from the core network, and the hook pressing the mute terminals must be translated into signaling flow recognizable to the application server. On the other hand, the customers’ using habit is inheritable, and the majority of the activated clients have been used to the provisioning mode of the PSTN services. So even if the IMS network will be accessed through the SIP terminal method, it requires the support for the original service flow.

    PES and PSTN Simulation Subsystem (PSS) both provide services compatible with PTSN. Their main difference lies on the access: the former accesses through nonSIP terminals, the latter through SIP terminals. Providing the same service in IMS network as in PSTN is their common goal. The technology has two key points: one is the correct translation of the traditional terminal behaviors into the SIP signaling at the access side, and the other is the capacity of the IMS Service Control (ISC) interface to completely fulfill the service of CS Network’s INAP/CAMEL/WIN protocol.

    In the architecture of TISPAN NGN, PSTN/SS end office will evolve into AGCF, and will be responsible for accessing CS terminals, and translating the traditional terminals’ special operations like hook pressing into SIP signaling compatible with IMS. After receiving the signals, SCSCF will trigger off relative Application Server (AS) for treatment.

    In the current standard, two treatment methods, including loose coupling and close coupling, are given. The loose coupling maps the hook pressing and the consecutive dialing into a unified PSS service flow according to the service configuration, enabling AS to be truly irrelevant to the service triggering method at the access side. This method makes PSS and PES services converged and unified; therefore, it is more appropriate to the development of FMC.

2 Security—Provision of IMSTraditional Service from the View of Bearer Network
The security problem rooted in IP network will gradually protrude with the deployment of IMS. To make the terminals intelligent and open brings hidden security problems. The security problems exist in two respects: security of terminals, and security of network. The former one refers to the terminals’ own security, such as virus spread and malicious attack. The latter one refers to the security of the serviceproviding equipment. These two are mutually influential: if the terminals are hijacked by the malicious code, it may become a disaster of the whole network.

    The Internet field already has plenty of security products which can carry out protection of nonspecific applications. The purpose of Session Border Controller (SBC) , which is widely applied in SIP networks, is to solve the security problem in SIP networks. Similar to the solution to the security problem in the IMS network, SBC equipment is also introduced to isolate the IMS core network from the access network and any third party network. In this way two safe domains are formed. The core network is relatively isolated, and its safety is easier to be guaranteed in a closed environment, but the access domain’s security remains a problem.

    The traditional telecom service users have been used to 24/7 nonstop services. The security problem brought by IP network will probably reduce the service connecting time, especially some government regulating service (such as emergency call) forced to provide in IP domain will face more severe challenges. On the other hand, because of legal monitoring for government’s control, the IMS’s load cannot have endtoend encryption. All the encryption and security measures can only be segment by segment, or they will end up at the border of the IMS core. To some degree, this reduces the available network’s security solutions. The security of the whole IMS network has to rely on the security equipment provided by the telecom operators whose liabilities and obligations would somewhat increase. Therefore, the cost of replacing traditional service with IMS increases.

    The security problems exist for fixed and mobile networks as well. The wireless network has an urgent requirement for encryption at its access side due to its air transmission characteristic at that side. There are also multiple options for the level that encryption applies:

    (1) Encryption in the data link layer: this is completed in the access network without the knowledge of IMS core network. The defect is that not all the infrastructure at the bottom layer is capable of encryption, so users who access through different media will have different effects.

    (2) Encryption in the IP layer, such as IP Security Protocol (IP Sec) based on Authentication and Key Agreement (AKA) adopted by 3GPP IMS: this can solve the inability of the access layer to support encryption, but it is still limited to the layer of call signaling, having no description on the encryption in the medium layer. Meanwhile, the encryption in the IP layer will have interconnection problems when passing through some equipment over the third layer, such as Network Address Port Translation (NAPT) equipment.

    (3) Encryption in the transport layer, such as Transport Layer Security (TLS) recommended in RFC3261: TLS’s security solution is the same as Hyper Text Transfer Protocol (HTTP) which is widely applied and validated. It can avoid problems which happen when passing through the equipment above Layer 3, such as NAPT, but it still has no regulation on the encryption in the medium layer.
(4) Encryption in the application layer, such as encryption or signature on the body of a SIP message: as above stated, this will lead to certain judicial actions unimplemented. It is accordingly forbidden.

3 The Promotion of Operation Mode to FMC Evolution
In recent years, the revenue of the telecom operation has shown two trends of change. One is the distribution of mobile traffic from fixed traffic, and the other is the distribution of voice service caused by VoIP and other IP application.

    To confront the first trend, fixed operators turn into the hybrid operators, engaging in operation of fixed services. But they are at a loss to the second trend in which the customers pay relatively less, and yet enjoy the traditional communication services that are worth times of cost. In this situation that the customer’s total expense is far lower than the total revenue, the operators have no better solution but to provide products of lower total cost but of higher total value. However, to radically solve the distribution of revenue, the only thing operators can do is to provide customers with products of higher customer value.
Speaking of providing information transmission service alone, the space for additional value increase is limited. Telecom operators have to integrate their products, as well as introduce new key elements of customer value. This is not only restricted in assembling traditional telecom products, but also to integrate other new elements (such as integrating communication and information services, even financing services, tourist and recreational services). To get these elements alone, the customers will have higher cost, but the cost can be greatly reduced through purchasing communication service packages. The cost mentioned includes both money and other costs like physical strength, energy and credit, where telecom operators, as the channel’s end controller, are not only capable of providing, but are also good at.

    This convergence of the first level will certainly have influence on the
operator’s internal organizational structure which is classified by the network type. First of all, for operational entity, the access mode of mobile or fixed network cannot be the foundation for organizational classification. Secondly, the products may be integrated. The majority of services after the convergence can be provided on both networks, as well as share the same service agreement for two types of access media, such as services crossing more than two types of access networks, like coringing and Voice Call Continuity (VCC). Meanwhile, converged network’s influence on operational mode is shown in the levels of service fulfillment as well as operation and maintenance. Although the network is converged in the control layer, the access mode and the terminals’ capability still has crucial influence on concluding service agreement. How to guarantee more flexible and complex product assembly in the level of Operation Support System/Business Support System (OSS/BSS) is the prerequisite for the convergence of operation network.

    The convergence of the second level is the biggest challenge that the operators are facing. It means that the traditional single role of telecom service providers cannot continue, but they have to participate into public life, providing more products demanded by the society, and acting as more active marketers, just like the finance industry does.

    As stated above, FMC will certainly have impact on the existing telecom operation system, and the reform of operation organizational structure will also decide the speed of realizing the convergence. The converged network plus the converged operation organizational structure cannot solve the fundamental challenges in current telecom operation. These challenges are the reduction of the marginal value because of the replacement of services and the distribution of revenue after telecom popularization. These problems need to further integrate communication into daily life.

4 Conclusions
FMC has become the development trend in the communication industry and is a great opportunity for telecom operators to transform and develop. The brand new service experiences brought by FMC will also give users more convenience and pleasure. Although now there are some key issues affecting the large scale commercial deployment, it is believed that after 3 to 5 years of development and maturity, problems like realtime services and network security will be gradually solved, and the operators and the communication industry will also set up their operation mode suitable to themselves. FMC will face a bright future.

References
[1] Wei Leping. Strategic Consideration on Next Generation Network [J]. ZTE Communication, 2004, 10 (1): 14.
[2] Xu Heyuan. Relations Among NGN, Softswitch and IMS [J]. ZTE Communication, 2006, 12 (5): 14.
[3] 3GPP TS 24.228. Signaling Flows for the IP Multimedia Call Control Based on SIP and SDP [S].
[4] TS 183 043 Ver.1.1.1. IMSbased PSTN/ISDN Emulation; Stage 3 Specification [S]. 2006.
[5] TS 182 012 Ver.1.1.1. IMSbased PSTN/ISDN Emulation Subsystem; Functional Architecture,
IMSbased Emulation [S].2006.

Manuscript received: 20061116