Strategic Thinking on Next Generation Network

1 Driving Forces of Developing Next Generation Network (NGN)
    (1) Revolutionary Technologies
    As fundamental technologies, microelectronics technologies will continue to develop by Moore’s Law in the next 10-15 years. The increase rate of optical transmission capacity has already slowed down from doubling every 9 months a few years ago to doubling every 14 months currently. But the development of optical transport technologies is still beyond Moore’s Law, and estimated to last at least for 5-10 years ahead. The number of mobile phone subscribers in the world is going to exceed that of wired subscribers. The rapid expansion of IP application and basic maturation of IPv6 technology are bringing IP technology into a new era. In short, breakthrough in revolutionary technologies has laid a solid foundation for the advent of NGN.

    (2) Fundamental Change of Traffic Composition 
    Over the past 100 years, telecom networks have been dominated by telephony traffic, and traditional networks based on circuit switching were basically competent for supporting the service. However, in the recent years,
IP-dominated data services are developing at a very fast speed and have broken the service pattern. Traffic of IP services has dominated telecom networks. Therefore, new next generation network architecture is needed to effectively support such burst-oriented data services.

    (3) Trend of Network Convergence 
    With the maturation of technological conditions, network convergence, especially at network edges, is becoming a major trend of telecom development. From the convergence of voice and data to the convergence of the wireless and the wired, and from the convergence of transport networks and various service networks to the convergence of "three networks" in the end are the inevitable trends of NGN development.

    (4) Market Opening and Market Competition
    Changes of market demands, market competition and regulation policies are driving the information industry into a comprehensive competition era. In order to reduce costs, develop services and provide subscribers with convenience, the partial and even whole convergences of multiple independent service networks have become unavoidable. Also the final convergences of services, terminals and even company’s organization frameworks are unavoidable. Therefore, a trend towards converged next generation networks will be inevitable.

2 Concepts and Characteristics of NGN
NGN generally refers to an IP-based converged or partially converged full-service network that simultaneously supports voice, data and multimedia services. On one hand, NGN is not a simple extension and superposition of existing telecom networks and IP networks, nor a network with a single new node or network technology. It is a total solution and a change of entire network architecture. On the other hand, NGN’s advent and growth are not a revolution, but an evolution.

    ITU-T sums up the main characteristics of NGN: it is a packet based network, and able to provide all services, including telecom services; it can make use of transport technologies with multiple bandwidths and QoS guarantee; its service-related functions are independent of its transport technologies; it allows subscribers to freely access different service providers; it supports the general mobility and allows to provide subscribers with constant and universally existing services. In short, the main goal of NGN is to simplify the convergence of networks and services.

    It can be seen that NGN involves very extensive contents, and is absolutely not limited to the Softswitch system only. From the viewpoint of a network, NGN involves all network layers from the backbone network, metropolitan area network, access network, customer premises network to various service networks. When the service network layer is concerned, NGN refers to the next generation service network. When the access network layer is concerned, NGN refers to various broadband access networks. As for the transport network layer, NGN often refers to next generation intelligent optical transport network. In a word, NGN in a general sense actually includes almost all new generation network technologies. The following 5 major strategic directions are the key to NGN development.

3 Evolution to NGN

3.1 Evolution to Softswitch-Based Next Generation Switching Network
The Softswitch system breaks the closed switching structure of the traditional circuit switching system. It uses a completely different horizontal combination mode, opens interfaces among various functions of switches and uses open interface and general protocols, forming an open and distributed system architecture for multi-manufacturer application. Softswitch hardware is decentralized, while service control and service logic are relatively concentrated, which allows service providers to select the optimal and most economical equipment combination to construct their networks, with low networking costs and easy network upgrade. Besides, it is easy to speed up the development, implementation and deployment of new services and applications, and helps quickly realize wide area service coverage with a low cost and promote voice and data convergence.

    Softswitch has the following 5 advantages. First, Softswitch fulfills a distributed communication and management by adopting an open architecture. Therefore, it has good structure scalability. Its application layer and media control layer have been separated from the hardware of the media layer and incorporated into an open and standard computing environment, thus allowing full use of commercial standard computing platforms, operating systems and development environments. Second, Softswitch implements the convergence of multi-service networks, and simplifies network hierarchy and structure and service provision across different networks such as the circuit switching network, packet network, fixed network and mobile network. High costs of constructing and maintaining multiple separated service networks and complication of upgrade are avoided. Third, by use of packet switching technology, utilization efficiency of network resources is greatly increased. The complexity caused by a lot trunks of mesh connection between switches and the bearing cost of service networks are reduced. Fourth, since Softswitch pricing can follow the software license mode, its investment would increase along with the growth of subscribers, which helps new telecom carriers and traditional carriers exploit new markets. Introduction of Softswitch also enables carriers to use IP networks of other carriers to quickly enter others’ markets without any limitation by account settlements. Last, Softswitch equipment takes up small footprint, which not only obviously improves room utilization, but also facilitates flexible implementation of nodes.

    The main shortcomings of using a Softswitch system are that the Softswitch technology is not matured enough, and that large-scale field application experiences are lacked, especially in multi-manufacturer interoperation, QoS guarantee of real-time services, unified and effective management of networks, service development and service profit-making capabilities.

3.2 Evolution to Next Generation Mobile Communication Network Represented by 3G Systems
Since 2003, the service revenue, number of subscribers and total length of local calls of Chinese mobile carriers has exceeded those of fixed network carriers. China is the first large country that has realized such a change in the world. In order to exploit new frequency spectrum resources, to realize global unified frequency bands, unified systems and seamless roaming to the largest extent, to meet market demands for middle and high speed data and multimedia services, to further improve frequency spectrum efficiency, to increase network capacity and reduce costs, 3G systems are inevitable development trends of mobile communications.

    Both Wideband Code Division Multiple Access (WCDMA) and CDMA2000, as two Frequency Division Duplex (FDD) systems of 3G, have shown a good development state since 2003. The number of global CDMA2000 subscribers had exceeded 112 million by the first half of 2004. Among them, the number of EV-DO subscribers reached 7.5 million. WCDMA systems also spring up and give chase. Their subscribers have exceeded 7 million, and 40 WCDMA networks have been officially put into operation in the world. Some WCDMA carriers have started to implement the R4 version with the Softswitch structure concept on their core networks. So far, the system hardware of WCDMA and CDMA2000 has worked stably, their software has been continuously upgrading, and there are more and more dual-mode cellular phones. It can be thought that the two systems have been basically matured. There are no significant differences between the two in technology and service capability. Their technical parameters and performances are relatively close except the difference in core network signaling, chip rate, base station synchronization and pilot frequency structure. They are all square in voice capacity, data capacity, and coverage, and have similar economic performance. CDMA2000 is currently leading in the market, but WCDMA is expected to become the dominant application system gradually in the future since it has been supported by more equipment manufacturers, chip developers and service application developers and has more powerful global roaming capability.
The development of TD-SCDMA, a Time Division Duplex (TDD) system of 3G, obviously lags behind that of WCDMA and CDMA2000 systems. The prime reason is its absolutely unfavorable position in fund and personnel investment for the research and development, which is caused by its standards failing to obtain extensive global supports. Carriers also pay great attention to the problems of high costs, strong interference and limited international roaming when TD-SCDMA is used to independently form a large network. However, the system not only combines time division, code division and space division technologies, but also adopts such new technologies as smart antenna, joint detection and uplink synchronization. Therefore, it has natural advantages in frequency spectrum efficiency and frequency spectrum flexibility. The core network and high-layer protocols of the wireless network of the TD-SCDMA system are exactly the same as those of the WCDMA system, so it can realize mutual complementation of advantages, hybrid networking with the WCDMA system. Binding with the WCDMA system, it mainly covers hot-spot areas and focuses on data services.

    In addition to technical factors, the development of 3G depends to a great extent on services, service implementation and the architecture of services. In order to adapt to the development of data services, requirements of new industry chains and service modes and to speed new service development, development of an integrated service platform with an open and horizontal structure is the key to 3G service growth. And the implementation of unified provision, unified accounting and unified security management is most important. As for service development, it cannot be neglected that voice and narrowband services are still the main revenue sources for mobile carriers in a quite long period of time. Various content services are just complementary services with continuous growth.
Along with 3G commercialization, Beyond 3G (B3G) or 4G technologies with higher rate, higher frequency spectrum efficiency, better coverage and more powerful service support capability have also been under research. It is planned to complete their frequency spectrum allocation in 2007, to complete main standards in 2010, and to put
B3G/4G systems into large-scale commercial use after 2015. The 4G system aims at obviously better system functionality and performance when compared to 3G. In the respect of functionality, the B3G/4G system plans to introduce new service platforms to implement seamless handover between different access means; In the respect of performance, its target transmission rate is maximal 100 Mb/s and average 20 Mb/s. The cost per bit will be hopefully reduced to 10% and even to 1% of that of the 3G system. It could realize the control of various mobile characteristics and data packet transmission with different QoS.

3.3 Evolution to IPv6-Based Next Generation Internet
Moreover, IPv4 is increasingly not able to meet the requirements of future development due to its intrinsic shortcomings in application limitation, quality of service, flexibility of management and security. That the Internet gradually shifts to the IPv6 based Next Generation Internet (NGI) is almost an inevitable major trend. 
IPv6 fundamentally solves problems of the address limit and tremendous routing tables existing in IPv4, and supports more effective mobile IP.

    First, IPv6 expands address length from 32 bits of IPv4 to 128 bits, providing almost unlimited public addresses and completely eliminating the address barrier in Internet development.

    Second, IPv6 protocol includes the mobile IPv6 part, which enables mobile terminals to realize free moving between different access media without changing their own IP addresses. Also routing optimization between any two terminals in the world can be implemented.

    Third, IPv6 simplifies management and maintenance of network nodes by a series of automatic discovery and automatic configuration functionalities. Plug and play can be fulfilled, which helps to support mobile nodes and applications of a lot of small home appliances and telecom equipment.

    Fourth, by use of IPv6, a large number of new applications can be developed, such as P2P services (for example, online chat and online game), 3G services, and home networks.

    Fifth, IPv6 implements priority by using stream classes and stream labels, which guarantees good network QoS.
Sixth, Its built-in IPSec and sending equipment own permanent addresses, which can not only implement
end-to-end encryption, but also easily identify the type of information-sending equipment, realizing real end-to-end security.

    Seventh, the hierarchic structure is used for the addressing of Ipv6. Thus, the routing efficiency is obviously enhanced and the number of routers is greatly reduced.

    Last, IPv6 defines the multicast functionality, simplifying the offering of stream media services. In short, IPv6 will become a service layer convergence protocol for evolution to NGN.

    IPv6-related technical standards have been basically formed, but actual IPv6 network deployment is very slow. The main reason is that IPv4 may cope with the address demand in about 5 years by using such measures as Network Address Translation (NAT). Besides, the IP address mode has a close relation with upper layer protocols and the network operating modes. The implementation of IPv6 needs not only to upgrade network layer protocols, but also to upgrade application software and to modify packet forward modules of routers, which almost involves all equipment of networks.

3.4 Evolution to Diversified Broadband Access Network
Recently, access network broadbandization goes very quickly at home and abroad. However, the access network is sensitive to its costs, regulation, services and technologies. So far, there is no recognized and absolutely dominating broadband access technologies yet. Although ADSL, HFC and Ethernet will globally keep being the 3 dominant broadband access technologies in the near future, ADSL applications has exceeded those of HFC and become the leading broadband access technology. Moreover, various new technologies are still continuously coming forth. Therefore, in a quite long period of time, diversified access technologies will co-exist and complement each other, competing and developing. Several new promising broadband access technologies are described below. 

    EoVDSL is a VDSL based on the Ethernet technology. First, EoVDSL integrates the characteristics of layer-2 Ethernet and VDSL physical layer. It has a good cost performance and high downlink rate (over 100 Mb/s). Its symmetrical transmission rate can reach 26 Mb/s. Second, EoVDSL can implement long distance transmission over the existing twisted pair, saving new Class 5 cable laying. Third, the transmission distance of EoVDSL is farther than that of Ethernet, which helps improve actual take-up rate of subscribers. Access equipment can be centralized to reduce the maintenance cost. Finally, EoVDSL has relatively low crosstalk and higher available twist pair due to its relatively low power density and reasonable arrangement of frequency spectrum, so it fits applications with dense users. The main shortcoming of EoVDSL is that its use of Ethernet at layer 2 asks for properly dealing with all basic problems that Ethernet has, such as manageability, security and QoS. Moreover, the longtime dispute about standards of the two line codes, DMT and QAM, affects its development. It is predicted that EoVDSL technology as a complementation to ADSL, will find wide applications after the recent settlement of its global standard. 

    Based on IEEE 802.11 protocols, Wireless Local Area Network (WLAN) actually is a wireless Ethernet. It supports relatively high rate (up to 2-11 Mb/s, and even to 54 Mb/s) and the networking is simple. Therefore, commercial users favor it. In order to apply the technology to the access network field, it is necessary to properly solve problems such as authentication and accounting, user management, user roaming, user and network security, user handover, equipment and network management and user access control. The core problem is its business model. Should it be used as a binding value-added service of the wired access to provide a set service? Or used as an independent new service to generate the cash flow? There is no answer to it yet. The former is more likely realizable.

    From a long-term viewpoint, the optical access network, especially the passive optical network is likely a relatively ideal solution. A new generation passive architecture—the GPON standard recently defined by ITU increases upstream and downstream rate up to 2.5 Gb/s, and uses the Generic Framing Procedure (GFP) to more effectively support various data services, making the passive optical network more attractive.

    The key problem of the optical access technology is that its comprehensive cost is very high while market demands for the transmission rate is not great enough. Therefore its development is limited at a relatively low speed. However, recent technology advancement, especially the advent and development of low cost Vertical Cavity Surface Emitting Laser (VCSEL), provides new driving forces for the development of optical access technologies. However, aiming at a mainstream access technology, it still needs to solve a series of problems in addition to equipment costs, including networking, splicing and connecting technology, test and laying and installation technologies. 

    From the viewpoint of network operation, it is very complicated and expensive to provide the long-term support and maintenance for different equipment in the same network. Therefore, with diversified access technologies, the establishment of a common access platform with a modularized architecture should be a development trend. That can simplify the network structure, reduce duplicate elements and components, decrease access costs and protect investment, accelerate service offering, and save costs for long-time evolution and technology replacement of networks. Specific implementation can be done by using public subscriber line cards, public open network interfaces, network management interfaces and other public subsystems, integrating various broadband and narrowband access technologies and providing various broadband and narrowband services.

3.5 Evolution to Next Generation Transport Based on Optical Internetworking
With major breakthroughs in technology and market driving, Wave Division Multiplex (WDM) systems develop at a very fast speed. Presently, the 1.6 Tb/s WDM system has been widely used for commercial purposes. However, although relying on WDM technology, a breakthrough in transmission link capacity has basically realized, common end-to-end WDM systems only provide original transmission bandwidth and need flexible nodes to implement efficient and flexible interworking capability. The existing Digital Cross Connection (DXC) system is very complicated. Its node capacity doubles about every 2-3 years, and cannot keep up with the growth of network transmission link capacity. Now people turn to optical nodes with the hope to further extend the capacity, i.e., Optical Add-Drop Multiplexer (OADM) and Optical Cross-Connect (OXC).

    As network traffic is convergent to dynamic IP traffic, a flexible and dynamic optical network infrastructure is indispensable. The up-to-date development trend is to introduce automatic wavelength configuration function, so called Automatic Switching Optical Network (ASON), enabling optical internetworking to shift from a static one to a dynamic one. ASON brings great advantages: network resources are allowed to be allocated dynamically to routers to shorten upgrade and capacity expansion time at the service layer; service offering and expansion are speeded up; maintenance and management costs are reduced; quick service restoration capability at the optical layer is available; operational support system software used in new technology configuration is reduced; man-made errors are reduced; and new wavelength services are introduced, such as the bandwidth-on-demand service, wavelength lease, bandwidth service with multiple classes, dynamic wavelength allocation lease service and Optical Virtual Private Network (OVPN).

    Of course, implementation of optical internetworking still requires solutions to a series of problems in hardware, software and standardization. But it has a bright future. Intelligent optical networks will become an important development direction and market opportunity of optical communication in a few years.

    Transition to the automatic optical switching network has two basic evolution structures, i.e., the overlay model and peer model. The overlay model, also called client-server mode, is a network evolution structure supported by international standardization organizations and quasi-standardization organizations such as ITU, Optical Interconnection Forum and IETF. It is also a model that most traditional carriers favor. Its basic thinking is to put control intelligence defined by the optical transport layer on the optical transport layer only, without interference from the client layer. Its greatest benefit is that a unified and transparent optical transport layer platform can be implemented, supporting multiple client layer signals with no limits to IP routers. Besides, requirements from the client layer are transferred to the optical service layer through interfaces, and client’s connection requirements are completed at the optical network layer. In such a way, network topologic details of the optical transport layer can be shielded. The model also allows the optical transport layer and client layer to evolve independently. Moreover, by use of sub-network division, carriers can not only make full use of their existing infrastructures, but also introduce new technologies into other parts of the networks without any limitation by existing infrastructures. In addition, as the model can use mature and standardized User Network Interfaces (UNI) and Network Node Interfaces (NNI), it is relatively easy to fulfill interoperation of optical networks in the near future and quickly implement commercial network deployment.

4 China Telecom’s Strategic Thinking on NGN
China Telecom is the largest fixed network carrier in the world. As a traditional telecom carrier, its main local fixed telephone service is undergoing a quick market losing with the growth of mobile, PHS and IP services. Under the situation, China Telecom urgently needs to search for new strategic approaches to reduce network costs, increase service revenue and develop new services. The advent and growth of NGN provides it an important opportunity in time. Therefore, China Telecom started to develop its ADSL broadband access network in a large scale in the second half year of 2001, and launched pilot projects of Softswitch-based next generation switching network in 4 cities in 2002. At present, the projects have gone into the commercial trial stage after technology and service tests had finished. This year, China Telecom has started constructing a nationwide IP/MPLS network with carrier-class service quality and IPv6 capability, as a bearer network to support the convergence of new generation services.

    Obviously, China Telecom’s understanding of NGN is not a simple use of the Softswitch system to upgrade its switching network but a long-term and all-around strategic understanding. China Telecom makes use of Softswitch as an opportunity to implement integrated technology transformation. NGN must have an open and distributed framework in architecture and sustainable development ability in service. It must be able to cover from the core to the edge, from the wired to the wireless and from the service network to the transport network. It must offer
multi-vendor environment in application environment and interoperation. In short, NGN is not a partial technical modification and renovation of existing networks, but an end-to-end converged total solution with a gradual evolution. This will be the main strategic transformation task of China telecom in the next 10-15 years.

Manuscript received: 2004-09-03