【正文】 e as their 2G technology. At time of writing,Japan is the most advanced in terms of 3G network deployment. The three incumbent operators there have implemented three different technologies: JPhone is using UMTS,KDDI has a CDMA2000 network, and the largest operator NTT Do Co Mo is using a system branded as FOMA (Freedom of Multimedia Access). FOMA is based on the original UMTS proposal, prior to its harmonization and standardization. The UMTS standard is specified as a migration from the second generation GSM standard to UMTS via the General Packet Radio System (GPRS) and Enhanced Data for Global Evolution (EDGE), as shown in Figure. This is a sound rationale since as of April 2003, there were over 847 Million GSM subscribers worldwide1, accounting for 68% of the global cellular subscriber figures. The emphasis is on keeping as much of the GSM network as possible to operate with the new system. We are now well on the road towards Third Generation (3G), where the network will support all traffic types: voice, video and data, and we should see an eventual explosion in the services available on the mobile device. The driving technology for this is the IP protocol. Many cellular operators are now at a position referred to as , with the deployment of GPRS, which introduces an IP backbone into the mobile core diagram below, Figure 2, shows an overview of the key ponents in a GPRS network, and how it fits into the existing GSM infrastructure. The interface between the SGSN and GGSN is known as the G interface and uses the GPRS tunneling protocol (GTP, discussed later). The primary reason for the introduction of this infrastructure is to offer connections to external packet networks, such as the Internet or a corporate Intranet. This brings the IP protocol into the network as a transport between the SGSN and GGSN. This allows data services such as or web browsing on the mobile device,with users being charged based on volume of data rather than time connected. The dominant standard for delivery of 3G networks and services is the Universal Mobile Telemunications System, or UMTS. The first deployment of UMTS is the Release ’99 architecture.In this network, the major change is in the radio access network (RAN) with the introduction of CDMA technology for the air interface, and ATM as a transport in the transmission part. These changes have been introduced principally to support the transport of voice, video and data services on the same network. The core network remains relatively unchanged, with primarily software upgrades. However, the IP protocol pushes further into the network with the RNC now municating with the 3G SGSN using IP. The next evolution step is the Release 4 architecture, Figure 4. Here, the GSM core is replaced with an IP network infrastructure based around Voice over IP technology.The MSC evolves into two separate ponents: a Media Gateway (MGW) and an MSC Server (MSS). This essentially breaks apart the roles of connection and connection control. An MSS can handle multiple MGW, making the network more scale able. Since there are now a number of IP clouds in the 3G network, it makes sense to merge these together into one IP or IP/ATM backbone (it is likely both options will be available to operators.) This extends IP right across the whole network, all the way to the is referred to as the AllIP network, or the Release 5 architecture, as shown in Figure 5. The HLR/VLR/EIR are generalised and referred to as the HLR Subsystem(HSS).Now the last remnants of traditional telemunications switching are removed, leaving a network operating pletely on the IP protocol, and generalised for the transport of many service types. Realtime services are supported through the introduction of a new network domain, the IP Multimedia Subsystem (IMS). Currently the 3GPP are working on Release 6, which purports to cover all aspects not addressed in frozen releases. Some call UMTS Release 6 4G and it includes such issues as interworking of hot spot radio access technologies such as wireless LAN.Like any CDMA system, UMTS needs a wide frequency band in which to operate to effectively spread signals. The defining characteristic of the system is the chip rate, where a chip is the width of one symbol of the CDMA code. UMTS uses a chip rate of , the UMTS air interface is termed ‘wideband’ CDMA.There are actually two radio technologies under the UMTS umbrella: UMTS FDD and TDD. FDD stands for Frequency Division Duplex, and like GSM, separates traffic in the uplink and downlink by placing them at different frequency channels. Therefore an operator must have a pair of frequencies allocated to allow them to run a network, hence the term ‘paired spectrum’. TDD or Time Division Duplex requires only one frequency channel, and uplink and downlink traffic are separated by sending them at different times. The ITUT spectrum usage, as shown in Figure 6, for FDD is 1920980MHz for uplink traffic, and 21102170MHz for downlink. The minimum allocation an operator needs is two paired 5MHz channels, one for uplink and one for downlink, at a separation of 190MHz. However, to provide prehensive coverage and services, it is remended that an operator be given three channels. Considering the spectrum allocation, there are 12 paired channels available, and many countries have now pleted the licencing process for this spectrum, allocating between two and four channels per licence. This has tended to work out a costly process for operators, since the regulatory authorities in some countries, notably in Europe, have auctioned these licences to the highest bidder. This has resulted in spectrum fees as high as tens of billions of dollars in some countries. The Time Division Duplex (TDD) system, which needs only one 5MHz band in which to operate,often referred to as unpaired spectrum. The differences between UMTS FDD and TDD are only evident at the lower layers, particularly on the radio interface. At hig