CDMA Technology
Members Sign-In
From Global To Universal?

By Alistair Urie, Standardization Director
and Stephane Breyer, Product Manager, Alcatel

The rollout of third generation UTRA (UMTS Terrestrial Radio Access) technology planned for 2002 and beyond, will see the introduction of CDMA-based access techniques and network planning methods into GSM-based networks. UTRA, or WCDMA (Wideband CDMA) as it is sometimes called (referring to one of its component techniques), will also be deployed as a third generation upgrade for some of the Japanese PDC networks, where GSM core network technologies will also be introduced. The Universal Mobile Telecommunications System(UMTS) is a third generation mobile communications system that is being developed as an evolution from the existing GSM system. Handover and service continuity between GSM and UMTS is being ‘designed in’ from the start with dual mode terminals and infrastructure hiding the complexity of the system from the end users.

The UMTS Terrestrial Radio Access (UTRA) air interface is based on the combined use of Japanese WCDMA and European TD-CDMA technologies. One original feature of the UTRA concept is the existence of two different modes: FDD (Frequency Duplex Division) and TDD (Time Duplex Division), each one based on a different, but related, access technology.

UTRA FDD is a ‘pure’ direct sequence (DS) CDMA system where the orthogonality between two different air connections is ensured by codes. It can be broadly compared to existing CDMA systems such as cdmaOne albeit with some differences relating to the implementation of the codes’ management and the need for synchronization. Key design characteristics are the use of asynchronous operation, with optional support for synchronous cells, and the use of time multiplexed downlink pilots.

Design characteristics
UTRA TDD introduces an additional ‘time orthogonality’ component: the variable bit rates are supported by a dynamic and independent allocation of both codes and timeslots (16 rather than the eight of GSM) in both the uplink and downlink. Key design characteristics are the use of burst mid-ambles to assist burst detection and use of time alignment to maintain the TDMA structure. Power control bandwidth is reduced compared to direct sequence schemes, and multi-user detection can be economically implemented. Since this mode can also operate without macro diversity, it opens the potential of unlicensed applications such as corporate applications and Cordless Telephone Systems (UMTS-CTS) for residential markets.

Both of these modes will use a common carrier spacing of 5 MHz and have a common chip rate (the most likely value will be 3.84 Mcps which is slightly down from the original rate of 4.096 Mcps).

FDD is adapted for symmetric traffic and macro-cells while TDD is better suited to asymmetric traffic (such as Internet surfing), for micro-cell and private and other indoor environments.

These two modes offer the opportunity to get the best efficiency with the same system whatever the conditions (wide area, urban, indoor coverage from outdoor, indoor, etc). It is likely that the "mature" public UMTS network will have a radio coverage made of two layers : a macrocell FDD offering bearers up to 384 kbps and a microcell TDD offering bearers up to 2 Mbps. This solution respects the various requirements from the IMT-2000 terrestrial radio interfaces for efficient user bit rates in a range of different environments. UTRA technology obviously uses the technology of soft and softer handover/ handoff, but as well as this, code management allows an easy use of adaptive antennas that introduce a level of ‘spatial’ orthogonality to decrease the level of interference for a given terminal and thus to increase the frequency efficiency. With a voice codec at 8 kbps, it is commonly agreed that an FDD carrier will support up to 100 users with an activity factor of 50 per cent.

What then are the network aspects of UMTS? The UMTS core network will be a direct evolution from GSM with the use of MAP mobility and CAMEL (Customised Applications for Mobile services Enhanced Logic) and MEXE (Mobile station Execution Environment) service protocols. In addition, SIM cards will be in the terminals and an architecture based on radio access element BTS and BSC/RNC (Base Station Controller/Radio Network controller), the circuit switching element MSC/VLR (Mobile Switching Center/Visitor Location Register), databases and service platforms such as HLR (Home Location Register), SMS-C (Short message Service Center) and SCP (Service Control Point). Finally an Internet-based packet data backbone with GSM-specific Serving and Gateway GPRS Support nodes (GSNs).

At a service architecture level UMTS will re-use all existing GSM services and will implement the Virtual Home Environment (VHE) concept using elements from CAMEL, MEXE and SIM toolkit.

As for deployment and roll-out, migration from GSM towards UMTS will follow a two-step process.

In the first step, 2G/3G core network equipment is deployed and the third generation radio equipment is installed as an overlay network. Roaming and handover/handoff between GSM and UMTS is part of the first commercial standard and will be realized using the MAP and GPRS protocols. Multi-standard terminals will be able to access both the national coverage of the 2G and the services of the 3G in "hot-spots". Coverage takes time and costs money; this solution allows the deployment UMTS only as and when it becomes necessary.

As a second step operators will slowly integrate more and more of their 2G coverage into combined 2G/3G equipment. The main drive will be to minimize operating cost once the initial cost and risk of 3G deployment has been achieved.

Of course, all this activity will have little meaning without defined standards. Until the end of 1998 standards for UMTS, and the related WCDMA and WP-CDMA radio proposals, were being progressed in parallel in the European Telecommunications Institute Special Mobile Group (ETSI SMG), Japan’s ARIB and TTC committees, Korea’s TTA and the North American T1P1 committee (with TIA TR46). In addition the ETSI SMG was working on both GSM phase 2+, covering features such as CAMEL, GPRS, SIM toolkit, AMR (Adaptive Multi Rate) and CTS, and the evolution of the GSM core network (and hence GSM MAP) for third generation needs.

The work on UTRA and the GSM-based UMTS core network is now concentrated in the newly launched Third Generation Partnership Project, or 3GPP. This body brings together the founding members ETSI, ARIB and TTC, along with North America’s T1 committee and Korea’s TTA to develop common technical specifications for third generation radio and networks. These member organizations are committed to individually publish standards based on the agreed 3GPP outputs covering terminals, radio access and core network aspects. 3GPP has four Technical Specification Groups (TSGs), open to participants from the member organizations, on Service and System Aspects (TSG-SA), Radio Access Network (TSG-R), Core network (TSG-N) and Terminals (TSG-T) and a Project Co-ordination Group (PCG) with representatives from the member organizations and market representative partners.

It is worth noting that a parallel organization, known as 3GPP2 , has been created to globalize the development of specifications for third generation mobile systems based on CDMA2000/IS-41 networks. This body, like 3GPP, has member organizations from TIA, ARIB and TTC and TTA. One interesting aspect of 3GPP2 is that they have established a dedicated TSG to develop specifications for the interworking of UTRA with an IS-41-based core network.

At a global level recommendations for third generation mobile systems are also being developed in the ITU-R TG8/1 group (for radio aspects) and ITU-T SG11 (for network aspects). TG8/1 is worth particular attention following the successful agreement in March 1999 on a common document covering recommendations on key characteristics of the IMT-2000 radio interface. This document (currently known as ‘IMT.RKEY’) unites the UTRA and CDMA2000 proposals for CDMA interfaces and the UWC136 proposal for a TDMA-based solution into a single multi-mode concept.

A timetable for the introduction of UMTS is already taking shape. UMTS systems will be deployed first in Japan where they will be introduced during 2001. The European Community has established a general rule for the opening of UMTS commercial operations on 1 January 2002 with field testing likely to begin one year earlier (the assumption is that a standard for a ‘commercial’ system will be available by the end of 1999). The process of allocation is to be defined during 1999 with the UMTS Forum recommending an allocation of up to 2x15 MHz paired and 5 MHz unpaired blocks of spectrum per operator, thus allowing up to four operators (these figures are typical ones and depend on local regulators).

Commercial deployment
The timing for licences is very different from one country to another. If you wanted a UMTS license in Finland for instance, you’re too late: the four licenses have already been allocated. The United Kingdom, Holland and Germany are expected to be the 1999 "allocators". Some countries have been or will be taking the "beauty contest" option (Finland) and others the "auction" option (UK) as happened in the US when PCS was introduced. France is following the EU schedule with rule definition during the next fall. Other European countries have already indicated their intention of allocating IMT-2000 spectrum to UMTS broadly following the EU timeframe. Asia, Africa and the Latin American countries are likely to then follow over the next five years with the timing decided by both market needs and the availability of the IMT-2000 band at 2 GHz. The first UMTS equipment will be installed during 2001 in Japan where it will be commercially deployed and at the same time in Europe where it will be deployed for test purposes. Commercial deployment both in Europe and elsewhere in countries that have chosen the UMTS concept will begin the year after.

After a slow start, involving operator trials, market tests and early adopters, it is expected that the market will take off after 2005 when the bulk of 2G users will start to migrate to the 3G systems. However, what activity takes place in re-farming of GSM 900 and 1800 frequencies into UMTS remains to be seen.

Standards and specification work is now running at a fast pace with many wide-ranging debates in 3GPP on fundamental aspects of the key network interfaces and architectures, the detailed stage one service descriptions and the need to seek agreements on dull but essential issues such as document styles, numbering and procedures. Work is currently targeted at completing a first set of specifications at the end of the year. These will be called "Release 99", following the GSM tradition. Specifications for UMTS are, however, still far from complete and so it is too early to be sure of what the "fine details" are. A number of current debates will affect this outcome, notably the radio harmonization process.

Although we refer to CDMA2000 and UTRA as two different systems, a number of important global operators have declared that they want a single global standard. Work towards this goal started last year in the ARIB ad hoc S committee and in various other fora. The latest step towards this goal has been the agreement in TG8/1 (outlined above) in which the multi-carrier (MC) component of the CDMA2000 proposal and the UTRA FDD and TDD technologies is referred to collectively in a common CDMA concept with FDD/MC, FDD/DS and TDD modes respectively. These three modes are still subject to change as the harmonization process continues over the rest of 1999 and at least some features of the Chinese TD-SCDMA proposal are integrated.

What is not yet clear is to what degree this harmonization process between the two CDMA camps will result in a genuine common standard as opposed to an ITU-R recommendation which could be said to blur the remaining differences. One possible outcome could see the eventual fusion of the 3GPP and 3GPP2 groups into a common forum but this will depend in the resolution of the "two network problem" as well. The so-called two network problem is the due to the historical split between the AMPS world, with its daughter standard IS-95 (cdmaOne) and IS-136 (TDMA) digital cellular systems and the common IS-41 (or as it is now called ANSI-41) and WIN core network, and the GSM world with MAP CAMEL and GPRS. For many operators and manufacturers the differences between these core networks is actually more significant and more important than last year’s CDMA wars. Many people have thus called for a longer term goal of merging these two network types together or at least opening a fully featured gateway between them.

Interworking and roaming between MAP and IS-41 networks has been demonstrated for simple voice calls. However, a number of unresolved problems arise due to fundamental differences in the way IN services are implemented (CAMEL for MAP and WIN for IS-41) and the different techniques used for packet data and Internet services. Work on this issue is currently taking place in joint TR45.2/TR45.3/T1P1 and TR46 meetings, with results due by mid 2000, and no doubt groups like 3GPP and 3GPP2 will soon need to start work on it as well.

The issue is even more complex when we consider the fact that the TDMA industry association UWCC is now working on the application of GSM’s GPRS core network to the TDMA /IS-136 systems and that ETSI has adopted the same radio technology, EDGE, as a 2G GSM enhancement. While this will open the route to GSM-TDMA roaming, at least for Internet services, it will actually result in what can be called "two and half different networks".

There is another "big issue" that needs to be resolved—IPR and the associated problem of limiting royalty payments. Fortunately, the simple blocking issue of IPR availability seems now to be resolved with both Ericsson and Qualcomm agreeing to respect ETSI, ITU and ARIB rules on essential patents. There remains, however, the far more complex task of reaching a multi-party agreement on the achievement of a low-cost IPR solution, although, after the agreement reached in January this year within the UMTS IP Association (UIPA) to define a common scheme comprising the patent platform supported by a PEE (patent evaluation entity) , it seems some progress has already been made.

The task before us now is to find a way to limit the effect of remaining differences on our end users so that the original goal of IMT2000 of "anytime, anywhere" can be achieved. The adoption of a CDMA-based UTRA is a step along this path since it opens the door to multi-mode UTRA/CDMA2000 terminals. But is there a market for such devices—and will operators take advantage of such market? That remains to be seen.