August 2000 European Institute for Research and Strategic Studies in Telecommunications GmbH Deliverable 2 Guidelines for the design of UMTS Access Networks Project P921-PF
Project P921 Services, applications and Goals Quality of Service The main target of this work was to UMTS is going to support a variety of services obtained by means of a link level simulator to develop recommendations and guide- and applications, using both circuit and pack- corrupt the application bit stream and to lines for UMTS network design and et switched access. In the framework of evaluate the degradation of the quality due to implementation. These guidelines and EURESCOM project P921 three kinds of the radio interface. recommendations are to support plan- applications have been selected for Quality of ners and operators in designing and Service analysis: audio retrieval, MPEG-4 video The results of the tests have shown a strong implementing efficient UMTS networks. download applications, and IP-based appli- impact of the UTRA interface on the Quality of Application testing Link Level Subjective Application Simulator Testing Error Application patterns performance UTRA character- QoS isation cations (web browsing, ftp). The objective of Service. For example, real time streaming the quality test was to assess the impact of of high quality music over UMTS requires a the UTRA (UMTS Terrestrial Radio Access) highly protected channel, at least when the interface on the selected applications. The test application is not using any error resilience method applied was to use the error patterns tools. Cell coverage Cell breathing: left lower, right higher cell traffic in UMTS One of the fundamental characteristics of CDMA systems implemented in UMTS is that the coverage range is intrinsically linked to the capacity of the system: the more traffic is carried by a cell, the smaller the coverage area of the cell becomes. This phenomenon is known as “cell breathing“, which shows the service area of one base station with different traffic loads in the system. This dynamic behaviour makes cell planning and network dimensioning a very complex process. Traditional static prediction methods are not Link level, considering the effects of the radio appropriate. Therefore simulation and statis- channel on individual bits transmitted in a tical modelling techniques have to be used. single communication. However, the system is very complex, with so System level, considering a number of cells many interactions, that the simulation has and mobiles, based on output parameters been split into two parts: from individual link simulations produced at link level.
UTRAN characterisationUTRAN, the UMTS Terrestrial Radio Access Resulting system working point for the 8 kbit/s voice serviceNetwork, operates in two modes, the UTRA as a function of mobile speed and number of users per slotFDD and the UTRA TDD mode. The UTRAN (downlink – Vehicular A channel)link level simulation results of P921 are givenfor the voice service, circuit switched data System working point [Eb/No @ BER = 0,1 %]service (LCD, Long Constrained Delay) and for 15packet switched data service (UDD Uncon- 8 users per slotstrained Delay Data) over the ETSI / ITU 14propagation channels (vehicular A/B, outdoor 13 4 usersto indoor A/B). The simulations include 12 per slotrealistic algorithms for closed loop power 11 1 usercontrol and pilot assisted channel estimation. per slot 10For the up-link channel, the antenna diversity 9technique has been implemented by doublingthe Rake receiver structure and using an equal 8gain combiner before decoding. Voice service 7was simulated at 8 kbits/s, and LCD and UDD 1 10 100 1000services at 64, 144 and 384 kbits/s. Mobile speed [km/h]Link budget Link-level simulation resultsand cell sizes Cell radius [km]The link budget is calculated by the following 2,0 UMTS speechprocedure:1. Uplink path loss evaluation 1,5 GSM 18002. Downlink power level evaluation at cell border 13. Downlink EIRP value evaluation per traffic UMTS LCD384 (Long Constrained channel (Effective Isotropic Radiated Power Delay 384kb/s) – i.e. how much power you would be trans- 0,5 mitting if transmitting in a perfect sphere) UMTS UDD480 (Unconstrained4. Downlink power evaluation per traffic 0 Delay Data 480kb/s) channel 21 22 23 24 25 26 27 28 29 305. Downlink path loss evaluation Average power of the mobile station [dBm]On the basis of the radio link results the UMTS(FDD component based on W-CDMA access)link budget has been evaluated for the case of a function of average power of the mobile sta- cell radius is greater than the one of GSMan urban environment. The cell radius of the tion and offered service (70 % cell load). The 1800. In contrast to this, the coverage in UMTSUMTS system has been compared with the UMTS cell radius is compared to the cell radius is smaller than in GSM 1800 systems for theone of GSM 1800. The figure presents results of a GSM 1800 system. It is worth noting that other services.from link level simulations: The coverage range in the GSM 1800 case the cell radius is notof UMTS services in the urban environment as related to the system load. The results show that, in the case of a voice service, the UMTSUTRAN architectureThe UMTS Radio Access Network is built RNC is connected to the Core Network (both that cells and RNCs are identified – normallyaround two new nodes and three new inter- packet and circuit domains) by the Iu inter- by the number of bits in the identities, butfaces (see the figure). The Node B is effec- face; RNCs are connected together with the sometimes hidden elsewhere in the protocoltively a UMTS “base station“, while a Radio Iur interface. Each Node B is connected to an definitions. There is currently no restriction ofNetwork Controller (RNC) is comparable with RNC by the Iub interface.There are some fun- the numbers of Nodes B in a Radio Networka GSM Base Station Controller (BSC). Each damental limits on the numbers of cells and Subsystem (RNS) or PLMN. RNCs that can be supported, due to the way
UTRAN architectureAccording to standardisation the limits are asfollows:s Maximum number of Cells in a PLMN Iu Iu 26,435,456s Maximum number of RNCs in a PLMN RNS RNS 4,096s Maximum number of Cells in an RNS RNC RNC 65,536 Iurs Maximum number of Nodes B in an RNS No limit defined in the standards Iub Iubs Maximum number of Cells in a Node B Node B Node B Node B Node B No limit currently defined in the standardsIn practice, the maximum numbers supportedby the vendors will vary and are likely to belower than the absolute limit stated here.Infrastructure sharingGiven the limited number of sites for new base operators. In contrast to the mechanical antennas and feeders, and assuming thatstations, and the cost of errecting new masts, issues, there should be no problem with the the structure is capable of withstanding thesite sharing between 3G and GSM is likely to co-location of W-CDMA and GSM900/1800 additional wind load. This has to be deter-be of importance, especially for existing sites. It should be possible to share the same mined on a case by case basis. headframe between GSM and UMTS, assum- ing there is sufficient space for the additionalHierarchical cell structuresUMTS, as GSM, supports the deployment of specified in UMTS will result in a minimum equipment designed to a later release of themicro cells within macro cells to provide obtainable cell radius, which is accentuated standards is available. These issues requireincreased capacity in traffic hot spots and when good line of sight is achieved. There is further investigation. There are two options forcoverage where previously none has existed. also some doubt about the suitability of the the choice of carrier for micro cells:However, there is some concern that the currently specified soft handover mechanism s Same carrier for micro/macro cellslimited dynamic range of the terminal power as for use in contiguous micro-cellular coverage s Different carriers for micro/macro cells areas. Therefore it could be that micro cells cannot be designed to perform optimally untilIncreasing the coverage areaThe UTRAN will support six sectored sites, 3 sectors – 900 beamwidth (left) compared to 6 sectors –which could maximise coverage and capacity 600 beamwidth (right)of UMTS sites. The basic principle is that byusing six narrow beam antennas, the coveragearea of a cell will be extended due to theincreased forward gain, and the capacity willbe double that of a three-sectored cell. Theuse of six sectors can lead to an increase inthe coverage area that is served by multiplecells (i.e. the soft handover region), depend-ing on the local propagation conditions andthe antenna pattern. The two figures show theoverlap between the antenna patterns. Thisdoes not match the soft handover regions, butit shows, how the overlap can increase, givencertain antenna beamwidths.
Conclusionss The number of services in a UMTS system coverage is intrinsically linked to the performance of the TDD mode is more influ- is substantially higher compared to GSM, capacity of the system. Cells are breathing; enced by the mobile speed than the FDD which makes the network design more the coverage range for voice varies between mode. For the voice service, the UMTS cell complex. Packet switched mode allows cost- 200 m and 1.4 km, depending on the radius is greater than the GSM 1800 one. effective transport of data, but requires QoS number of users. Traditional static predic- Data services with data rates higher than control. Some applications such as voice or tion methods for network planning are not 384 kbit/s have a lower cell radius com- real-time video require throughput with a applicable. pared to GSM 1800. guaranteed data rate and maximum delay. s Two link level simulators (W-CDMA and TD- s The Project has reviewed available system Mobile communication applications have to CDMA) have been developed in the project level simulators, and established scenarios be designed according to the user mobility, to evaluate the radio performance of UTRA. for system level simulations. A future pro- the radio environment (user speed and The main outcome of link level simulations ject is envisaged to analyse these scenarios. coverage radius), the application topology, is the system working point, the minimum and the user terminal requirements. Eb/No (ratio between energy per bit and A more detailed version of this deliverable is Current applications content, e.g. JPEG, noise). Voice services have an almost available at: does not allow missing data. constant system working point with respects UMTS radio interface has a strong impact to the mobile speed in the range of http://www.eurescom.de/ on the QoS of applications, requiring an 3-250 km/h. Data services (LCD & UDD) public/projects/P900-series/ error-resistant mechanism to obtain the are more sensitive to the mobile speed and p921/P921.htm required QoS level. In a CDMA network to the propagation environment. The linkAbout P921 Publications resulting from this work:EURESCOM Project P921-PF started on 1. D. Wake and R. E. Schuh, IEE Electronics 3. Ralf E. Schuh and David Wake, Proceedings,23 February 1999 with a planned duration of Letters, vol. 36, no. 10, pp. 901-902, 2000. IEEE International Conference on Third18 months. The total budget was 100 MM. 2. D. Wake and R. E. Schuh, Technical Digest, Generation Wireless Communications, IEEEAdditional information can be obtained from: International Topical Meeting on Microwave 3g Wireless2000, San Francisco, Silicon Photonics – MWP’99, Post deadline paper, Valley, USA, ISSN No. 1529-2592 (2000), http://www.eurescom.de/ Session F-12, pp. 9-12, ISBN 0-7803-5558- pp. 48 – 51, June 14 - 16, 2000 public/projects/P900-series/ X, Melbourne, Australia, November 17 – 19, p921/P921.htm 1999.The Project team: Project Members Name Company Email Josef Noll (Project Leader) Telenor email@example.com Jon Harris BT firstname.lastname@example.org Milan Jankovic Community of Yugoslav PTT email@example.com Borislav Odadzic firstname.lastname@example.org Armando Annunziato CSELT – Telecom Italia Group email@example.com Enrico Buracchini firstname.lastname@example.org Bruno Melis email@example.com Anne-Gaële Acx France Télécom firstname.lastname@example.org Jean-Francois Chaumet email@example.com Nicolas Guerin nicolas.Guerin@rd.francetelecom.fr Georgos Agapiou OTE firstname.lastname@example.org Dimitrios Xenikos email@example.com Amparo Sanmateu T-Nova firstname.lastname@example.org Ignacio Berberabana Telefónica I+D email@example.com Héctor González firstname.lastname@example.org Fernando Martinez email@example.com Jorge Montero firstname.lastname@example.org Arild Jacobsen Telenor email@example.com Tor Jansen firstname.lastname@example.org Tore Arthur Worren email@example.com Uwe Herzog (Project Supervisor) EURESCOM firstname.lastname@example.org