wcdma radio access network dimensioning for multiple .wcdma radio access network dimensioning for...

Download WCDMA Radio Access Network Dimensioning for Multiple .WCDMA Radio Access Network Dimensioning for Multiple Services ... Derivation of system capacity on generic basic become ... In

Post on 01-Feb-2018

225 views

Category:

Documents

7 download

Embed Size (px)

TRANSCRIPT

  • WCDMA Radio Access Network Dimensioning for Multiple Services

    Igor S. Simi, Ericsson d.o.o, V. Popovia 6, Beograd

    igor.simic@ericsson.com I. INTRODUCTION One of the most important characteristics of WCDMA is the fact that power is the common shared resource. This makes WCDMA very flexible in handling mixed services and services with variable bit-rate demands. Other aspect of the pooled downlink output power resource is an impact on the radio access network (RAN) design. To diverse users different amount of maximum output power can be assigned depending on its service and its downlink interference situation. Since all the users share the same power and the same frequency, the origination of new calls, or the re-negotiation of the existing ones modifies the transmitted power in the uplink and downlink affecting the quality of service (QoS) of all the users. The power and frequency sharing results in a soft capacity characteristics; no block exists in the system. In principle new users can be accepted in the system, tolerating a QoS degradation at cell edge. For this reason in WCDMA, the radio network planning and radio resource management algorithms are intended to minimize the transmitted power in the uplink and in the downlink, in order to achieve the full exploitation of system capacity and performance. The RAKE receiver attempts to recover as much power as possible from the cells in the mobile's active set (the cells it is in soft handoff with). Any non-recovered power is interference as far as the connection in question is concerned and will degrade the achieved QoS. Power control outer loops will attempt to compensate, thus increasing interference and lowering capacity. When there is low load in the system, the users do not generate much interference. The majority of the interference is non-power controlled, i.e. background noise or interference from the non-power controlled downlink broadcast channels. For the low load cases, the coverage is higher since the users in a cell do not generate much interference. During high load on the other hand, the power-controlled interference from the users constitutes the majority of the total interference and the maximum coverage is reduced. Further, since the majority of the interference is power controlled, changes in number of users will now have a larger impact on the system. The quality based power control leads to a trade-off between coverage and capacity. Different operators will use UMTS to solve their diverse needs. The usage of data services will depend on services and terminals available, operators profile including their choice of charging and the competing media at that time.

    From end-user and application point of view four major service classes can be identified and separated into: Real time applications Conversational class, where the QoS have to preserve

    time relation (variation) between information entities and to have a low delay (voice, video, CS data);

    Streaming class, where the QoS have to preserve time relation between information entities (video or audio streaming);

    Non-real time applications Background class, where the destination is not expecting

    the data within a certain time but with preserved payload content (email, messaging);

    Interactive class, where a request/response pattern is of importance and the payload content must be preserved (WWW, ftp, telemetry).

    When a user equipment (UE) wants to establish a connection to the core network (CN), regardless of which applications to be used, the UE will ask for a RAB with a set of Quality of Service (QoS) attributes. The RAB is a point-to-point connection between a UE and the core network. The most important attributes are: QoS class: Conversational, streaming, interactive or

    background Maximum bit rate: Highest bit rate desired Guaranteed bit rate: Lowest bit rate acceptable The RAB connection is realized as a radio bearer connection between the UE and RNC and an Iu bearer connection between RNC and CN. Table 1 shows a method of how applications can be mapped onto radio bearers.

    Application RAB class Radio bearer UL/DL Voice Conversational 12.2 kbit/s + 3.4 kbit/s SRB Video

    telephony Conversational Conversational1 64 kbit/s +

    3.4 kbit/s SRB Packet data (web, e-mail

    ftp, etc)

    Interactive, background,

    streaming

    32 kbit/s (FACH) 64/64 kbit/s+3.4 kbit/s SRB

    64/128 kbit/s+3.4 kbit/s SRB 64/384 kbit/s+3.4 kbit/s SRB

    V.90 Modem Conversational 57.6 kbit/s + 3.4 kbit/s SRB

    Voice + packet data

    Conversational + (interactive or

    background)

    12.2 kbit/s + 64/64 kbit/s + + 3.4 kbit/s SRB

    Table 1 Mapping of typical applications to available radio access bearers (RAB) There are strong relation between network costs and services used in dimensioning. Therefore RAN design strategy is very

  • important and should be considered first. Several concepts could be used: 1) For 2G operators the best solution is reuse of existing

    sites. It might be very inefficient to begin with a network for low-date rate services and make it tighter later on.

    2) Key service dimensioning strategy where key service is typical service the company has based its 3G strategy upon. Key service should be significant improvement over 2G services and have to be available everywhere. Since coverage area depends strongly of the service, it must also be taken into account when choosing the key service (Vodafone data services and 3s video telephony).

    3) RAN planning for predicted traffic demand and multiple services - variable load dimensioning methodology. Method is presented in section III.

    II COVERAGE AND CAPACITY ESTIMATION UPLINK CAPACITY Derivation of system capacity on generic basic become increasingly difficult with variability of services. The time consuming simulation are usually used for obtaining capacity figures. In this section rough capacity estimate is presented mainly for dimensioning purposes. Each mobile user equipment (UE) m, must be received in the base station with a signal Sm that produces the target (S/I)m. It is assumed that all mobiles requires the same target (S/I)m = m and have perfect power control.

    mthothown

    mm

    mtot

    m

    m SNIIS

    SIS

    IS

    ++==

    =

    (1)

    where Nth is thermal noise power spectral density, Iown is total received power from mobiles in own cell and Ioth is total received power from mobiles in other cells + interference from other sources. UE is performing transmitting power update in order to maintain the Eb/Io ratio constant. The ratio between Eb/Io and signal to interference ratio (S/I)m can be expressed by:

    p

    b

    m GIE

    IS 0/=

    (2)

    where required Eb/I0 is ratio between energy per bit and spectral interference density. From (1) and (2) can be expressed:

    tot

    b

    pI

    IEGS

    0/1

    1

    += (3)

    Further if M mobiles is connected to own cell

    tot

    b

    pth

    M

    iithtot I

    IEG

    MFNSNI

    0

    1

    /1

    )1(+

    ++=+= =

    (4)

    where F is the ratio between the interference coming from neighbouring cells and the own cell interference Ioth/Iown. From (4) it is possible to calculate the noise rise, the ratio between interference caused by other UEs and the thermal noise:

    th

    thoth

    pole

    th

    totul N

    NI

    MMN

    II +

    ==1

    1 (5)

    Mpole is a theoretical maximal number of UEs attached to a cell. This number can not be reached since the interference in the cell would be infinite. From (4) and (5) with assumption Iul is infinite Mpole is expressed as:

    +=

    ob

    ppole IE

    GF

    M/1

    1 . (6)

    For several services i.e. to several different targets noise rise expression can be generalised:

    th

    thoth

    Kpole

    K

    polepole

    th

    totul N

    NI

    MM

    MM

    MMN

    II +

    ==

    ,2,

    2

    1,

    1 ...1

    1 (7)

    In WCDMA analysis, it is expected to define the cell loading. For single service load is defined as:

    poleMMLoading = (8)

    where M is the number of simultaneous users in the cell. For a multi-service system where the services utilize different types of RABs, the equation can be written as:

    ...M

    MM

    MM

    MLoading

    ,pole,pole,pole

    +++=3

    3

    2

    2

    1

    1 (9)

    where Mn is the number of simultaneous users for the n RAB Mpole,n is the uplink pole capacity for the n RAB. UPLINK COVERAGE Conventional link-budget (maximum path-loss for which a system should be planned) could be determined for uplink.

    lpathmax = PUE Bsens. + PCmarg IUL - LNFmarg Gb (10) where lpathmax is the maximum path loss due to propagation [dB]. PUE is the maximum UE output power [dBm]. Bsens is the Node B sensitivity [dB] PCmarg is the power control margin [dB] LNFmarg is the log-normal fading margin [dB]. IUL is the noise rise [dB] Gb is the sum of all UL gains and loses at Node-B and

    terminal, including antenna gains and body loss. The unloaded Node B has the sensitivity level without any interference contribution from other UEs, and can be expressed as:

    0/log10 NERNNB bInfoftsens +++= [dBm] (11)

  • where Eb/N0 is the ratio between bit energy and noise spectral density [dB] Nt is the thermal noise power density (174 dBm/Hz) Nf is the noise figure RInfo is the information bit rate

    From (10) and (11) it can been seen that bit-rate and noise raise have strongest influence on the uplink budget

Recommended

View more >