traffic management strategies for multi carrier

Practical Traffic Management Strategies for Multi-Carrier CDMA

Nortel Confidential and Proprietary 06/15/99

Practical Traffic ManagementStrategies for Multi-Carrier CDMA

(Issue 0.3)

Muhieddin NajibDept. 2779

Core RF Engineering

Wireless Solutions

NORTEL NETWORKS

June 1999


Nortel Confidential and Proprietary 1 06/15/99

Table of Contents

1 Introduction............................................................................................................ 31.1 Scope .............................................................................................................................3

1.2 Audience........................................................................................................................3

1.3 Document Revision History ..........................................................................................3

1.4 Contact ..........................................................................................................................3

1.5 References .....................................................................................................................3

1.6 Acknowledgment...........................................................................................................3

2 Overview of Traffic Management Methods for Multi-carrier CDMA .................... 42.1 Introduction ..................................................................................................................4

2.2 Traffic Management Using the IS-95 Hashing Function .............................................52.2.1 Channel Assignment.......................................................................................................... 52.2.2 Deployment Recommendations and Possible Uses ............................................................. 62.2.3 Advantages........................................................................................................................ 72.2.4 Disadvantages ................................................................................................................... 7

2.3 Traffic Management Using GSR Messaging................................................................82.3.1 Channel Assignment.......................................................................................................... 82.3.2 Possible Uses..................................................................................................................... 9

2.4 Traffic Management Using Neighbor Pilot’s Ec/Io......................................................92.4.1 Idle-frequency handoff between two Carriers ..................................................................... 92.4.2 Deployment Considerations ............................................................................................... 9

2.5 Traffic Management Using MCTA ............................................................................102.5.1 Introduction......................................................................................................................102.5.2 Channel Assignment.........................................................................................................112.5.3 Cell ID and the Pilot Data Base (PDB)..............................................................................112.5.4 Populating of SBSC Databases .........................................................................................122.5.5 Carrier Determination Algorithm (CDA)...........................................................................122.5.6 A Special Case .................................................................................................................142.5.7 Recommended Datafill – Configuring the MCTA .............................................................15

3 Capacity Considerations for Traffic Management Methods................................. 173.1 Erlang Capacity for MCTA........................................................................................17

3.1.1 Erlang Capacity for MCTA with up to Three Carriers .......................................................173.1.2 Erlang Capacity of MCTA with more than Three Carriers .................................................20

3.2 Erlang Capacity of Hashing and GSR Functions.......................................................20

3.3 Impact of Paging on MCTA Capacity........................................................................203.3.1 One RF carrier with Half-Rate Paging (Baseline Scenario) ................................................213.3.2 One RF carrier with Full-Rate Paging ...............................................................................213.3.3 Two or Three RF carriers with one Half-Rate Paging ........................................................213.3.4 Two or Three RF carriers with one Full-Rate Paging.........................................................21

4 Traffic Management in Hybrid (Mobile/Fixed) Networks.................................... 224.1 Traffic Management in Fixed Wireless Networks .....................................................22



4.2 Traffic Management in Hybrid Systems ....................................................................224.2.1 First Approach (Preferred): Mixing Fixed and Mobility Traffic Across All Carriers...........224.2.2 Second Approach: Separating Fixed and Mobility Traffic on Different Carriers.................23

4.3 MCTA Erlang Capacity of Hybrid Systems ..............................................................24

5 Deployment Issues and Recommendations........................................................... 255.1 One Paging Channel and Excess Delay ......................................................................25

5.2 More than 3 Carriers Situation ..................................................................................255.2.1 Hashing in Conjunction with MCTA.................................................................................26

5.2.1.1 Four Carriers................................................................................................................265.2.1.2 Five Carriers ................................................................................................................26

5.3 When One Carrier is Wilted.......................................................................................27

5.4 MCTA on Multi-Carrier Border Cell Sites................................................................285.4.1 Problem Statement ...........................................................................................................285.4.2 Deployment Considerations ..............................................................................................28

5.5 Disabling the MCTA Feature on Metro Cell Sites .....................................................305.5.1 Two Carriers ....................................................................................................................305.5.2 Three Carriers ..................................................................................................................315.5.3 In-door Metro Cell (Four Carriers)...................................................................................31

5.6 MCTA Deployment for Hot Spot Areas.....................................................................31

5.7 MCTA Restriction on Using PN Offsets.....................................................................32

6 Acronyms.............................................................................................................. 33



1 Introduction

1.1 ScopeThe intention of this document is to discuss several different methods of trafficmanagement for multi-carrier CDMA systems. The advantages, applications andlimitations of each method are discussed. Deployment recommendations andpossible uses of each method are provided. Finally, the RF capacity of all methodsis discussed and the impact of paging on air-interface capacity is quantified.

1.2 AudienceThis document is intended for Nortel Networks CDMA RF Engineering teams in allregions.

1.3 Document Revision History

IssueNo.

ReleaseDate Reason(s) for Reissue

Author &Department

0.1 04/29/99 Initial draft for internal review Muhieddin Najib(2779)

0.2 05/05/99 For review by Core RF Engineering,Product Line Management, CDMAPerformance Group, SystemEngineering and RF Engineering (US)

Muhieddin Najib(2779)

0.3 05/26/99 For release to audience Muhieddin Najib(2779)

1.4 ContactThis document is maintained by Nortel Core RF Engineering. For moreinformation, contact Muhieddin Najib at ESN 444-2404.

1.5 References[1] Brian Troup and Thomas Snellbaker, MCTA Simply Stated, 2nd edition.[2] Reid Chang, Distributing Traffic Among Two Carriers, version 0.0.[3] Nishith Tripathi, Multicarrier Traffic Allocation, Issue 1.0.[4] Sarvesh Sharma, CDMADPLY AA03 - MCTA Deployment Guidelines, pp. 185 –

198, Sec.6.Also, Multi-Carrier Deployment Guidelines, pp. 209 – 217, Sec.8.[5] Muhieddin Najib, “CDMA FWA RF Planning Guidelines”, Issue 0.7, Sec. 3.[6] Yves Choiniere, A presentation on “Multi Carrier Traffic Allocation (MCTA)”,

March 1999.

1.6 AcknowledgmentI wish to thank Brian Troup and Farhad Bassirat for the helpful discussions andinvaluable comments.



2 Overview of Traffic Management Methods for Multi-carrier CDMATraffic can be balanced across the carriers either when the mobiles are in idle modeor during call setup. Idle-mode balancing can be achieved either usingGlobalServiceRedirection or CDMAChannelList in conjunction with Hashingfunction at the mobile station. Traffic balancing during call setup can be done usingMulti-Carrier Traffic Allocation (MCTA) feature.

Throughout this document, it is assumed that F1 is the primary (or common) carrierfrequency. F2, F3, etc are the added carrier frequencies and referred to as secondarycarriers. The following sections describe the methods in more detail.

2.1 IntroductionThere has been a phenomenal demand for increased voice capacity as well as needfor fax and data services. In areas where the traffic demand is high and with onlyone RF carrier per cell site, the BTS RF resources saturate resulting in callblocking. One way to alleviate this problem is to deploy multiple carrierfrequencies in the same geographical region.

Each CDMA carrier occupies 1.25 MHz. For PCS band, the spectrum allocated toBlocks A, B and C is 15 MHz both forward and reverse links. The spectrumallocated to Blocks D, E and F is 5 MHz both forward and reverse links. Therefore,the number of CDMA carriers that can be accommodated in Blocks A, B and C is11, while Blocks D, E and F can accommodate 3 carriers.Table 1 summarizes the CDMA frequency assignments for the PCS band.

For the cellular spectrum, there are two designated blocks: A and B. Both blocksoccupy 12.5 MHz for the forward and reverse links. The channel spacing for thecellular band should be based on 1.26 MHz because the IMF and Combiners aredesigned to work with 1.26 MHz. The maximum number of CDMA carriers thatcan be accommodated in each Block is 9. Table 2 summarizes the CDMAfrequency assignments for the cellular band. (There are restrictions on using all 9carriers as noted in the footnotes).

Table 1CDMA preferred Set Frequency Assignment for PCS Band

Transmit FrequencyBand (MHz)Block

Designator PersonalStation

BaseStation

Preferred Channel Number

A 1850-1865 1930-1945 25, 50, 75, 100, 125, 150, 175, 200, 225, 250, 275

D 1865-1870 1945-1950 325, 350, 375

B 1870-1885 1950-1965 425, 450, 475, 500, 525, 550, 575, 600, 625, 650, 675

E 1885-1890 1965-1970 725, 750, 775

F 1890-1895 1970-1975 825, 850, 875

C 1895-1910 1975-1990 925, 950, 975, 1000, 1025, 1050, 1075, 1100, 1125, 1150, 1175



Table 2CDMA preferred Set Frequency Assignment for Cellular Band Class 0

Transmit Frequency Band (MHz)BlockDesignator Mobile Station Base Station Preferred Channel Number1

A”AA’

824.025 – 835.005844.995 – 846.495

869.025 – 880.005889.995 – 891.495 31, 73, 115, 157, 199, 241, 283, 691, 10192

BB’

835.005 – 844.995846.495 – 848.985

880.005 – 889.995891.495 – 893.985 384, 426, 468, 510, 552, 594, 758, 7353, 7772

A-Band primary channel: 283 B-Band primary channel: 384A-Band secondary channel: 691 B-Band secondary channel: 777

If the system deploys more than one CDMA carrier, then users are assigned to oneof the available carriers using either Hashing function, GSR messaging or theMulti-Carrier Traffic Allocation (MCTA). In the following subsections, we providean overview of the functionality of these methods and in the next section, wediscuss the capacity considerations associated with each method.

2.2 Traffic Management Using the IS-95 Hashing Function

2.2.1 Channel AssignmentUnder normal operation conditions, uniform distribution of mobile stations amongthe available RF carriers is desired. The Hashing function achieves randomdistribution of mobiles among the RF carriers, while the mobiles are in the idlemode. Using this method, the mobile hashes when it powers up or when it enters aregion where the channel list message broadcast by paging channel indicatesavailability of different frequencies.

Figure 1 depicts a typical deployment scenario of two RF carriers F1 and F2.During idle mode, when the mobile leaves the F1 region and enters the (F1+F2)region, it will find the Channel List message of the primary carrier paging channelcontaining the two frequencies. It will use the Hashing algorithm to determinewhether it should stay in F1 or idle handoff to F2. The Hashing function makesdecision based on mobile’s IMSI number and the frequencies available in channellist message, which divides the traffic equally between the available carriers.

Using the Hashing function, the mobile chooses the CDMA carrier according to thefollowing equation:

1 CDMA deployment has to start with either the primary or secondary channels. The first CDMAchannel requires 1.77 MHz (1.25 MHz is used for traffic and 0.52 is used as guard bands).Additional CDMA carriers can be deployed without guard bands, spaced 1.26 MHz apart (i.e., 42AMPS channels).2 Requires frequency coordination with non-cellular carriers.3 Requires frequency coordination with A-Block carrier.



CDMA Channel Number =

N × ((40503 × (L ⊕ H ⊕ DECORR)) mod 216 ) / 216 + 1

where x is the largest integer less than or equal to x, N is number of the availableRF carriers in the channel list, HASH_KEY is equal to the 32 least significant bitsof IMSI_S1 + 224 × IMSI_S2) and DECORR = 0. In the above equation, the WordL is the bits 0-15 of HASH_KEY and Word H is the bits 16-31 of HASH_KEYwhere bit 0 is the least significant bit of HASH_KEY.

Example:Depending on the content of the Words L and H, the function (L ⊕ H ⊕ DECORR)takes values between 0 and 216. Suppose that for one mobile (L ⊕ H ⊕ DECORR)= 64 and there are 3 carriers in the channel list message (i.e., N = 3). In this case,the Channel Number = 3 × ((40503 × 64) mod 216 ) / 216 + 1 = 2.

(For more information on the Hashing algorithm, refer to IS-95, section 6.6.2.2.4.)

2.2.2 Deployment Recommendations and Possible UsesFor the two-carrier case, both frequencies (F1 and F2) should be datafilled to bothcarriers’ paging channel (Channel List Message) in all sectors within the (F1+F2)region. (All frequencies should be listed in the same order in the Channel ListMessage on all BTSa). The primary carrier frequency F1 should be datafilled onlyoutside the shaded region of Figure 1. The procedure can be also extended to twomore than two carriers.

The multi-carrier region (F1+F2) should be deployed as a cluster with at least twotiers of cell sites. This is because, during the call, the mobiles cannot be re-distributed to other carriers in order to offload the overloaded carrier. If the callduration is long and mobility is high, then there is a tendency to overload theprimary carrier F1. Because a call initiated from F1 region entering (F1+F2) regionwill stay in F1 (soft handoff). A call initiated from F2, when leaving the region willbe handed down to F1 (hard handoff); and if it returns to (F1+F2) region again, itwill stay with F1 (soft handoff).

If the (F1+F2) region is large enough so that the border cells only occupy a verysmall portion of the entire (F1+F2) region, or if the mobility is not high, or if thecall duration is not too long, this load unbalancing effect may be neglected.However, if the effect of load unbalancing is not desired, one may considerintentional steering of more traffic toward the secondary carrier (s) during the idlemode so that the overall loading of the two carriers can be approximately balanced.However, the IS-95 Hashing function can only distribute traffic evenly among Ncarriers. On the other hand, the GSR messaging can be used to provide unevendistribution of idle mode traffic between two carriers.



F1 + F2RTD RTD

RTD RTD

RTD RTDF1

F1 F1

F1

Idle Handdown Boundary

Hard Handoff Boundary

BorderSectors

BorderSectors

Figure 1. Classification of traffic regions for two RF carriers.

One possible use for the Hashing function is in conjunction with the MCTA withmore than three carriers. In this case, the Hashing algorithm can be used to hashamong all available carriers. Several MCTA’s are then enabled over several groupsof carriers, each with three or less carriers (see Section 5.2.1 for more details).

2.2.3 Advantages• Does not consume any additional BSC/BTS/RF resources.• Does not require the overlying frequencies to be on the same BSC. This is

significant when the number of available ports on a BSC is limited. However,this may not apply to the Metro Cell.

2.2.4 Disadvantages• For the Legacy product, the channel list is not settable. Thus, any change to the

channel list requires a BTS download. Consequently, erratic service (or noservice) will be encountered if the channel list contains a frequency which iseither wilted or out of service. The mobile will keep looking for the frequencywhich is not there and it will not move on to another frequency. All frequencies



included in the channel list must be blossomed at all times. Failure to do so willresult in mobiles showing “no service”. This problem is alleviated in the MetroCell BTS since the channel list is settable.

• Hashing requires paging on all carriers, which leads to a decrease in the availabletraffic channel power due to the power being used in paging channels.

• Does not take the excess capacity of other carriers into account when a call isbeing setup. Consequently, the carrier loading may not be balanced althoughmobiles are evenly distributed.

2.3 Traffic Management Using GSR MessagingSimilar to the method of Channel List + Hashing functions, traffic can also bedistributed across two carriers by the Global Service Redirect message (GSR) plusthe Access Overload Class (ACCOLC) of the mobile. A GSR message is sent overthe paging channel to redirect mobiles to different carriers within the system, toanother system, or another service provider such as AMPS. Unlike Hashingalgorithm, GSR can provide uneven distribution of traffic between two carriersonly. This is further explained in section 2.3.1.

2.3.1 Channel AssignmentAll sectors within the F1+F2 region are datafilled with GSR with some percentageof the mobile stations REDIRECT_ACCOLCr subfields turned “on” (the subfieldcorresponding to the Access Overload Class of the mobile station set to ‘1’ in theREDIRECT_ACCOLCr field of the received message). Mobiles with ACCOLCfield being turned on are directed to F2 by the GSR. Others stay on F1. Sincemobile stations that are not for test or emergency use are assigned to overloadclasses ACCOLC 0 through ACCOLC 9, one can send 10%, 20%, … , 80%, 90% ofmobiles to F2. For example, with subfields corresponding to classes 0, 1, 2, 3turned off, the rest turned on, one can keep about 40% of mobiles in F1, and directthe rest 60% to F2.

If the GSR message is turned on, it should be the only mechanism used to steertraffic, so Hashing function should be turned off, i.e., F1’s Channel List should notcontain any other frequency.

Thus, using GSR and turning on a percentage of REDIRECT_ACCOLCr fields, itis possible to control the percentage of idle mobiles that are redirected to F2.

For more information on GSR message, please refer to Section 6.6.2.2.6 of the IS-95 standards.



2.3.2 Possible Uses1. During normal operation of the system to redirect mobiles to another system or

another service provider such as AMPS. This is required when such services orsystems are collocated in the same coverage area.

2. During maintenance of the system to redirect mobiles to a different carrierwithin the system.

3. On multi-carrier border sites. Referring to Figure 1, idle-mode mobiles leavingthe (F1+F2) region on F2 are directed to go to F1. This method is preferredover the idle hard hand-off method that is based on the “Neighbor Pilot’sEc/Io”.

2.4 Traffic Management Using Neighbor Pilot’s Ec/Io

2.4.1 Idle-frequency handoff between two CarriersThe previous two methods steer the traffic in a fixed fashion, either according tomobiles IMSI or according to their ACCOLC.

This method uses inter-frequency idle mode handoff to adaptively steer idlemobiles over two-carrier sectors. An idle handoff occurs when a mobile station hasmoved from the coverage area of one base station into the coverage area of anotherbase station during the Mobile Station Idle State. If the mobile station detects aPilot Channel signal from another base station that is sufficiently stronger than thatof the current base station, the mobile station determines that an idle handoffshould occur.

Each multi-carrier BTS site is datafilled to include its opposite in the idle neighborlist, i.e., datafill the overlapping sector in F2 as F1’s neighbor, and data fill F1 asF2’s neighbor. During inter-frequency idle mode handoff, the traffic is steeredtoward the carrier with better Ec/Io, and the carrier with lighter traffic loading willhave a better Ec/Io.

Initially, mobiles moving into (F1+F2) region will stay in F1. If F1 is overloadedits Ec/Io will degrade. At the point (Ec/Io)F2 > (Ec/Io)F1 + 3 dB, all idle mobilessatisfying this condition will handoff toward F2. If F2 is overloaded, and F1’s Ec/Iobecomes 3 dB better, the idle handoff will go toward the other way (see Figure 2).

2.4.2 Deployment Considerations1. Initially, all mobiles will stay on F1 until the Ec/Io of F2 at their locations

becomes 3 dB above the Ec/Io of F1. When this happens, all idle mode mobileswill hand-up to F2 until its Ec/Io degrades by another 3 dB compared to that ofF1. Since a hysteresis width of 6 dB between handing down and up can only becaused by a large number of users, the instantaneous loading between the twocarriers appears unbalanced, although a balanced loading can be achieved over



a wide observation window. Therefore, this method is best suited for high denseurban areas where traffic loading is usually high.

2. If it is desired to have higher percentage of mobiles be idle handoff toward F2,then one may intentionally make the (Ec/Io) of F2 stronger. For example, onemay reduce the pilot gain of F1 in the (F1+F2) region, so that in this region the(Ec/Io)F2 is 1 - 2 dB better than (Ec/Io)F1, initially.

3. The 3-dB threshold used to hand-off from one carrier to another is mobilevendor specific. Some vendors may implement different numbers while othersmay not entirely support such feature.

4. This solution is not recommended at this time, but listed here for thecompleteness of the subject.

F2 (Ec/Io)F2

F1 (Ec/Io)F1

If (Ec/Io)F2 > (Ec/Io)F1 + 3 dB, F1 => F2If (Ec/Io)F1 > (Ec/Io)F2 + 3 dB, F2 => F1

Figure 2. Traffic Management Using Neighbor Pilot’s Ec/Io.

2.5 Traffic Management Using MCTA

2.5.1 IntroductionEfficient use of the additional carriers is utilized if the arrival traffic is assigned tothe carrier with the least interference. The Nortel Multi-Carrier Traffic Allocation(MCTA) feature can dynamically balance traffic across multiple RF carriers servedby the same system. MCTA operates at call setup and supports up to 3 RF carriersper sector. For the MCTA to work, the multi-carrier BTSs (F1, F2 and F3) shouldbe on the same BSC. This feature has become available since the introduction ofNortel Base Station Software – version 7.0 (NBSS7.0). It works with a single



Metro Cell BTS, several Legacy BTSs and a combination of Legacy and MetroCell BTSs.

The MCTA uses the excess forward link capacity as a criterion to assign calls to RFcarriers. When a new call arrives, the carrier with the maximum available (i.e.,remaining) power is assigned to the call and the traffic is balanced among carriers.MCTA assigns traffic to carriers during call origination.

In the following sections, we provide a brief description of the MCTA functionalityand datafill. For more details, please refer to references [1] and [4].

2.5.2 Channel AssignmentAt call setup time, MCTA selects which carrier will service the call. The selectionprocess is accomplished as follows: During new call or hard hand-off setup, theSBS Selector sends a “capacity request” message to each Base-station TransceiverSubsystem (BTS) configured as part of a MCTA cell site. Each BTS responds tothe request with its capacity data in a “capacity response” message. Upon receivingthis message, the Selector Bank Subsystem Controller (SBSC) uses the CarrierDetermination Algorithm (CDA) to select the best carrier, using a combination offrequency priority, excess capacity threshold, and the excess capacity provided inthe capacity response message. CDA then returns the address of the BTS(BTSCRM Address), which supports the carrier chosen to service the call. Thesame procedure applies if a Metro Cell is servicing the sector, except that the MetroCell can support multiple carriers. Note that once the call has been setup, MCTAfeature is no longer involved.

2.5.3 Cell ID and the Pilot Data Base (PDB)To configure a cell site as a MCTA cell site, all sectors of each carrier covering aparticular area must have the same cell ID (where the Cell ID is the cell numberand the sector ID). This does not mean there is only one Pilot Data Base (PDB)record needed for that particular sector. Because the area is being covered bymultiple frequencies, and frequency is part of the Extended Basestation ID (EBID),a PDB record is needed for each frequency being used. The MCTA is automaticallyactivated when more than one Extended Base ID entry in the PilotDatabase has thesame Cell Id and sector number.

When populating a PDB, multiple records will share the same Cell ID if the cellshave co-located BTSs or a Metro Cell with multiple frequencies. This is simplybecause the Extended Base ID (EBID) contains four fields: CDMA channelnumber, frequency band (800 or 1900 MHz), Cell Number, and Sector ID. Cell IDis the Cell Number plus the Sector ID. Since CDMA frequencies of BTS-1 differfrom that of BTS-2 in co-located BTS cells (or the Metro has multiple frequencies),there are multiple PDB records per same Cell ID.



2.5.4 Populating of SBSC DatabasesPilot Database (PDB) record can be modified to include the Frequency Priority andthe Capacity Threshold. When populating a PDB, multiple records will share thesame Cell ID if the cells have co-located BTSs or a Metro Cell with multiplefrequencies. Pilot Database (PDB) records are populated through the Base StationManager (BSM) by running the datafill script files which are run from the BSMCommand Line Interface (CLI) window. In the mean time, whenever a PDB recordis added, a corresponding Cell ID Database record is also automatically derived byextracting the Cell ID, ACN address, EBID, priority and threshold from the PDBrecord. If the Cell ID record already exists, the ACN Node ID, the EBID, thepriority, and the threshold are added to the array and the number of ACN NodeID’s are incremented by one.

2.5.5 Carrier Determination Algorithm (CDA)The CDA determines the best carrier using a combination of Frequency Priorityand threshold response retrieved from the Capacity Request Response Message.When the SBSC sends the Capacity Request to all BTSs it starts time-out timer.This timer is configurable from 0 to 255 ms (milli-seconds). A value of 100 ms isrecommended. The following attributes are added for each PDB record to helpCDA do its job:

• MCTA Priority (when datafilled at the AdvancedFA(n) MO)4 or FrequencyPriority (when datafilled at the Pilot Database MO)5 – A lower value means thecarrier has a higher preference. This datafillable parameter is used to determinethe order in which the carriers should be analyzed. The capacity requests are sentsequentially starting with the BTS with the highest priority carrier first (this doesnot mean that the capacity response from the BTS with the highest prioritycarrier will always be received first). The preferred carrier is analyzed first and ifit meets the capacity criterion (if its relative capacity is positive), it will beselected first. If the carrier with highest priority can not be selected, the test willbe performed on the second highest priority carrier, and so on. If all relativecapacities are negative, then the carrier with the largest relative capacity isselected (see also flowchart in Figure 3).

- Recommended value for even loading: 0 . (This is also the defaultvalue).

• MCTA Threshold (when datafilled at the AdvancedFA(n) MO)4 or CapacityThreshold (when datafilled at the Pilot Database MO)5 – Represents theminimum value which Excess Forward Link Capacity (EFLC) in units of callsshould drop below before selecting the next preferred frequency. (i.e., thenumber of calls to admit to a high priority frequency before selecting the nextpriority frequency).

4 If only one Metro Cell is used.5 If (1) Legacy BTSs are used, (2) More than 1 Metro Cell BTS is used and (3) Legacy and MetroCell BTSs are used.



- Recommended value for even loading: 64 (This value causes allRelative Capacity responses to have a negative number). This is also thedefault value.

Note:EFLC represents the available power in the carrier (i.e., the difference between totalavailable forward link power and total power being used by existing calls andoverhead channels). The EFLC is then converted to number of calls according tothe formula:

lPowerfficChanneAverageTraoldkingThreshLoCallBlocEFLC

EstimateCapacity−= (1)

The blocking threshold has the effect of reserving some of the available forwardtransmit power in a sector. The AverageTrafficChannelPower is determined as

in Use Channels Traffic ofNumber PowerTx Channel Traffic Total ousInstantane=lPowerfficChanneAverageTra (2)

Find the preferred frequency ( F ) using the priority entries

Calculate Capacity Estimate for eachfrequency in a sector

is Capacity Estimate of (F)

> MCTA threshold

of F

Select frequency ( F )

Find the next preferred freq. using the priority entries

Are there morefrequencies available?

Select the frequency with the largest Relative Capacity (voting stage)

Y

N

Y

N

Figure 3. Flowchart of CDA algorithm



The Algorithm makes “Capacity Requests” to all BTS serving the sector on whichthe call attempt was made. Each carrier is assigned a capacity threshold. Itspurposes are to:

• Identify desired level of usage in F1 before MCTA allocate calls in F2• Reduce hard handoff by maximizing F1 use at the multi-carrier border sites

The CDA then calculates the Relative Capacity values between the estimatedabsolute capacity of each BTS and the MCTA Threshold. i.e.,

Relative Capacity = Estimated Capacity - MCTA Threshold (3)

CDA will then select the first positive Relative Capacity value received. If noCapacity Response has a positive Relative Capacity, CDA waits (time permitting)for Capacity Responses from all BTSs serving that sector to return. If all theRelative Capacity responses are negative, it selects the frequency that has thehighest priority. If all priorities are equal, then CDA chooses the least negativevalue of the Relative Capacity responses and the call is setup on that carrier.In case that a “time-out” is reached and the SBSC has not received all of the BTSCapacity Responses, the SBSC will use available responses for the CarrierDetermination Algorithm. All late capacity responses will be ignored.

2.5.6 A Special CaseAs indicated by equations (1) and (2), the capacity estimate calculation fails whenthere are no traffic channels in use in the specified sector. In this case, the capacityestimation is based on the channel element provisioned as follows:

}in UseNot BTSin Elements Channel Traffic Total

,dProvisione Sectors ofNumber

dProvisione Elements Channel TrafficCurrent min

=EstimateCapacity (4)

The “Total Traffic Channel Elements Provisioned” refers to all configured trafficchannels both free and in use.

If there are no calls on any of the carriers, the number of CE available is used toselect the carrier. The computed Estimated Capacity = Number of CE available.Therefore, if one of the carriers has more channel elements available than the other,the traffic would be redirected onto that carrier (assuming all carriers have the samepriority) until a few calls become established on that sector. Note that the TrafficChannel Elements (TCE) are assigned on a per carrier basis (even with the MetroCell), but can be pooled as one common source for all sectors (up to three sectors).

ExampleConsider carriers F1, F2 and F3 with capacity thresholds T1, T2 and T3 andpriorities P1, P2 and P3 where P1 is the highest priority and P3 is the lowest



priority. If a capacity response C1 is received from F1 such that C1 > T1 (positiveRelative Capacity, which indicates a light load) then the call is allocated to carrierF1 without waiting for a response from the other carriers. However, if there is nocapacity response from C1 or if the capacity response from C1 is such that C1 <T1, then the carrier with the highest capacity and priority is selected.

2.5.7 Recommended Datafill – Configuring the MCTAThere are two main configurations for the MCTA:

1. Even Loading: In this configuration, the traffic is distributed equally amongthe available carriers. To achieve this, the Capacity Threshold is datafilled at 64(this value causes the Relative Capacity to be a negative number) and theFrequency Priority is datafilled at 0 (no preference). This configuration isrecommended to be used in systems where the second or third carriers aredeployed over a large area.

2. Sequential Loading: Used to fill up the first carrier then the second and so on.This configuration can be used to load up the underlying frequency to minimizethe amount of HHO from one frequency to another. The Frequency Priority canbe used to give the underlying carrier a higher preference followed by thesecond carrier and so on. The value of Capacity Threshold can be adjusted upor down to change the value of the “Relative Capacity” of the sector. Blockingstatistics can be used to fine tune this value. This configuration would likely beused at a border site where a frequency overlay ends.

Figure 4 illustrates how the “Even” and “Sequential” datafill can be used toconfigure the MCTA in the core and multi-carrier border sites with three carriers.

F3F2F1

F2F1F1

Datafill: Sequential(Multi-carrier border sites)

F1: Priority = 0 Threshold = 5F2: Priority = 1 Threshold = 5F3: Priority = 2 Threshold = 5

Datafill: Sequential(Multi-carrier border sites)

F1: Priority = 0 Threshold = 5F2: Priority = 1 Threshold = 5

Datafill: N/A Datafill: Even

F1: Priority = 0 Threshold = 64F2: Priority = 0 Threshold = 64F3: Priority = 0 Threshold = 64

F3F2F1

Moving away from core of a network



Figure 4. Even and Sequential datafill for the MCTA.



3 Capacity Considerations for Traffic Management MethodsCapacity is an important factor that is used to determine which traffic managementmethod should be deployed. In this section, we provide a summary of the Erlangcapacity of the different traffic management methods discussed above.

3.1 Erlang Capacity for MCTA

3.1.1 Erlang Capacity for MCTA with up to Three CarriersEven though the maximum number of trunked channels remains unchanged, theoverall system capacity is enhanced by MCTA. This section provides performanceimprovement in Erlang capacity for both mobility and fixed networks.

The following method should be used to calculate the MCTA erlang capacity for 2or 3 RF carriers and any given GoS and number of users/sector/carrier.

GivenGoS = pCleared Call Model (i.e., Erlang-B formula)Number of users/sector/carrier = NNumber of single-pool carriers = M. Currently, the maximum number ofcarriers that can be pooled together using the MCTA algorithm is 3. Thus,Mmax = 3.

step 1: Calculate X = PPG * MIF,where X = percentage Erlang improvement with MCTA

MIF = MCTA Improvement Factor (i.e., the factor of MCTAimprovement vs. perfect pooling).

MIF = 0.55 for mobility applicationsMIF = 0.5 for FWA applications

PPG = Perfect Pooling Gain percentage improvement= (Erlangs with perfect pooling for M*N circuits/Erlangs without

pooling) –1(Use Erlang-B table to find the Erlang value that corresponds to the

desired number of circuits and GoS)

step 2: Calculate Y = M * E1C * (1 + X),

where Y = actual erlangs with MCTA improvementM = number of carriers - only valid for 2 or 3 carriers, beyond that you

must break up the carriers into groups of 3 or lessE1C = erlangs for 1 carrier, 1 sector (e.g. 6.6 for 1% blocking, 13

available circuits/users)



Table 3 provides the MCTA Erlang capacity for mobility CDMA systems for 2 and3 CDMA RF carriers and for two commonly used grade of service figures.Similarly, Table 4 provides the MCTA Erlang capacity for fixed wireless accessCDMA systems. The results in Table 3 and Table 4 assume that all the carrierscover the cell completely.

Table 3MCTA Erlang Capacity for CDMA Mobility Systems



Table 4MCTA Erlang Capacity for CDMA Fixed Wireless Access Systems

2 Carriers 3 Carriers 2 Carriers 3 Carriers15 18.28 28.88 19.97 31.3116 19.9 31.36 21.69 33.9417 21.53 33.87 23.42 36.5818 23.19 36.41 25.16 39.2319 24.86 38.95 26.92 41.9120 26.53 41.52 28.68 44.5921 28.23 44.11 30.46 47.3022 29.93 46.7 32.24 50.0123 31.63 49.3 34.03 52.7324 33.35 51.91 35.83 55.4725 35.07 54.55 37.63 58.2026 36.81 57.19 39.44 60.9527 38.56 59.85 41.26 63.7228 40.3 62.51 43.09 66.4929 42.05 65.17 44.92 69.2730 43.81 67.85 46.75 72.0531 45.58 70.53 48.59 74.8432 47.35 73.23 50.44 77.6433 49.13 75.92 52.29 80.4434 50.91 78.63 54.14 83.2535 52.7 81.35 56.00 86.0736 54.48 84.06 57.86 88.8837 56.28 86.79 59.73 91.7038 58.08 89.52 61.60 94.5439 59.88 92.25 63.46 97.3740 61.69 94.99 65.34 100.2041 63.5 97.74 67.22 103.0542 65.31 100.49 69.10 105.9043 67.13 103.24 70.99 108.7544 68.95 106 72.87 111.6045 70.77 108.77 74.76 114.4546 72.6 111.54 76.65 117.3247 74.43 114.3 78.55 120.1848 76.26 117.08 80.44 123.0449 78.1 119.86 82.34 125.9150 79.93 122.64 84.24 128.7951 81.78 125.43 86.14 131.6652 83.61 128.21 88.05 134.5453 85.46 131.01 89.96 137.4254 87.31 133.81 91.86 140.3055 89.15 136.6 93.77 143.18

GoS = 1% GoS = 2%Number of

Users/Sector/ Carrier



3.1.2 Erlang Capacity of MCTA with more than Three CarriersAt this time, the MCTA is limited to 3 carriers per traffic pool. If the capacitydemand requires deploying more than 3 carriers, then the carriers must be dividedinto two or more traffic pools, each with three or less carriers. The total Erlangcapacity of all carriers is the sum of the Erlang capacities of the individual MCTAgroups, each with 3 or less RF carriers. The distribution of the incoming trafficacross the different MCTA groups can be achieved through GSR or Hashingfunction.

3.2 Erlang Capacity of Hashing and GSR FunctionsBoth Hashing and GSR functions distribute user traffic randomly among theavailable RF carriers. The selection of the RF carrier occurs while the mobile is inthe idle mode without any consideration for the loading factor of the individualcarriers. Thus, if there are N users demanding service from a certain base stationemploying M RF carriers, then each carrier will be assigned N/M users, on theaverage, regardless of the amount of loading of each carrier. (Note that the noiserise of different RF carriers carrying the same number of calls could be quitedifferent, especially on the forward link). Suppose that at one time instant the call isassigned to the carrier that has reached its capacity limit. In this case, the call willbe blocked. In contrast, the call is less likely to be blocked using the MCTAfunction since a new call is always assigned to the carrier with less loading. Thus, ifit desired to keep blocking at the same level as that without the MCTA, thennumber of calls per carrier can be increased. The relationship between blocking andnumber of calls (or Erlangs) is approximately governed by the Erlang-B formula.

Since Hashing and GSR cannot reduce blocking, number of calls/carrier is the sameregardless of number of carriers. We conclude that there is no improvement inErlang capacity due to Hashing or GSR function. The Erlang capacity of severalRF carriers is thus the sum of the Erlang capacities of the individual carriers.

3.3 Impact of Paging on MCTA CapacityTypical pilot power setting is about 13% of the total usable HPA power. Thepaging channel power is 4.5 dB below pilot power for half-rate paging (4800 kbps).This represents approximately 5% of the total usable HPA power. For full-ratepaging (9600 kbps), the paging channel power is 1.5 dB below pilot power. Thisrepresents approximately 10% of the total usable HPA power. Assuming thatcapacity is forward-link limited and is proportional to the remaining power at theHPA (total – overhead power)6, the difference between full-rate and half-ratecapacity is 5%. Several deployment scenarios are identified below:

6 We assume that RF capacity is not affected by the reduction in interference due to pagingsuppression or power reduction, since, when the HPA reaches its capacity limit, an equivalentamount of interference is induced in the system whether it is caused by paging or traffic.



3.3.1 One RF carrier with Half-Rate Paging (Baseline Scenario)The capacity numbers that are communicated internally and externally are based onthe assumption that paging is running at half rate.

Example 3.3.1:For 8k vocoder and 8k EVRC, number of users/sector/carrier = 20 (for mobilityusers assuming tri-sectored sites, 2% achieved FER, Ec/Io of > -12 dB, etc). For1% GoS, Erlang/sector/carrier = 12.03.

3.3.2 One RF carrier with Full-Rate PagingCapacity (number of peak users in the sector) is reduced by 5% by going from half-rate paging to full-rate paging.

Example 3.3.2:For example 3.3.1, number of users/sector/carrier = 19 and Erlang/sector/carrier =11.23.

3.3.3 Two or Three RF carriers with one Half-Rate PagingThis scenario is typically associated with the MCTA where 2 or 3 carriers areserved by one paging channel. In this case, the capacity of the second and thirdcarriers is increased by 5%.

Example 3.3.3:For example 3.3.1, number of users/sector = 20 for the first carrier and 21users/sector for the second and third carriers. Using the MCTA, the averageErlang/sector/carrier = 13.82 for the two-carrier case and 14.60 for the three-carriercase.

3.3.4 Two or Three RF carriers with one Full-Rate Paging7

If one half-rate paging cannot accommodate all paging traffic required for two- orthree-carrier MCTA, then a full-rate paging channel should be used. In this case,the capacity of the first carrier is reduced by 5% and the capacity of the second andthird carriers is increased by 5%, all with respect to a half-rate paging capacity.

Example 3.3.4:For example 3.3.1, number of users/sector = 19 for the first carrier and 21users/sector for the second and third carriers. Using the MCTA, the averageErlang/sector/carrier = 13.40 for the two-carrier case and 14.31 for the three-carriercase.

7 Usually, the paging channel capacity is higher than the CM capacity. Consequently, pagingchannel load is not of a major concern.



4 Traffic Management in Hybrid (Mobile/Fixed) Networks

4.1 Traffic Management in Fixed Wireless NetworksWhen the traffic demand in a FWA network exceeds the capacity of one CDMAcarrier, one or two additional carriers are deployed above the common carrier (orcarriers) at different cell sites. The MCTA method should be used to distributetraffic across the carriers. Even if the overlaying carrier (s) do not provide acontiguous coverage, the MCTA should always be used at all sites implementingmore than one carrier including the multi-carrier (MC) border sites. This is becausethere are no hard hand-off problems associated with FWA networks.

4.2 Traffic Management in Hybrid SystemsFor hybrid CDMA networks with more than one RF carrier available, there are twopossible approaches to distributing the hybrid traffic across the CDMA carriers.One approach is to assign the fixed traffic to one set of RF carriers and the mobilitytraffic to another set of carriers. The second approach is to mix both traffic types onthe same RF carrier (s).

4.2.1 First Approach (Preferred): Mixing Fixed and Mobility Traffic Across AllCarriersIf the additional carriers cover the whole or a large portion of the geographical area,then mixing the traffic on all available carriers is recommended. In this case thetraffic allocation should be done through the MCTA algorithm. This approach ismotivated by four reasons:

1. First, this approach tends to maximize the benefit of the MCTA as comparedwith the case where the fixed traffic is assigned to one set of carriers and themobility traffic is assigned to another set of carriers.

2. Second, the traffic resources can be shared more effectively. Since the peakhour for mobile and fixed users occurs generally at different times slots and forthat matter they load the resources at different times of the day. Consequently,the MoU per carrier is maximized.

3. Third, the ability of reducing the hardware limits on capacity. Hardware limitsinclude, but not limited to, channel element, T1/E1 and BTS control processor.Since FWA capacity is expected to be higher than mobility, distributing thefixed and mobility traffic over the same carrier (carriers) will balance the trafficdistribution among the carriers.

4. Using this method, there is no need to come up with a new feature for trafficmanagement in hybrid networks.



4.2.2 Second Approach: Separating Fixed and Mobility Traffic on DifferentCarriersIn some deployment scenarios, it may become necessary to separate fixed andmobility traffic from one carrier to another. For example, if the additional carriersare deployed at isolated sites, then we face the problem of gray zones. This problemoccurs at the cell edge of an isolated carrier due to the lack of the soft hand-off gainin the reverse link. The coverage in the forward link, on the other hand, improvessubstantially due to the lack of the out-of-cell interference. As a result, a gray-zonearea representing the coverage imbalance between the links is created. Thissituation is depicted in Figure 5. To solve this problem, the fixed users in theisolated carrier (s) should be assigned to the additional carrier (or carriers using theMCTA). The mobile users should only be placed on the common carrier (orcommon carriers using the MCTA). This solution reduces the problems associatedwith border cells (i.e., hard hand problems) making it an attractive solution whenthe percentage of mobile users is small.

To achieve this traffic assignment, the fixed users in the isolated carriers areassigned specific access classes (for example access class 0 and 1). The GSR on thecommon carrier (F1) can be used to send all terminals with access class 0 and 1 tothe additional carrier (F2, F3, … ). In order to maximize the Erlang capacity, thefrequency selection should be done using the MCTA feature, which could beachieved with only one paging channel for up to 3 carriers.

Figure 5. Illustration of frequency deployment in hybrid networks withnoncontiguous coverage.

Coverage areaof commoncarrier F1 forfixed andmobility users

Coveragearea of F2for fixedusers

Coveragearea of F3for fixedusers

Gray-zone area



Mobility and fixed users in areas that have only the common carrier must thenshare that carrier. If there are more than one common carrier, then fixed and mobileusers can be assigned to either carrier. If it is desired to separate them, then theGSR can be used to direct fixed and mobility traffic to their desired carriers.

4.3 MCTA Erlang Capacity of Hybrid SystemsIf the users are 100% mobiles or 100% fixed, then the Erlang capacity that can beachieved with the MCTA is given Table 3 for mobility and Table 4 for fixed. Theprocedure of calculating the MCTA Erlang capacity for more than 3 carriers isgiven in Section 3.1.2. For hybrid systems, the Erlang capacity is the weightedlinear sum of the Erlang capacities of the individual mobile and fixed systems. Theweights are calculated according to the percentages of the mobility and fixed usersand also according to their Erlang usage per subscriber. Detailed procedure andexamples are given in [5, Sec. 3.4].



5 Deployment Issues and Recommendations

5.1 One Paging Channel and Excess DelayDeploying a second carrier without Paging/Sync (i.e., all mobiles are idle on F1)will cause mobiles to take 3 to 8 seconds before they can return to F1 after a call isestablished on F2. The phone may show "No Service" for 3 to 8 seconds, whichmay not be desirable to the customer. In this case, paging on F2 may not beavoided.

5.2 More than 3 Carriers SituationAt this time, the MCTA can support only 3 carriers per traffic pool. If the designrequirements dictate deploying more than 3 carriers, then the traffic resourcesshould be divided into several groups, each with 3 or less carriers. The distributionof the incoming traffic among the different MCTA groups can be achieved throughthe Hashing.

In order to maximize the benefit of MCTA, one should try to maximize the numberof carriers per pool, although this cannot be always achieved (see section 5.2.1).Table 5 provides a procedure of how to group the carriers.

Note that each group of RF carriers managed by an MCTA requires at least onepaging channel. In order to reduce the effect of paging on capacity, paging shouldbe transmitted at half rate. If half-rate paging cannot handle all required traffic forthe 3 carriers, then full-rate paging should be considered.

Table 5Optimal distribution of carriers across several traffic pools for amaximum number of carriers per pool of 3. This arrangement

tends to maximize the Erlang capacity of the cell siteNUMBER OF CARRIERS PER GROUPTotal Number

of Carriers Group 1 Group 2 Group 3 Group 41 1 0 0 02 2 0 0 03 3 0 0 04 3 1 0 05 3 2 0 06 3 3 0 07 3 3 1 08 3 3 2 09 3 3 3 0

10 3 3 3 111 3 3 3 2



Example:Suppose it is desired to deploy 5 CDMA carriers in one particular cell site, then thisrequires 2 traffic pools one with 3 carriers and the other with 2 carriers. For thisscenario, 60% of the incoming traffic should be directed towards the first group and40% should be directed towards the second group8.

Suppose that the system is designed for 13 users per sector per carrier at 1% GoS(assume mobility). Then, from Table 3, the total Erlang capacity of the cell site is24.39 (for 3 carriers) + 15.27 (2 carriers) = 39.66 Erlangs. This number should becontrasted with 5 x 6.61 = 33.05 Erlangs, which is the Erlang capacity of the cellsite without MCTA. Thus, for this example, the MCTA provided 20% increase inErlang capacity over the conventional method (i.e., without pooling the trafficresources).

5.2.1 Hashing in Conjunction with MCTAThere are several possibilities of distributing traffic among the available carriers,but we discuss here only the recommended solution for each scenario. The designstrategy should attempt to achieve the following goals:

1. Maximize the number of RF carriers in each MCTA group according to Table5. This tends to improve the capacity performance of the MCTA.

2. Distribute the traffic evenly across the carriers. If the this goal cannot beachieved, then one should try to allocate more traffic to the overlaid carriers(i.e., more traffic should be allocated to the carrier that has the largest coverage,followed by the carrier that has the second largest coverage, and so on).

5.2.1.1 Four CarriersTwo carriers (e.g., F1 and F2) in one MCTA group and the other two carriers (F3and F4) in another group. Using Hashing function, one paging channel per trafficpool is required. For example, the Channel Lists of both F1 and F3 should have{F1,F3}. So, 50% of the idle-mode traffic is placed on F1 and the other 50% isplaced on F3. The MCTA is then enabled on the traffic groups (F1, F2) and (F3,F4) by requiring the Cell ID of the carriers in each group to be the same in theExtended Base ID entry of the PilotDatabase.

5.2.1.2 Five CarriersTwo solutions are recommended based on number of paging channels:

1. Two Paging Channels for the Five Carriers: Two carriers (e.g., F1 and F2)in one MCTA group and the other three carriers (F3, F4 and F5) in anothergroup. Using Hashing function with two paging channels, the Channel Lists ofboth F1 and F3 should be datafilled with {F1,F3}. So, 50% of the idle-mode

8 This may not be achieved if it is desired not to have one paging channel per carrier (see Section5.2.1.2)



traffic is placed on F1 and the other 50% is placed on F3. The MCTA is thenenabled on the traffic groups (F1, F2) and (F3, F4, F5) by requiring the Cell IDof the carriers in each group to be the same in the Extended Base ID entry ofthe PilotDatabase.

The problem with this configuration is that carriers F1 and F2 are eachassigned 25% of the calls, while carriers F3, F4 and F5 are each assigned16.67% of the calls. This solution is recommended if two paging channels forthe five carriers are proven to be enough and if it is desired to place more callson carriers F1 and F2 (this may be desirable in order to avoid the problemsassociated with hard hand-off).

2. Three Paging Channels for the Five Carriers: Two carriers (e.g., F1, F2 andF3) in one MCTA group and the other three carriers (F4 and F5) in anothergroup. Using Hashing function with three paging channels, the Channel Listsof F1, F3 and F5 are datafilled with {F1, F3, F5}. So, 33.33% of the idle-modetraffic goes to each of F1, F3 and F5. The MCTA is then enabled on the trafficgroups (F1, F2, F3) and (F4, F5) by requiring the Cell ID of the carriers in eachgroup to be the same in the Extended Base ID entry of the PilotDatabase.

In this configuration, carriers F1, F2 and F3 are each assigned 22.22% of thecalls, while carriers F4 and F5 are each assigned 16.67% of the calls. Thus, thisconfiguration has a better traffic balance than the previous one (with theoverlaid carriers taking the extra loading).

5.3 When One Carrier is WiltedThe impact of wilted carrier on the MCTA functionality varies depending on howthe MCTA is implemented. Below, we identify three methods of implementing theMCTA on non-border sites and discuss the pros and cons of each method and theeffect of wilted carrier on the MCTA performance.

1. One Paging Channel per Carrier with the Frequency Number Listed inits Own Carrier (Recommended Method):The result of this configuration varies depending on the product type:

Non-Metro Cells: Using this method, F1 is only datafilled in the channel listof F1 and F2 is only datafilled in the channel list of F2. This way, if either F1or F2 goes out of service, then no mobile will go to the carrier which has goneout-of service. This is the preferred method, since it guarantees a normaloperation of one carrier when the other one is wilted. The downside of thismethod is that we are wasting a paging channel power in addition to acomplete non-usage of the secondary carrier(s). Please refer to section 3.3 fora discussion and examples of the impact of paging on the MCTA capacityperformance.

Metro Cells: This method cannot be used due to the limitation of onecommon channel list on all carriers.



2. One Paging Channel per Carrier with both Frequencies in the ChannelList of both Carriers (Not Recommended):The second approach is to include both F1 and F2 on the channel list of bothcarriers. In this case, the Hashing function is enabled. The problem with thisapproach is that if F2 goes down and a mobile is hashing on F2, it cannot findit and goes back to F1. Again, it sees both carriers on the channel list, hashesto F2, cannot find it, will come back to F1, and re-tries to hash to F2,... and theprocess will end up in a loop. The other problem is the capacity loss due to thetransmission of two paging channels.

3. One Paging Channel for only the Primary Carrier:The third approach is to have one paging channel on F1 only (with F1 on itschannel list only). This method improves the capacity of F2, since no power isdelegated to paging. The downside is that if F1 goes out of service, then nomobile will be able to access either F1 or F2.

5.4 MCTA on Multi-Carrier Border Cell SitesThe MCTA should also be used in border cells, but in conjunction with the Multi-Pilot Hard Hand-Off (MPHHO) in order to alleviate the border cell issues. MCTAborder cells need no special configuration and fit in nicely with the existingdeployment. This makes it easy to change the MCTA boundaries as needed.However, the multi-carrier border cells need to be datafilled as CELL_BORDERcell-type in the Pilot Database (PDP) on the SBS if the RTD is to be used as ahard-hand-off trigger.

5.4.1 Problem StatementWith the current implementation of MCTA algorithm, there is no distinctionbetween call setups due to new calls or hard hand-off calls. Consequently, somecalls could drop at multi-carrier border cells if MCTA places a hard hand-off callonto the same frequency that it was hard handing-off from.

In addition, since the outer tier of F2 doesn’t have neighbor cells in the outwarddirection, the coverage of F2 border sector will extend way beyond that of co-locating F1. This problem is known as a gray-zone and is depicted in Figure 6. If amobile stay on F2 and move far beyond the RTD boundary and make a call, oncethe call is setup on F2, it will immediately be hard handoff down to F1. However,the mobile’s location may no longer be within the coverage area of F1 so there is ahigh probability of HHO failure, as shown in Figure 6.

5.4.2 Deployment ConsiderationsTo solve these problems, the multi-carrier border sites should be implemented withthe following considerations:



1. Extend the RTD boundary far enough to force hard hand-off (from F2 to F1) tooccur with the neighbor cells (those without MCTA and assuming that there isonly one carrier in the neighbor cells). In Figure 1, this is given by the outertier (border sectors facing outward away from the F1+F2 region). This way, theloop condition can be reduced (i.e. call will not go to F2). At the same time,MPHHO can be used to improve HHO performance.

F2

F1

RTD HHOBoundary

F1 + F2Region

BorderSectorsHHO to F1

(Drop Call)

Figure 6. The foot-print of border sector of F2 is bigger than that of F1. If a mobilemoves way beyond RTD boundary and still makes an access on F2, after the callsetup it may be HHO to a wrong F1 cell and the call drops.

F2

F1

RTD HHOBoundary

F1 + F2Region

F1F1

F1

Idle HandDown to F1Neighbors

BorderSectors

Figure 7. The foot-print of border sector of F2 can be reduced by reducing digitalgain of its overhead channels, so that it will cover beyond RTD, but within the foot



print of F1. The mobile should perform idle hand down to one of the F1 neighborcells before it moves too far beyond the RTD boundary.

Note:The RTD boundary and the paging channel coverage limit should be within thereverse-link coverage limit of the mobile.

2. Reduce the overhead gain of the F2 border sectors, so that their coverage areawill shrink in, but still covers beyond the RTD boundary, as shown in Figure 7.Please note that this solution is not always recommended and could beprohibited if the system noise rise in the forward link results in too muchdegradation in the pilot Ec/Io.

As mentioned above, MPHHO can be utilized to alleviate the problems createdwhen MCTA is enabled at multi-carrier border sites. MPHHO allows mobile tohard handoff into up to six target cells not in the MCTA region. With multipletarget cells, it is possible to delay the hard hand-off long enough such that themobile can move out of the underlying sector and then hard hand-off to thesurrounding tier of cells. In this case, the mobile can perform hard handoff with abetter chance of surviving.

Note that if the first entry in the target list is non-MCTA site, then the MCTAalgorithm is disabled for that HHO. If the first entry in the target list is an MCTAsite, then MCTA will be invoked.

5.5 Disabling the MCTA Feature on Metro Cell SitesIn some Metro Cell sites, it may be desired to disable some RF carriers from theMCTA functionality either temporarily due to some maintenance issues orpermanently due to RF problems or operational as dictated by the customer. It isimportant to note that the MCTA should be used where it is possible, since theMCTA is the only traffic management method that dynamically balances trafficacross carriers.

5.5.1 Two CarriersTo disable the MCTA and still use both frequencies, the cell IDs must be madedifferent on both carriers. This method requires modifying scripts by hand to allowdifferent cell IDs per sector. The downsides of this approach are the obvious loss ofthe MCTA capacity gain and, for Redundant 2-carrier Metro Cells, the cells mustbe split, and as such, two T1's must be provided instead of just one. (Although, it ispossible to daisy-chain them which would save some cost).



5.5.2 Three CarriersIf it is desired to disable the MCTA on one carrier (e.g., F3) and keep it for theremaining carriers (e.g., F1 and F2), then the cell IDs on carriers F1 and F2 must bemade identical, while F3 should be datafilled with a different cell ID. In this case,one paging channel for F3 and another paging channel for {F1, F2} are needed.

5.5.3 In-door Metro Cell (Four Carriers)The indoor Metro Cell supports four carriers. The MCTA can only be enabled on amaximum of three carriers (e.g., F1, F2 and F3). Thus, if it is desired to disable theMCTA on one carrier (e.g., F4), then the cell IDs on carriers F1, F2 and F3 must bemade identical and the cell ID for F4 is made different. This scenario also requiresone paging channel for F4 and another paging channel for {F1, F2, F3}.

Note that with four carriers, it is also possible to run the MCTA in conjunction withthe Hashing function on two groups of two carrier each. Please refer to Section5.2.1.1 for details.

5.6 MCTA Deployment for Hot Spot AreasThe MCTA features the ability of assigning different frequency priorities with thedifferent carriers. One possible use of this feature is to reduce unnecessary hardhandoffs under lightly loaded conditions by making the MCTA algorithm allocatecalls on a single frequency. Such benefit is obtained only if the additional carriersare deployed to cover isolated areas (e.g., hot spots areas) as shown in Figure 8.

Thus, under lightly loaded conditions, carrier F1 is assigned higher priority in orderto keep the calls on F1 (even within the hot spots) as they move out of the hot spotarea, thereby allowing calls to perform soft handoff.

Figure 8. MCTA deployment in hot spot areas.

F1 only(Assigned

higherpriority)

F1, F2

F1, F2

F1, F2

F1, F2



However, in such cases where the entire coverage area is overlaid with theadditional frequency, such an approach may not be necessary. Under suchcircumstances, the frequencies should be assigned equal priorities to reduce callsetup delays.

5.7 MCTA Restriction on Using PN OffsetsAll co-located MCTA sectors should use the same PN Offset. This restriction isdue to the limitation of the Channel Assignment Message in IS-95-A, which doesnot have a field for PN offset. Consequently, if the Channel Assignment Message isused to redirect the mobile to a different frequency, the mobile “re-tunes” to theassigned frequency and attempts to search for the same PN offset. This is alsomandated as per IS-95-A section 7.1.3.2.1 which states, “The same pilot PNsequence offset shall be used on all CDMA frequency assignments for a given BaseStation.”



6 Acronyms

ACCOLC Access Overload ClassACN Application Communication NetworkBSM Base Station ManagerBTS Base Station Transceiver SubsystemCDA Carrier Determination AlgorithmCDMA Code Division Multiple AccessCLI Command Line InterfaceEBID Extended Basestation IDEFLC Excess Forward Link CapacityFWA Fixed Wireless AccessGoS Grade of ServiceGSR Global Service RedirectionHHO Hard Hand-OffHPA High-Power AmplifierMC Multi-CarrierMCTA Multi-Carrier Traffic AllocationMoU Minute of UseMPHHO Multi-Pilot Hard Hand-offNBSS Nortel Base Station SoftwarePCS Personal Communication ServicesPDB Pilot Data BaseRF Radio FrequencyRTD Round Trip DelaySBSC Selector Bank Subsystem Controller

traffic management strategies for multi carrier

Documents

multicarrier cdmaissue

channel assignment

possible uses

deployment recommendations

deployment considerations

muhieddin najibdept

neighbor pilots ecio

idlefrequency handoff