PCI RACH - Planning_Topics

Planning Topics    

PCI Planning PRACH Planning UL DM RS Planning TAC Planning

Company Confidential 1 © Nokia Siemens Networks

Presentation / Author / Date

PCI Planning



PCI Planning What is the PCI? • •

Physical Layer Cell Identity (PCI) identifies a cell within a network There are 504 Physical Layer Cell Identities -> PCI is not unique! Physical Layer Cell Identity = (3 × NID1) + NID2 NID1: Physical Layer Cell Identity group. Defines SSS sequence. Range 0 to 167 NID2: Identity within the group. Defines PSS sequence. Range 0 to 2

•

PCI is not the E-UTRAN Cell Identifier (ECI) • ECI is unique within a network • ECI does not need to be planned. ECI value is set by the system

• Physical Cell Identity is defined by the parameter phyCellID: Parameter

Object

Range

Default

phyCellID

LNCEL

0 to 503

Not Applicable



PCI Planning Planning Overview •

PCI planning is analogous to scrambling code planning in UMTS: • A UE should never receive simultaneously the same identity from more than a cell • Maximum isolation required between cells with the same PCI • Neighbour cells should not have the same PCI (collision free planning) • Neighbours of neighbours cell should not have the same PCI (confusion free planning)

•

Additionally, PCI planning needs to follow the ‘PCI modulo’ rules: modulo3, modulo6 and modulo30 • If mod3(PCI) rule is true then mod6(PCI) and mod30(PCI) are true • If mod6(PCI) is true then mod30(PCI) is true • If mod6(PCI) is not true then mod3(PCI) is not true • If mod30(PCI) is not true then mode6(PCI) is not true

•

There should be some level of co-ordination across international borders when allocating PCIs • To avoid operators allocating the same identity to cells on the same RF carrier and in neighbouring geographic areas



PCI Planning Impact in Reference Signal Positions (1/2) • Reference signals are used for channel estimation, cell selection, cell reselection and handover

• The PCI determines the position of the cell specific reference signals (RS) in frequency domain – Position of RS in time domain is fixed: slots 0 and 4 of the PRB – Each RB reserves REs for 4, 8, or 12 RS depending on whether this is 1, 2, or 4 antenna ports, respectively



PCI Planning Impact in Reference Signal Positions (2/2) • RS in frequency domain can have 6 different positions per PRB across two groups – RS positions are repeated after two consecutive Groups Physical Layer Cell Identity = (3 × NID1) + NID2 NID1: Physical Layer Cell Identity group. Determined by SSS sequence. Range 0 to 167 NID2: Identity within the group. Determined by PSS sequence. Range 0 to 2

Resource elements allocated to Reference Signals



PCI Planning modulo3 (PCI) Rule

Rule: • Avoid assigning to the cells of one eNB PCIs with the same modulo 3 Reason: • PSS defines NID2. There are 3 NID2 in a group so PSS is generated using 1 of 3 different sequences • If two cells of the same eNB have the same mod3(PCI) it means they have the same NID2 (i.e. 0, 1 or 2) and the same PSS sequence – PSS is used in cell search and synchronization procedures: Different PSS sequences facilitate cell search and synch procedures



PCI Planning MIMO 2x2 • When using 2 antennas the number of RS is doubled • The position of the RS within each antenna pair (Ant0, Ant1) is fixed • With MIMO case, not following mod3(PCI) implies RS occupies the same REs • RS SINR is poor reducing the achievable throughput

RE used as RS in Ant0 are unused in Ant1 and vice-versa Company Confidential 8 © Nokia Siemens Networks


PCI Planning ‘Modulo3 (PCI)’ Rule •

modulo 3 rule should be extended to the neighbour cells outside the same eNB – Difficult to avoid mod3 collision in real networks as Mod3 is limited to 3 values (e.g. the cells of the same 3 sector site)

FDD case: • eNBs are not frame synchronised so even if two neighbour cells from different eNBs transmit the same PSS sequence/use same RE for RS it is likely that they don’t interfere in time TDD case: • Frame synchronised: Bad SINR from RS if inter-site cells have same mod3(PCI) • Tests show DL throughput is affected. Solution: Good planning to reduce overlapping areas • Trade off: RS-RS interference vs. RS-PDSCH interference – RS-RS interference: causes channel estimation degradation -> affects throughput – RS-PDSCH interference: causes data symbol puncturing lowering effective coding rate -> PDSCH throughput is also affected Company Confidential 9 © Nokia Siemens Networks


TD-LTE PCI mod3 overlap between sites Original scenario: PCI 45 and PCI47

Modified scenario: PCI 400 and PCI403

• •

Test between 2 sites with one cell each Original PCIs (left) where changed to PCIs (right) so both sites have same mod3 (PCI)=1

Effects: • SINR reduction: 17 to -2dB • Throughput is only reduced from 17Mbps to ~14Mbps

• More info: LTE Optimization Training (RF measurement and Optimization chapter): •


https://sharenetims.inside.nokiasiemensnetworks.com/O pen/426475080


Impact of PCImod3 collision on tput, TD-LTE •

Case: UE at the border of two cells who have the same PCImod3, RSRP from both cells = -67dBm in both measurement cases (only PCI changed)

•

NSN 7210 TD dongle, 2.6GHz, 10MHz bandwidth

16 14

tput, Mbps

12 10

no PCImod3 collision

8

PCImod3 collision

6 4 2 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53

seconds



PCI Planning

‘Modulo6(PCI)’ Rule Rule: • If mod3(PCI) can’t be fulfilled, avoid assigning the same mod6(PCI) to the cells of the same site Reason: • 1Tx case: PSS sequence is not unique within the cells of a site but its position the in frequency domain is still different -> not RS interference

• 2Tx case: RS to RS interference can not be avoided. The only way to avoid it when using MIMO2x2 is with the mod3(PCI) rule Summary: • For 2Tx case the cells of the same site should have different mod3 (PCI). For 1Tx case the mod6(PCI) should be different • Reason: To have frequency shifts for RS of different cells as they are framesynchronized (cells of the same site) and avoid RS interference in DL. Company Confidential 12 © Nokia Siemens Networks


PCI Planning:

‘Modulo30(PCI)’ Rule ‘Modulo 30’ Rule: • If mod6(PCI) can’t be fulfilled, avoid assigning the same module30(PCI) to the cells of the same site Reason: • mod30 is required in other planning areas like the UL Demodulation reference signal planning Example • There are 30 groups of sequences ‘u’ for PUSCH. Each cell within a site should have sequences from different groups • If the PCIs for cells of the same site have different mod30 then ‘u’ (group sequence number) is different and it is not necessary to plan the grpAssigPUSCH parameter

u  PCI  grpAssigPU SCHmod 30 Company Confidential 13 © Nokia Siemens Networks


PCI Planning Recommendations, wrap up Id = 0

In priority order, number 1 most important (all four should be fulfilled, ideally)

Id = 2

Id = 6 Id = 8

Id = 1 Id = 3

1. Avoid assigning the same PCI to

Id = 5

Id = 9 Id = 11

neighbour cells Id = 4

2. Avoid assigning the same mod3 (PCI) to ‘neighbour’ cells

Id = 7

Id = 10

Example 1 PCI Identity Plan

3. Avoid assigning the same mod6(PCI) to ‘neighbour’ cells

4. Avoid assigning the same mod30 (PCI) to ‘neighbour’ cells

Example 2 PCI Identity Plan Company Confidential 14 © Nokia Siemens Networks


PCI Planning 6 sector sites •

In 6 sectors sites is not possible to assign PCIs with different modulo 3 as we have 6 cells and only 3 different possibilities

•

If increasing sectorisation (from 3 to 6 sectors) then every second group of identities should be allocated within the initial plan –

To allow eNodeB to be allocated identities from two adjacent groups when the number of cells is increased from 3 to 6

Rule: • Planning should be done assigning PCIs from two consecutive groups and avoiding that the consecutive cell (i+1) has the same modulo 3(PCI) • By assigning PCIs from two consecutive groups the ‘module6’ rule is followed



PCI Planning Methods •

Manual • Valid for small amount of sites (e.g. trials) • No need for additional tools, just follow the rules considering the site distance and cell azimuths

• Atoll or other planning tools (e.g. Asset) • PCI planning supported • NetAct Optimizer • PCI planning supported • NSN Internal tools (e.g. Alpha, MUSA) • Alpha: https://sharenet-ims.inside.nokiasiemensnetworks.com/Open/434150579 • MUSA: https://sharenet-ims.inside.nokiasiemensnetworks.com/Open/428210505



PRACH Planning



PRACH Planning Principle PRACH configuration two cells must be different within the PRACH re-use distance to increase the RACH decoding success rate

PRACH transmission can be separated by: • Time (prachConfIndex) – PRACH-PUSCH interference: If PRACH resources are separated in time within eNB – PRACH-PRACH interference: If same PRACH resources are used for the cells of an eNodeB. – PRACH-PRACH interference is preferred to PRACH-PUSCH interference so prachConfIndex of the cells on one site should be the same

• Frequency (prachFreqOff) – Allocation of PRACH area should be next to PUCCH area either at upper or lower border of frequency band, however should not overlap with PUCCH area – Avoid separation of PUSCH in two areas by PRACH (scheduler can only handle one PUSCH area) – For simplicity use same configuration for all cells

• Sequence (PRACH CS and RootSeqIndex) – Use different sequences for all neighbour cells Company Confidential 18 © Nokia Siemens Networks


UpPTS: Uplink Pilot Timeslot. TDD specific

Preamble Formats

• 3GPP (TS36.211) specifies 4 random access formats for FDD and TDD plus an additional format (Format 4) specific for TDD that uses the UpPTS

• FDD: Only Formats 0 and 1 are supported in initial releases (up to RL30)

• TDD: Only Formats 0 ,1, 2 and 4 are supported in RL15TD Recommendation: • Select Format0 for cell 6

9.38 10  3 10  1.4km 2 8

ranges <14.53 km

• Select Format1 for cell ranges <77.34 km

• Select Format2 for cell ranges <29 km (TDD only)

• Select Format4 for cell ranges <1.4 km (TDD only)



Preamble Format 4 TDD-only • 3GPP (TS36.211) specifies a special random access format 4 for TDD • Preamble format 4 is allocated in UpPTS increasing the UL throughput as more resources can be reserved in the normal UL subframes for PUSCH CP

Sequence

TCP 14.5s

TSEQ  133 s

• Maximum cell radius with preamble format 4 is about 1.4km • Restrictions when using preamble format 4: • Only root sequences 0-137 are allowed (rootSeqIndex 0….137) • Allowed values of prachCS are 4...6. Setting prachCS =6 gives the maximum cell range <1.4km

• prachFreqOff must be set to 0 with preamble format 4 • prachHsFlag must be set to false with preamble format 4 Company Confidential 20 © Nokia Siemens Networks


PRACH Configuration Index prachConfIndex (FDD)

Extract of the random access preamble configurations table (only for supported preamble formats 0 and 1)

• The parameter defines the Allowed System Frame for random access attempts, the Sub-frame numbers for random access attempts and the Preamble format

• Supported values in RL10 up to RL30: – For Preamble Format 0: 3 to 8 – For Preamble Format 1: 19 to 24 • RACH Density indicates how many RACH resources are per 10ms frame. • Only RACH density values of 1 and 2 are supported up to RL30.E.g. – RACH density=1 Only one random access attempt per frame – RACH density=2 Two random access attempts per frame

Recommendation: Configure the same PRACH configuration Indexes at cells belonging to the same site. E.g.:  3 or 4 or 5 if RACH density=1 and 6 or 7or 8 if RACH density=2 (Preamble Format 0)



PRACH Configuration Index prachConfIndex (TDD) Supported values in RL15TD: • Preamble Format 0: 3 to 7 • Preamble Format 1: 23 to 25 • Preamble Format 2: 33 to 35 • Preamble Format 4: 51 to 53 Restrictions: • If tddFrameConf=1 no limitation for prachConfIndex • If tddFrameConf=2, prachConfIndex is limited to 3, 4 and 6 or 51 to 53 • If tddSpecSubfConf is set to ‘7’ (ssp7), prachConfIndex is restricted to 3…7 or 51…53 • If tddSpecSubfConf is set to ‘5’ (ssp5) prachConfIndex is not restricted

Recommendation: Configure the same PRACH configuration Indexes at cells belonging to the same site. E.g.:  3 if tddFrameConf=1 and Preamble Format 0 Company Confidential 22 © Nokia Siemens Networks


RACH Density • Based on the expected RACH procedures per second and the maximum collision probability of the RACH preambles it is possible to estimate the RACH density as follows:

• Recommendation: use PRACH density 1 for start

• Since PRACH performance measurement counters are available it will be possible to evaluate the amount of PRACH / RACH procedures in time and adapt /optimize the settings

• Features: RL30 PRACH Management SON feature: an aspect of this feature is to adjust the PRACH density to the traffic in the cell -> not in RL30 Company Confidential 23 © Nokia Siemens Networks


x

UE pcoll

(ex _ Rach _ Load )  p UE  (1) * 64 * ln 1  coll *100 100   = maximum collisiion probability [%]

Ex-RACH_Load = expected RACH Procedures per sec 0.5 ≤ x 0.5 < x 1
≤1 ≤2 ≤3 ≤5

=> RACH Density = 0.5 => RACH Density = 1 => RACH Density = 2 => RACH Density = 3 => RACH Density = 5 => RACH Density = 10

PRACH Frequency Offset prachFreqOff • Indicates the first PRB available for PRACH in the UL frequency band • PRACH area (6 PRBs) should be next to PUCCH area either at upper or lower border of frequency band to maximize the PUSCH area but not overlap with PUCCH area • Parameter is configured based on the PUCCH region i.e. its value depends on how many PUCCH resources are available. • If PRACH area is placed at the lower border of UL frequency band then: PRACH-Frequency Offset= roundup [PUCCH resources/2]

•

If PRACH area is placed at the upper border of the UL frequency band then: PRACH-Frequency Offset= NRB -6- roundup [PUCCH resources/2]

NRB: Number of Resource Blocks

• TDD specific: prachFreqOff =0 when preamble format 4 is used Company Confidential 24 © Nokia Siemens Networks


PRACH Cyclic Shift PrachCS • PrachCS defines the configuration used for the preamble generation. i.e. how many cyclic shifts are needed to generate the preamble • PrachCS depends on the cell size – Different cell ranges correspond to different PrachCS • Simplification: To assume all cells have same size (limited by the prachConfIndex)

Recommendation: Select PrachCS based on the cell range E.g. if estimated cell range is 15km then PrachCS: 12 If all cells in the network are assumed to have same cell range them PrachCS is the same for the whole network



PrachCS and rootSeqIndex Formats 0, 1 and 2 • PrachCS defines the number of cyclic shifts (in terms of number of samples) used to generate multiple preamble sequences from a single root sequence • Example based on PrachCS=12 -> number of cyclic shifts: 119 – Root sequence length is 839 so a cyclic shift of 119 samples allows ROUNDDOWN (839/119)= 7 cyclic shifts before making a complete rotation (signatures per root sequence)

• 64 preambles are transmitted in the PRACH frame. If one root is not enough to generate all 64 preambles then more root sequences are necessary – To ensure having 64 preamble sequences within the cell it is necessary to have ROUNDUP (64/7)= 10 root sequences per cell Company Confidential 26 © Nokia Siemens Networks


Preamble formats 0 ,1 and 2

PrachCS and rootSeqIndex TDD specific (Format 4) • PrachCS defines the number of cyclic shifts (in terms of number of samples) used to generate multiple preamble sequences from a single root sequence • Example based on PrachCS=6 -> number of cyclic shifts: 15 – Root sequence length for preamble 4 is 139 so a cyclic shift of 15 samples allows ROUNDDOWN (139/15)= 9 cyclic shifts before making a complete rotation (signatures per root sequence)

• 64 preambles are transmitted in the PRACH frame. If one root is not enough to generate all 64 preambles then more root sequences are necessary – To ensure having 64 preamble sequences within the cell it is necessary to have ROUNDUP (64/9)= 8 root sequences per cell

Preamble format 4 Company Confidential 27 © Nokia Siemens Networks


PRACH Cyclic Shift rootSeqIndex (FDD) • RootSeqIndex points to the first root sequence to be used when generating the set of 64 preamble sequences. • Each logical rootSeqIndex is associated with a single physical root sequence number. • In case more than one root sequence is necessary the consecutive number is selected from the 838 available until the full set is generated

Recommendation: Use different rootSeqIndex across neighbouring cells as a mean to ensure neighbour cells will use different preamble sequences Company Confidential 28 © Nokia Siemens Networks


Extract from 3GPP TS 36.211 Table 5.7.2.-4 ( Preamble Formats 0-3). Mapping between logical and physical root sequences. Logical root sequence number

Physical root sequence index (in increasing order of the corresponding logical sequence number)

0–23

129, 710, 140, 699, 120, 719, 210, 629, 168, 671, 84, 755, 105, 734, 93, 746, 70, 769, 60, 779 2, 837, 1, 838

24–29

56, 783, 112, 727, 148, 691

30–35

80, 759, 42, 797, 40, 799

36–41

35, 804, 73, 766, 146, 693

42–51

31, 808, 28, 811, 30, 809, 27, 812, 29, 810

52–63

24, 815, 48, 791, 68, 771, 74, 765, 178, 661, 136, 703

….

…..

64–75

86, 753, 78, 761, 43, 796, 39, 800, 20, 819, 21, 818

810–815

309, 530, 265, 574, 233, 606

816–819

367, 472, 296, 543

820–837

336, 503, 305, 534, 373, 466, 280, 559, 279, 560, 419, 420, 240, 599, 258, 581, 229, 610

PRACH Cyclic Shift rootSeqIndex (TDD) • Same recommendation applies in case of TDD – Use different rootSeqIndex across neighbouring cells means to ensure neighbour cells will use different preamble sequences • Differences are: – rootSeqIndex is limited to 0…137 when preamble format 4 is used – the table for mapping of logical to physical root sequence numbers:

Extract from 3GPP TS 36.211 Table 5.7.2.-5 ( Preamble Formats 4). Mapping between logical and physical root sequences.



PRACH Planning Wrap Up Steps: - Define the prachConfIndex • Depends on preamble format (cell range) • It should be the same for each cell of the network - Define the prachFreqOff • Depends on the PUCCH region • It can be assumed to be the same for all cells of a network (simplification) - Define the prachCS • Depends on the cell range • If for simplicity same cell range is assumed for all network then prachCS is the same for all cells - Define the rootSeqIndex • It points to the first root sequence (838 sequences for FDD and 138 possible for TDD) • It needs to be different for neighbour cells across the network • rootSeqIndex separation between cells depends on how many are necessary per cell (depends on PrachCS) Company Confidential 30 © Nokia Siemens Networks


Exercise

• Plan the PRACH Parameters for the sites attached in the excel

• Assumptions: – PUCCH resources = 7 – Cell range = 5 km (all cells have same range) – One PRACH opportunity for 10ms – 20MH BW – FDD



PRACH Management Feature (LTE 581) RL30 and RL25TD • Automatic assignment of PRACH parameters during the initial eNB autoconfiguration process using NetAct Optimizer (i.e. PRACH auto planning): • prachCS • prachConfIndex • rootSeqIndex • prachFreqOff • Assignment done for all cells of an eNB considering own cell data and configuration data from ‘surrounding’ eNBs Feature delimitation • No PRACH / RACH optimization Based e.g. on counter or PM counter results • In RL30 runs only once during initial auto-configuration process: only new eNBs in planned state can use it . It is not possible for actual (upgraded) RL30 eNBs Benefit • No manual PRACH planning for new eNBs/cells required More info: https://sharenet-ims.inside.nokiasiemensnetworks.com/Overview/D433080674 Company Confidential 32 © Nokia Siemens Networks


UL Reference Signal Planning - UlseqHop - UlGrpHop - grpAssigPUSCH - ulRsCs - Sequence Group Number (u)



UL Reference Signal Overview

UL DM RS allocation per slot for Normal Cyclic Prefix

Types of UL Reference Signals • Demodulation Reference Signals (DM RS) – PUSCH/PUCCH data estimation • Sounding Reference Signals (SRS) – Mainly UL channel estimation UL (RL40) DM RS is characterised by: • Sequence (Zadoff Chu codes) • Sequence length: equal to the # of subcarriers used for PUSCH transmission • Sequence group: ▪ 30 options ▪ Cell specific parameter

• Cyclic Shift: UE and cell specific parameter Company Confidential 34 © Nokia Siemens Networks


UL DM Reference Signal Need for Planning Issue: • DM RS occupy always the same slot in time

UL DM RS allocation per slot for Normal Cyclic Prefix

domain • In frequency domain DM RS of a given UE occupies the same PRBs as its PUSCH/PUCCH data transmission • Possible inter cell interference for RS due to simultaneous UL allocations on neighbour cells – No intra cell interference because users are separated in frequency – Possible inter cell interference

Scope of planning: • DM RS in co-sited cells needs to be different TDD case: Since sites are frame synchronised cells should be planned as if they were sectors of the same site. Same recommendation as for FDD applies. Company Confidential 35 © Nokia Siemens Networks


RS Sequences and RS Sequence Groups Sequence Group Id, ‘u’

• RS sequences for PUSCH have different lengths depending the UL bandwidth allocated for a UE • 30 possible sequences for each PRB allocation length of 1-100 PRBs • Sequences are grouped into 30 groups so they can be assigned to cells • Sequence group number ‘u’:

u  PCI  grpAssigPU SCHmod 30



grpAssigPUSCH: group assignment for PUSCH Range [0…29], step 1

Cyclic Shift •

• •

Additional sequences can be derived from a basic sequence by applying a cyclic shift Cyclic shifts of a Fourier transform of an extended ZC sequence are fully orthogonal The actual UL reference signal cyclic shift ncs used by UE is different for every 0.5ms time slot





(1) ( 2) ncs  nDMRS  nDMRS  nPRS (ns ) mod 12

Cell-specific static cyclic shift defined by LNCEL/ulRsCs and broadcast on BCCH ulRsCs 0 1 2 3 4 5 6 7

ndmrs1 0 2 3 4 6 8 9 10


TTI-specific cyclic shift signalled to UE on PDCCH DCI0 in each uplink scheduling grant (defined by scheduler) DCI0 CS field 000 001 010 011 100 101 110 111

Pseudorandom cyclic shift offset that changes every time slot. Depends on the PCI, slot number ns and u via LNCEL/grpAssigPUSCH

ndmrs2 0 6 3 4 2 8 10 9


cinit

cell  N ID     32  u  30 

UL DM Reference Signal Hopping Techniques • Sequence Hopping – Intra-Subframe hopping between two sequences within a sequence group for allocations larger than 5PRBs – Only enabled if Sequence Group hopping in disabled – Not enabled in RL10/RL20/RL30: ulSeqHop= false

• Sequence Group Hopping – In each slot, the UL RS sequences to use within a cell are taken from one specific group – If group varies between slots: Group hopping – Group Hopping not enabled in RL10/RL20/RL30: UlGrpHop = false ▪ Group is the same for all slots

• Cyclic Shift Hopping – Always used – Cell specific cyclic shift added on top of UE specific cyclic shift Company Confidential 39 © Nokia Siemens Networks


Planning

From Theory to Practice… (1/2) Theory: • It should be possible to assign to the cells of one site the same sequence

group ‘u’ and ‘differentiate’ the sequences using different cell specific cyclic shifts i.e. allocating different ulRsCs

Remember!: Cyclic shifts of a Fourier transform of an extended ZC sequence are fully orthogonal Company Confidential 40 © Nokia Siemens Networks


Planning

From Theory to Practice… (2/2) PCI Practice: 75

grpAssigPusch 0 29

sequence id u 15 15

ulRsCs 0 4

76 • It doesn’t seem to work • UL Throughput gets considerably affected if UL traffic in neighbour cell

– From 40 Mbps to ~ 22 Mbps in the example



cinit 79 79

Planning New rule • Allocate different sequence group u for every cell, including cells of the same site – Cross-correlation properties between sequences from two different groups are good because of sequence grouping in the 3GPP spec

• ulRsCs does not matter (it is only relevant for sequences within one seq group u)



Planning Results • UL Throughput still suffers from UL interference in neighbour cell but the effect is lower

PCI 75 76 Company Confidential 43 © Nokia Siemens Networks


grpAssigPusch 0 0

sequence id u 15 16

ulRsCs 0 0

cinit 79 80

Pros an cons of the ‘new’ planning rule • [+]: Results seem to be better • [+]: Less parameters to plan, only PCI planning needed – UlRsCs only relevant when using sequences of the same group – ‘u’ will be different if PCI modulo 3 rule is followed. In that case ‘grpAssigPUSCH’ value is not relevant

u  PCI  grpAssigPU SCHmod 30

• [-]: Reduced group reuse distance compared to the case of assigning the same group per each site



UL DM RS Planning Wrap up • Principle: DM RS needs to be different in cells of the same eNodeB • Current planning approach: – Assign different sequence group number ‘u’ to the cells of the same site. Range: [0…29]. grpAssigPUSCH can be constant =no need for planning u  PCI  grpAssigPU SCHmod 30

– If cells of the site follow the PCImod3 rule, the sequence group number ‘u’ will be different

– If PCImod3 rule is not followed, check PCImod30 rule ▪ If problems use grpAssigPUSCH to differentiate the ‘u’ - sequence group number-

– If same ‘u’ has to be used in neighbouring cells and cannot be changed using

grpAssigPUSCH then assign different ulRsCs to the cells of a site. Range [0…7]



UL DM RS Planning example •

Using grpAssigPUSCH to tune PCI based sequence allocation in case of PCImod30 collision Delta_ SS = grpAssigPUSCH PCI = 3 Dss = 0 u=3

PCI = 30 Dss = 29 u = 29

PCI = 9 Dss = 0 u=9

indoor eNB

If grpAssigPUSCH=0 then u=0 interfering with the cell below. grpAssigPUSCH is used to avoid this PCI = 2 Dss = 0 u=2

eNB #2 PCI = 0 Dss = 0 u=0

PCI = 4 Dss = 0 u=4

PCI = 5 Dss = 0 u=5

PCI = 11 Dss = 0 u = 11

PCI = 6 Dss = 0 u=6

eNB #1 PCI = 1 Dss = 0 u=1

PCI = 10 Dss = 0 u = 10 PCI = 12 Dss = 0 u = 12

eNB #5

eNB #3

PCI = 8 Dss = 0 u=8 Company Confidential 46 © Nokia Siemens Networks

eNB #4


PCI = 7 Dss = 0 u=7

PCI = 14 Dss = 0 u = 14

PCI = 13 Dss = 0 u = 13

Tracking Area Planning



Introduction (1/2) • When the UE is in idle mode its location is known by the MME with the accuracy of a tracking area

• • • • •

Each eNodeB can contain cells belonging to different tracking areas One cell only belongs to one tracking area code (TAC) A tracking area can be shared by multiple MME Tracking Area Identity (TAI) = PLMN ID (mcc, mnc) + TAC all broadcasted in SIB1 Reserved TAC values: 0000 and FFFE( in hex) i.e. 0 and 65534 S1 Application Protocol Paging Message extracted from 3GPP TS 36.413

Tracking areas are the equivalent of Location Areas and Routing Areas for LTE



Introduction (2/2) • The normal tracking area updating procedure is used when a UE moves into a tracking area within which it is not registered

• The periodic tracking area updating procedure is used to periodically notify the availability of the UE to the network (based upon T3412)

• Tracking area updates are also used for • registration during inter-system changes • MME load balancing

Further details in 3GPP TS 24.301

• Large tracking areas result in • Increased paging load • Reduced requirement for tracking area updates resulting from mobility



Planning Guidelines • Tracking areas should be planned to be relatively large (100 eNodeB, 3 cells/eNodeB) rather than relatively small

• Their size should be reduced subsequently if the paging load becomes high • Tracking areas should not run close to and parallel to major roads nor railways. Likewise, boundaries should not traverse dense subscriber areas

• Cells which are located at a tracking area boundary and which experience large numbers of updates should be monitored to evaluate the impact of the update procedures

• Existing 2G and 3G location area should be used as a basis for defining LTE tracking area boundaries (?: see next slide)



Radio network configuration recommendations Extracted from CSFB Training Material:

https://sharenet-ims.inside.nokiasiemensnetworks.com/Open/438643378

• NSN recommends to use different Location Areas Identities in LTE (4G) access than in the 2G or 3G. The recommendation is e.g. for the following reasons: – MSS pooling concept requires that LTE (4G) Location Area identities are separated from 2G and 3G Location Area Identities. – When the 2G, 3G and LTE(4G) uses overlapping Location Area Identities, and when the CSFB is made to same MSS/VLR in which the LTE terminal is registered, the SGs association remains active in MSS/VLR after CSFB is made. It causes for a short time period after CSFB call is ended, that the LTE terminal is not reachable via SGs interface , because of many CSFB capable LTE terminals do not to listen LTE (4G) radio while camping in 2G or 3G radio. ▪ CSFB MSC Server is able to paging over the A/Iu interface in case paging over the SGs fails (terminal is hanging in 2G/3G after CSFB call and new MT call is coming). This cause some delay to call setup time. ▪ When the 2G/3G and LTE (4G) Location Area Identities are different, LTE terminal would be forced to initiate Location Update procedure always when changing the radio access from 2G or 3G to LTE (4G) and vice versa. With this concept, LTE terminal would be always reached in the current location without any delay.

 Summary: with this recommended concept, LTE terminal would be always reached in the current location without any delay. Company Confidential 51 © Nokia Siemens Networks


Tracking Area Lists •

A UE can be registered in multiple tracking areas to avoid unnecessary tracking areas updates at the tracking area borders. This is done via the TA list i.e. a list of allowed TA delivered to the UE in the attach and TAU procedures

•

TA list can contain a maximum of 16 different tracking area identities (TAI)

•

MME supports maximum 8000 TA lists

•

The TA list is configured in the MME TA1LSTNX.xml file

•

If the same TAI belongs to multiple TA Lists. The MME will send to the UE (during attach or TAU) the TA List with lowest value



Example of TA1LSTNX.xml file showing two TAI

•

Unclear if TA List configuration is radio planning or EPC task

Tracking Area Planning RAN sharing case •

In case of RAN sharing, recommendation of re-using existing LA from 3G/2G is not valid as TAC is the same for all PLMN Ids

• LNCEL: tac has multiplicity one i.e. no multiple entries possible • LNCEL: furtherPlmndIdL allows up to 5 entries • Together with primary PLMN ID (LNBTS: mcc, mnc & mncLength) there can be up to 6 PLMN Ids)

• Feature RAN sharing Multi Operator Core Network (MOCN-LTE4) currently supports only 2 PLMNs • Planned Feature RL50 (LTE1051) will support up to 6 operators MOCN



PCI RACH - Planning_Topics

Recommend Documents