1- What are the main KPI to measure the performance of 3G cell - Accessibility ( RRC , RAB , CSSR) - Retainability (speech , Video , PS DCR) - Mobility (SHO , IRAT HO success rate) 2- What resources affect HSDPA throughput in 3G system - (DL power, DL code and transport capacity) 3- What parameter tuning can be done to improve HSDPA throughput in any 3G cell - increase the DL channelization codes for HSDPA - changing the scheduling algorisms 4- How can we reach 21 Mbps in P7 - 15 codes in DL and 64 QAM 5- what is the usage of the following signaling messages in RRC protocol - Actives setup updates (ADD/Remove/Replace RL in SHO) - RB reconfiguration (channel switching between Cell_DCH and Cell_FACH RRC stats) - Physical channel reconfiguration (IF HO) 6- what is the use of GPEH tool in Ericsson system - tool used to record RAN and internal events in Ericsson system and the tracing files can be analyzed by TEMS visualization 7- what types of congestion can affect the services accessibility in any 3G cell - DL power ( AMR - Directed retry - reducing High R.99 RAB users SFxx parameters) - UL/DL CE ( reducing High R.99 RAB users SFxx parameters) - DL code (reducing static codes for HSDPA –AMR-Directed retry) - Transport capacity 8- what is the difference between RSCP and EC/No measures for pilot channel - RSCP is received signal code power for CPICH channel - Ec/No is The received energy per chip divided by the power density in the band . it reflects the quality of CPICH channel 9- what is the difference of using 2nd carrier and high power amplifier in expanding the capacity for any 3G cell - 2nd carrier gives capacity in DL power and DL codes
- High power Amplifier gives capacity in DL power only 10- what is the max bit rate that can be achieved in UL when using 10ms EUL and 2 ms EUL - 1.5 Mbps for 10 ms EUL - 5.76 for 2 ms EUL 11- how many HSSCCH channel can be configured in HSDPA cell ( - Four that allows four users per TTI
3G KPI’s and optimization: 1. LOW CSSR: 2. When a low CSSR is detected on a cell the first thing to check is if Admission Control is rejecting the RRC/RAB setup attempt (pmNoReqDeniedAdm) or if it is failing after admission (pmNoFailedAfterAdm). For high pmNoReqDeniedAdm refer to the “Admission Control” sections below. For high pmNoFailedAfterAdm refer to the “Failure After Admission” sections below. Causes: Admission Control: DL Power 1. Long term solutions are to increase the power capability of the sector 2. re-engineering the site to reduce feeder lengths 3. The short term solution is to reduce the traffic carried by the site Admission Control: DL Channelisation Codes 1. This will typically affect the PS Interactive R99 (DCH/FACH) CSSR worse than the Speech CSSR as the PS Interactive R99 RAB requires channelisation codes at a lower spreading factor (using more of the code tree). 2. In the P4 software release a cell that supports R99 and HSDPA typically has 5 spreading factor 16 DL channelisation codes reserved for HSDPA. This means that approximately 32% of available codes are reserved for HSDPA. When this is the case it is common for DL channelisation code congestion too occur 3. The long term solution is to add another cell in the coverage area to take some of the traffic; this may be achieved by introducing a second carrier, another sector, or another site 4. The short term solution is to reduce the traffic carried by the site
5. the required solution is sectorisation of the inbuilding antenna system or implementation of a second carrier frequency Admission Control: Hardware Usage (Channel Elements) 1. too few channel elements available 2. The channel element capacity of an RBS may be software limited (according the software license configured for the RBS) or hardware limited (according to the TXBs and RAXBs installed in the RBS). 3. Always compare channel element usage to the channel element capacity of the Wcell
4. LOW DCR: 1. If a cell has poor retainability it is typically due to either 2. missing neighbour definitions (WCDMA and/or GSM) 3. overshooting cell(s), PILLOT pollution. 4. a misbehaving neighbour site 5. a hardware/software fault or a misconfiguration. 6. It is also possible that there is some external source of interference (such as a
microwave link on the same frequency) affecting the retainability.
However, in the majority of cases the factors that affect the Speech retainability will also affect the retainability of the other RABs. When a high speech DCR is detected on a cell the first thing to check is the type of drops occurring as indicated
Soft Handover Drops Soft HO failures due to missing neighbour or Two common reasons are a neighbouring cell that is misbehaving (often due to faulty hardware/software) or a misconfiguration resulting in a failure to perform an inter-RNC SOHO across the Iur interface Missing Neighbour Drops UL Synchronisation Drops 1. missing IRAT neighbour relation definitions resulting in the connection “hanging on” to the 3G network until the call is dropped when it would be better served
handing the call over to the 2G network. This may be especially true for cells on the border of the 3G coverage area 2. By identifying such areas any missing 2G neighbour relations may be added; or perhaps a misconfiguration discovered, such as having an IRAT neighbour relation defined in an RNC towards a 2G cell that is not defined as an outer cell in the 3G MSC Server 3. . Another means to improve the situation may be to lower the thresholds used to trigger IRAT HOs 4. For example, triggering compressed mode at Ec/No=-11dBm instead of -12dBm may prevent drops as calls are handed earlier to the GSM network that typically has better coverage than the 3G network 5. . It is recommended to resolve any SOHO and missing neighbour drop problems before attempting to resolve UL sync drops as often the cause of such drops will resolve UL synchronisation drops too. 2. Other Reasons for Drops
If none of the above reasons for a poor DCR may be established, then it is likely to be a more complicated problem to resolve; often relating to a software/hardware fault, or perhaps an external source of interference in the area
HSDPA Considerations for HSDPA: DL Channelisation Codes 1. i.e. ∼32% of the available downlink channelisation codes are reserved for HSDPA 2. feature called HSDPA Dynamic Code Allocation making it possible for the HSDPA Scheduler to only use the channelisation codes available after R99 usage 3. This may help with downlink channelisation code congestion, but at the expense of reduced throughput for HSDPA users. It is still possible to reserve a limited set of codes for HSDPA 3. Considerations for HSDPA: Iub Bandwidth
1. the capacity of the Iub interface between RBS and RNC becomes an important factor effecting HSDPA performance and customer perception
2. i.e. across the Iub interface the R99 traffic has priority over the HSDPA traffic 3. As the R99 traffic on the site increases, so the Iub bandwidth available for HSDPA decreases 4. To alleviate the problem some traffic may be offloaded to GSM 5. but ultimately additional Iub bandwidth is required unless an additional site, perhaps an inbuilding site at a shopping centre, may be commissioned to carry part of the load.
HSDPA throughput counters
Overall Flowchart for Capacity Management
IRAT cell reselection: From 2g to 3g: QsearchI(VOICE)/P(GPRS): 7 enabled ( MS will look for 3G neighbours
measurements) FddQualmin= -10 EcNo (threshold: meaning if 3G cell is becoming better with good
quality .. then reselect to 3G cell—lower values (-8,-6) means that reselection will be delayed as mush better quality is required on 3G side.)
From 3G to 2G: qQualmin/ Qrxlevmin: -16/ -109dBm (The minimum required quality level in the cell
(Ec/No).) SsearchRAT=2-4dB
GSM measurements will happen when it is not possible to maintain call with good quality: CPICH EcNo < Ssearch_RAT + Qqualmin (default: EcNo < -14 dB)
IDLE mode trials: 1) Improved DCR a. Qrxlevmin: -115dBm à -109dBm b. Qqualmin: -18dB à -15dB c. (SsearchRAT=2dB; fddQmin=-12dB) (on border cells) EcNo < -10dB 2) Objective: Improve the RRC Connection Access Complete Ratio, and as a result the CSSR, by improving the required radio coverage conditions at the edge of coverage. The initial synchronization will be made easier with a better coverage. Also the possibility of inter-system ping-pong will be reduced with a higher hysteresis between technologies.
3) RRC Connection Access Failures on LAC-border cells decreased AdjsQOffset2=2dB and 4dB has been applied on the ADJS at the LAC border.The default was 0dB
The objective of the tested parameter changes is to reduce the amount of Location Updates, and to ensure that they are performed in better radio conditions in the target cell. Two parameter values will be tested. Statistics monitored: RRC Connection Access Complete Ratio, Amount of RRC Connections due to registration and inter-RAT. The RRC Completion Ratio might also improve.
4) Soft Handover Parameters for NRT Objective: Reduce the Soft Handover Overhead for Non-Real Time Services, and as a consequence reduce the load and congestion occurrence
Statistics monitored: Soft Handover Overhead, DCH Setup Failure due to Iub AAL2 Trans for PS data background, Background RAB Completion Ratio, PS Backg Allocated DL
•
•
In Lugano, the tested change decreased the Soft Handover Overhead for NRT, without degrading the RAB Completion ratio •
Addition Window=4dB à 2.5dB
•
Drop window=6dB à 4dB
In Biel there is a very small decrease •
•
The RAB Setup Failures were probably due to RNC issues, they have appeared with RAN04 and disappeared with CD2.1
Proposal •
Create FMCS #6 with Add=2.5dB / Drop=4dB on all RNCs
•
Apply NrtFmcsId=6 to de-congestion sites or areas
5) Trial 4: Downlink traffic volume measurement high threshold Objective: Reduce the number of 128kbit/s à 384 kbit/s upgrades for little data volumes. Avoid inefficient use of Iub and WSP capacity. Duration: 2 weeks Statistics monitored: PS Backg Allocated traffic for 128 kb/s and 384 kb/s in DL, Radio Bearer Reconfigurations, NRT DCH Requests in DL Changes: o
Downlink traffic volume measurement high threshold (TrafVolThresholdDLHigh): 1024 (1KB) à 4096 (4KB)
HO Between 2G and 3G:
From 2g to 3g: QsearchC: 7 enabled ( MS will look for 3G neighbours measurements in dedicated
mode)
minimum CPICH Ec/Io level (MET) = -8—10 db (threshold: meaning if 3G cell is becoming
better with good quality .. then HO to 3G cell—lower values (-8,-6) means that HO will be delayed as much better quality is required on 3G side). With this parameter you define the minimum CPICH Ec/Io level of an adjacent WCDMA RAN cell for an inter-system handover attempt. The threshold level must be exceeded before the BSC is allowed to trigger a handover attempt towards the adjacent WCDMA RAN cell. From 3G to 2G: HHoRscpThreshold =-105 (Normal)/ -100 (Fast ISHO) HHoEcNoThreshold=-12 (Normal)/ -10 (Fast ISHO)
The parameter HHoRscpThreshold determines the absolute CPICH RSCP threshold which is used by the UE to trigger the reporting event 1F. When the measured CPICH RSCP of all active set cells has become worse than or equal to the threshold in question, the RNC starts inter-frequency or inter-RAT (GSM) measurements in compressed mode for the purpose of hard handover.
dU setting:
Related Parameters:
For RSCP: HHoRscpThreshold Related parameters HHoRscpTimeHysteresis HHoRscpCancel HHoRscpCancelTime GSMcauseCPICHrscp IFHOcauseCPICHrscp
-105 (Normal)/ -100 (Fast ISHO) 100ms/1280ms(dU) -102 dBm 640ms/ 320(dU) enabled/disabled enabled/disabled
For EcNo: HHoEcNoThreshold Related parameters
=-12 (Normal)/ -10 (Fast ISHO)
HHoEcNoTimeHysteresis
640ms/640ms(dU) -9 dB 100ms/320ms(dU) enabled/disabled enabled/disabled
HHoEcNoCancel
HHoEcNoCancelTime GSMcauseCPICHEcNo IFHOcauseCPICHEcNo
HHoRscpTimeHysteresis=100ms •
The parameter HHoRscpTimeHysteresis determines the time period during which the CPICH RSCP of the active set cell must stay worse than the threshold HHoRscpThreshold before the UE can trigger the reporting event 1F.
GSMcauseCPICHrscp of FMCG •
This parameter indicates whether a handover to GSM caused by low measured absolute CPICH RSCP of the serving cell is enabled.
IFHOcauseCPICHrscp of FMCI •
The parameter indicates whether an inter-frequency handover caused by low measured absolute CPICH RSCP of the serving cell is enabled.
HHoRscpCancel = -102 dBm (3dB difference is needed between trigger and cancel
parameters.) If the inter-frequency or inter-RAT (GSM) handover caused by low measured absolute CPICH RSCP is enabled, the RNC starts the inter-frequency or GSM measurement in compressed mode when all active set cells have triggered the reporting event 1F for CPICH RSCP. The RNC cancels the event 1F of an active set cell, if the CPICH RSCP measurement result of the active set cell becomes better than or equal to the threshold HHoRscpCancel and the UE transmits the corresponding event 1E triggered measurement report to the RNC.
HHoRscpCancelTime =640 ms •
The parameter determines the time period during which the CPICH RSCP of the active set cell must stay better than the threshold HHoRscpCancel before the UE can trigger the reporting event 1E.
AdjsEcNoOffset.( CPICH Ec/No Offset) The CPICH Ec/No Offset determines an offset value, which the UE adds to the CPICH Ec/No measurement result of the neighbouring cell before it compares the Ec/No value with the reporting criteria.
AdjsDERR(Disable Effect on Reporting Range)This parameter indicates whether the neighbouring cell is forbidden to affect the reporting range (addition/drop window) calculation, if it belongs to the active set.
ISHO process:
Call setup in 3G
RRC CONNECTION SETUP TIMERS:
1.1
Test Procedures TEST CASES
DRIVE TEST METHODOLOGY
Radio Design Performance Check
Drive Test to be done using two UEs and a Scanner. (MS1-in idle Mode and; MS2 –in Dedicated Mode). Both UEs are Locked to 3G Only.
Coverage and quality performance check
1. Coverage performance of each sector (Best Server RSCP)
Call Performance Check.
Short calls: Static test on each sector. Call duration is 20 seconds and 10 seconds idle.
2. Quality performance of each sector(Best Server EcNo)
Long Calls: Drive around within the coverage objective of each sector.
Handover Performance Check
Handover performance to be verified by driving in clock wise and anti clock wise around the Site, for all three sectors. Then perform handover test towards its surrounding neighbors (at least 3 neighbors)
Data Performance Check
R99 & HSPA test for the following: 1. PDP context activation (5 attempts) 2. Throughput test (based on 2MB downloaded file from the FTP server.
Supplementary Services Check
3 SMS/MMS per cell at different locations with 10 sec break after each SMS/MMS
Acceptance Criteria:
Using R99 384kbps bearer, average FTP DL throughput should not be less than 340kbps. Using HSDPA bearer, average FTP throughput should not be less than 3Mbps (Without transmission limitation). Test method: The data transfer will be from a test FTP server. The file format should be a txt or data file. Acceptance Criteria: The SHO between the sectors of same site should be successful The SHO between the sites should be successful. Site acceptance performance test should be performed in static mode and in good RF conditions, CPICH_RSCP > -80dBm & CPICH_EcNo > -8dB Acceptance Criteria: The scramble configuration is consistent with the planned configuration. No cross TX feeder observed.
1.2
Drive Test Route The drive to be performed considering that the Scrambling code of the desired WCDMA cell is the primary server for most of the time.
Sector 1
Coverage observation
Quality observation
The average coverage within 800m is about -75dbm. No obstruction observed
The average Ec/No is larger than -10
Sector 2
The average coverage within 1Km is about -75dbm. No obstruction observed
The average Ec/No is larger than -10
Sector 3
The average coverage within 1Km is about -75dbm. No obstruction observed
The average Ec/No is larger than -10
Coverage performance – Scanner mode
Acceptance Criteria: If no special note is for a cell, the related CPICH RSCP should be larger than –70dBm within 100m from the Node B.
1.3
Quality performance – Scanner Mode (EcIo)
Acceptance Criteria: If no special note is for a cell, the related CPICH Ec/Io should be larger than –10dB within 100m from the Node B.
1.4
Coverage Performance – Dedicated Mode (CPICH RSCP)
Acceptance Criteria: If no special note is for a cell, the related CPICH RSCP should be larger than –75dBm within 100m from the Node B.
1.5
Quality Performance – Dedicated Mode (Ec/No)
Acceptance Criteria: If no special note is for a cell, the related CPICH Ec/Io should be larger than –10dB within 100m from the Node B.
1.6
Active Set Size Plot
Acceptance Criteria: The plots of Active Set (AS) size=3 should not be more than plots of Active Set size=2 or 1.
EVENTS:
hese different events that a UE reports.. like: Event 1x: related to intra freq events. Event 2x: related to interfreq measurement Event 3x: related to inter-RAT measurement where x can be a, b, c, d, e, f these are the 3 important categories... There description is as follow:
Event 1A:
"A Primary CPICH enters the reporting range", it is defined by the EcNo or can be RSCP lev; depends on the critera whether it is EcNo or RSCP. In du case it is EcNo and parameter related to event 1a is "ADDITION WINDOW" which is set to be either 3db or 4 db here. Simply we can say that addition of a cell in active set is triggered by event 1a.
Related Parameters EVENT 1A: • AdditionWindow • AdditionTime • AdditionReportingInterval • ActiveSetWeightingCoefficient • MaxActiveSetSize are used in case of event 1A
(4 dB default) (100ms: 6 internal value) (0.5 sec : 2 internal value) (0)4 (3)
Event 1B: "A primary CPICH leaves the reporting range" similar to the above but is for deletion of a cell from active set, Drop Window is set to be either 5 or 6db here.
Related Parameters EVENT 1B: • DropWindow • Drop Time • ActiveSetWeightingCoefficient
(6 dB) (640 ms: 12 internal value) (0)
are used in case of event 1B.
Event 1C: "A non-active primary CPICH becomes better than an active primary CPICH" it is basically for replacement of a CPICH in AS (active set). AS size is 3 which means Max no of CPICH in soft hand over can be 3. Now suppose there are 3 cells in AS and a 4th pilot becomes stronger than any of the 3 in AS, only replacement can be a solution which is triggered by event 1c. This is taken care by a 2db replacement window and of course hyst as well.
Related Parameters EVENT 1B: • ReplacementTime • ReplacementWindow • ReplacementReportingInterval are used in case of event 1C.
(100ms: 6 internal value) (2dB: default) (0.5 sec:2 internal value)
Event 1D: Change of best cell Event 1E: A Primary CPICH becomes better than an absolute threshold...it is basically related to cancellation of Hard handover..related params in nokia are HHOEcNoCancel or if criteria is RSCP then HHORSCPCancel. Event 1F: "A Primary CPICH becomes worse than an absolute threshold"....basically triggers HHO to GSM in nokia ( wheras huawei is using event 2D to do the same)...related params are HHOEcNoThreshold or HHORSCPThreshold...if event 1f is reported by allcells in AS, UE goes into compressed mode and if all criteria met..triggers HHO to GSM. So this is about event 1 description....all these events can be seen/analyzed in layer 3 messages...
Huawei Interview questions: Kind of question asked by Huawei during interviews:
• • HSPDA : High speed downlink packet access (HSDPA), it is like in GSM we have edge/gprs for data service. In 3G HSPA (HSDPA, HSUPA) do provide high speed data. • 2: HSUPA: High speed uplink packet access, it provides high speed PS data in uplink like HSDPA does in the DL.
•
3:BTS
and RNC if possible hardware description ( Boards and elements used):
Huawei is using BTS3900, which is attached here..
• 4:Capacity calculations what is the SHO overhead, how many channel elements will be needed, how big will be the Iub, based on presented traffic model SHO overhead is around 30-40%, 384 CEs are configured on each cell, Iub based interface here is IP based with radios upto 21Mbps per sector.
•
5:Differences
between planning based on coverage and on capacity
Main idea is same as is in 2G, here one has to be more careful to control overshoots to have good EcNo and less or no pilot pollution.
•
6:What
does it mean a noise raise of 3db? 50% coverage reduction
This is Noise rise at Node B ie in the UL, which corresponds to 50% loading of the network. the %age is of pole capacity which is basically theoretical max cap.
•
7:What
is the expected noise floor in uplink and in downlink?
The term noise floor is related to UL, in term of RTWP it is less than -100dbm, with loading noise floor keep rising..
• 8: Site search cycle (from nominal plan site selection, site survey site ranking, final selection, final report, acquisition) Just similar to 2G... • After integration
• 9: Site health checks: (Alarms, site acceptance drive test. Footprint check, SC check) HSDPA & R-99 Throughput, sector swap, hand over tests, IRAT HOs, reselection, functionality test that sites coverage is normal, VSWR and other alarm checks. 10:Explain criteria’s thresholds and process of cluster acceptance: Example 80% of sites radiating without problems Cluster acceptance depends on customers mood:) Generally we have to have CDR less than 2%, RAB set up sucess rate is 98% for all types of RABs
• 11: What sort of checks to do in a Swap case taking in consideration that you need to have the same KPI’s before and after 12:Description of used tools Actix, Tems, M2000 for different reports generation, LMT for changing/config params, U-Net planning tool. Nastar for viewing RNC logs (data) dump and other RNC data can be loaded directly to Nastar and from there all states can be viewed.
• 13: Elaborate on Monte Carlo simulations and on Okomura Hata propagation model Okumara Hata Model is on the same lines as is in 2G or in theory.
Monte Carlo Simulations are tool based analysis to analyze access failures under different loading conditions which are configured before simulations are run.
CHIP: The bits in the spreading code are called chips.
GMSK shifts one bit at the time, QPSK two bits at the time, and 8-PSK three bits at a time
The effective bandwidth for WCDMA is 3.84 MHz, and with guard bands the required bandwidth is 5 MHz. The guard bands are needed to reduce the interference between different 5 MHz WCDMA carriers. Soft HO capacity gain: As a conclusion it can be stated that soft and softer handovers consume radio access capacity because the UE is occupying more than one radio link connection in the Uu interface. On the other hand, the added capacity gained from the interference reduction is bigger and hence the system capacity is actually increased if soft and softer handovers are used.
Codes - What and why? A code is a specific sequence of bits applied to data to scramble the information.
A symbol is an information unit, transmitted via the radio interface. Downlink each symbol represents two bits. One bit of the code signal used for signal multiplying is called a chip. The code signal bit rate, which is hereafter referred to as the chip rate, is fixed in WCDMA, being 3.84 million chips per second (Mcps/s). With this chip rate the size of one chip in time is 1 / 3 840 000 seconds.
Spreading factor K = 2k, For instance, if k = 6, the spreading factor K gets the value 64, indicating that one symbol uses 64 chips in the WCDMA radio path. Another name for spreading factor is processing gain (Gp), and it can be expressed as a function of used bandwidths: The scrambling codes are divided into 512 code sets, each of them containing a primary scrambling code and 15 secondary scrambling codes.
puncturing
In some cases, we are able to perform so-called in case the bit rate is somewhat higher than certain allowed bit rate. Puncturing means that we remove some of the redundant bits from the channel coding, thus reducing the bit rate down to the wanted level.
Rate matching After the error protection, the baseband data rate is matched to the bearer bit rates used in the UMTS radio interface. The data rates are given with the available channelisation codes, resp. with the given spreading factors. In the figure below, you can see different bit rates that may be applied for user data. The bit rates range between 15 and 960 kilobits per second.
Please note that these bit rates are not the same as the application or user data rates. To exemplify this, let us imagine a speech call at 12.2 kilobits per second. These bits must undergo channel coding in order to enable error correction. After this channel coding (convolutional coding), the bit rate is around 24 kbit/s. Now we need to rate match this data to the closest allowed bit rate in UMTS, which in this case is 30 kbit/s (as you can see from the figure above). We repeat some of the encoded data bits according to a certain pattern in order to increase the bit rate to the wanted level.
UMTS channel structure
Radio Resource Control RRC states
When we have a dedicated channel open for a subscriber (for example, if we are using video), we say that subscriber is in Cell_DCH state. (The DCH is derived from the name of the channel in the air interface). In this state the UE is sending measurement reports to the network, thus the system can control the dedicated bearer and perform handovers. If the mobile is only sending small pieces of information, for example intermittent Internet based traffic or for signalling, then the RRC can be in a mode known as Cell_FACH (the FACH stands for Forward Access Channel) and is different from the previous state as no dedicated channel is used. The network does not perform handovers as the mobile moves from one cell to another. The UE just informs the
network of its current location. Depending on the bearer we have and how it is being used, the RNC will move the RRC between the different states. In addition to the Cell_FACH, if the network finds that the bearer is not being used for a long time, it can move the connection to a Cell_PCH mode (Paging Channel), where the mobile is still know to a cell level but can only be reached via the PCH. In this state the UE is using a Discontinous Repetition Function (DRX) to save battery. Again, unlike in the Cell_DCH, as the subscriber moves, the mobile informs the RNC which cell it has moved to. The final state is the URA_PCH. This state is similar to the Cell_PCH. But, instead of monitoring the connection on a cell level, it is now on a RNC level. URA stands for UTRA Registration Area and the UE monitors the broadcast channel for URA identities.
Admission control The main task of admission control is to estimate whether a new call can have access to the system without sacrificing the bearer requirements of existing calls. Thus the AC algorithm should predict the load of the cell if the new call is admitted. It should be noted that the availability of the terrestrial transmission resources is verified, too, meaning that there is no limiting factor in the rest of the UTRAN either. Based on the admission control, the Radio Network Controller (RNC) either grants or rejects the access.
Planning uplink admission control PRX_Target (Receive Power level) value. The area from 0 to this value is known as the planned load. Once the load is approaching this value, traffic reason handovers (TRHO) are performed.
As UMTS traffic is variable and constantly changing, it is more than feasible that
the traffic admission may exceed the PRX_Target. To handle this situation, there is a second level used called the PRX_TARGET_BS. This is a parameter used by the BTS to stop situations of congestion. Once this value is reached, the BTS takes actions to reduce the load in the cell.
Code allocation Both scrambling and channelisation codes used in the Uu interface connections are maintained by the RNC. In principle they could be maintained by the BTS, but then the system would experience lack of radio resource control, namely soft handovers, which will be explained later in this module. When the codes are maintained by the RNC, it is easier to allocate Iub data ports for multipath connections. The Uu interface requires two kinds of codes for proper functionality. A part of the codes used must correlate with each other to a certain extent, and the others must be orthogonal (they do not correlate at all). Every cell uses one scrambling code. As you already know, this code acts like a cell ID. Under every scrambling code the RNC has a set of channelisation codes. This set is the same under every scrambling code. The BCH information is coded with a scrambling code value, and thus the UE must first find the correct scrambling code value first in order to access the cell. When a connection between the UE and the network is established, the channels used must be separated. The channelisation codes are used for this purpose. The information sent over the Uu interface is spread with a spreading code per channel. Spreading code by definition is the same as scrambling code x channelisation code.
Channelisation code allocation and handovers
Fragmentation of code tree phenomenon where the probability of the blocked branch of the code tree increases too much and thus it starts to prevent new accesses to the system. For example, if an active call uses high bit rate over the Uu interface, the spreading factor value in use is small. It furthermore means that a very high-level branch of the code tree is blocked (see the figure below). When this call is finished and simultaneously new calls access the system, the blocked code tree branch is not “released” before the new accesses. In this situation the system wastes capacity because the code channels allocated for new calls are not necessarily allocated in the best possible way. As codes are released in different branches, the tree can become fragmented and the RNC should always try to reorganise the tree to make the best use of the resources. Therefore in UMTS networks, it is possible that the channelisation codes could change during a connection. Also, if the scrambling code in the uplink (that is, the user) is being used by another person in another RNC as the subscriber performs a soft handover, the
handover is refused and the serving RNC must allocate a new scrambling code to the subscriber.
Power control Inaccuracy in power control immediately increases interference, thus decreasing the capacity of the network.
Open loop power control
When the UE accesses to the network, the initial level for accessing is based on an estimate. This estimate in turn is based on the signal level received from the Node B when the UE is in idle mode. The basis for the UE estimate is the downlink power level that the UE detects from the physical channel CPICH. Closed loop power control
When the radio connection is established, the power control method is changed. During the connection, the method used is called the closed loop power control. Within this method, the Node commands the UE either to increase or to decrease its transmission power with the pace of 1.5 kHz (1500 times per second) in the FDD mode. (In the TDD mode closed loop power control is performed 100 to 800 times a second.) The decision whether to increase or decrease the power is based on the received Signal-to-Interference Ratio (SIR) estimated by the Node B. Outer loop power control
Due to the macro diversity (the UE is simultaneously attached to the network through more than one cell), the RNC must be aware of the current radio link conditions and quality. The RNC knows the allowed power levels of the cell and target SIR.
Packet scheduler Packet scheduler is a general feature, which takes care of scheduling radio resources for non-real-time (NRT) radio access bearers for both uplink and downlink directions. The gap between real-time (RT) traffic and the load target of the cell can be filled by the packet scheduler. An active set is a list of cells, through which the UE has a connection to the network, that is, through which the radio link set-up has been made. This is, the UE may have active radio connection between itself and the network through three cells simultaneously. In soft handover, the UE is connected to (at least) two Node Bs at the same time. In the uplink direction, the two signals come via the base stations to the Radio Network Controller (RNC). In the RNC the signal to be transported forward to the core network is selected. The selection is done frame by frame for the speech, and in smaller blocks for data
slotted/ compressed mode The possibility to perform an inter-system handover is enabled in the UMTS by a special functioning mode, slotted mode. When the UE uses Uu interface in the slotted mode, the contents of the Uu interface frame is “compressed” in order to open a time window, through which the UE is able to peek and decode the GSM BCCH information.
Intersystem handover from GSM
The handover is described in the figure 44. • The UE/MS creates a measurement report that the BSC evaluates to make the handover decision. • If the BSC decides to hand over to a UTRA cell resource reservation messages are sent to the UTRAN. • The UTRAN acknowledges the resource reservation and provides a UTRAN handover command. • The BSC sends the GSM intersystem handover command to the UE. In this command is included a UMTS Handover to UTRAN command which contains all the information needed to set up a connection to the UTRA cell. Since the amount of UTRA configuration information might be too large for a GSM message the message actually contains reference number to UTRA parameters not the real values. • The UE completes the procedure by a Handover to UTRAN complete message to the RNC. • As a last stage the RNC commands resources to be released by the BSC.
Intersystem handover from UTRAN
UMTS has already been deployed widely but there will still be reasons to perform Intersystem handovers to GSM. This could be because of services used, coverage or even traffic reasons.
Based on the measurement report including both UTRAN and BSS values the RNC makes the handover decision. • Resource reservation messages are sent to the BSC. • The BSC acknowledges the resource reservation and includes a GSM handover command. • The RNC sends an Intersystem handover command message to the UE, included in this message is the GSM Handover command.
• The UE switches to GSM RR protocol and sends the GSM handover access message to the BSC. • The BSC finally initiates resource release with message to the UTRAN
Micro diversity Referring to the soft handover and active set, there are two terms describing the handling of the multipath components: micro diversity and macro diversity. Micro diversity means the situation where the propagating multipath components are combined in the Node B.
Macro diversity Because of the fact that the UE may use cells belonging to different Node Bs or even different RNCs, the macro-diversity functionality also exists on the RNC level. The following picture presents a case in which the UE has a 3-cell active set in use and one of those cells is connected to another RNC. In this case, the Node Bs do signal summing concerning the radio paths of their own. In the RNC level, the serving RNC evaluates the frames coming from the Node Bs and chooses the best signal to send towards the CN domains. Summary of code usage in the uplink and downlink direction:
In the downlink direction, we need to be able to make a difference between different cells. Therefore the scrambling code is used for this purpose. Since we also must make a difference between different users within the cell, the channelisation code is used for this purpose. It also means that we will only use one dedicated physical channel in the downlink direction, and both signalling information (such as power control commands) and application data must be mapped onto this physical channel. In the uplink direction, we do not need to separate between cells. It means that we can utilise the scrambling code for separating between the different users. Following the same logic as earlier, it means that we can use the channelisation code to separate between different channels.
CDMA sequencing – a way to spread information Sequencing as a term refers to how the information to be transferred over the radio path with CDMA technology is spread over the defined frequency band. There are two basic alternatives: Frequency Hopping (FH) and Direct Sequencing (DS). In UMTS Release 99, there are two WCDMA modes: • FDD mode FDD stands for frequency division duplex. Two separate 5 MHz frequency bands are used – one for uplink transmission and another one for downlink transmission. • TDD mode TDD stands for time division duplex. Hereby, one frequency band is used both for uplink and downlink transmission.
RNC tasks This subsection gives a summary of the main tasks and functions: WCDMA radio resource management
Radio resource management of channel configurations, that is, how many traffic channels and signalling channels can be used in the RAN. This is done in connection with the radio network planning. • Management of traffic channels and stand-alone dedicated control channels can be further divided into radio resource management that attends to code allocation, admission control, channel release, load control, power control and handover control. • Handovers are controlled by the RNC, but can be initiated by the mobile station (MS) or the RNC. A handover can be one or more of the following types: − Soft, softer and hard HO with intra- and inter-RNC handovers − Code switching and code type switching HO. •
Telecom functionality
The telecom functionality includes tasks that are much related to the mobility and session management of subscribers and their connections: • Location and connection management • Ciphering • Indication of blockage on the channels between the RNC and the MSC • Allocation of the traffic channels between the RNC and the base stations • ATM switching and multiplexing • ATM transmission on SDH or PDH
• •
GPRS tunnelling protocol (GTP) towards the packet core network Security functions
Maintenance Operation
High Speed Downlink Packet Access (Rel.5) High Speed Downlink Packet Access (HSDPA) is a feature based on a downlink shared channel reserved for transfering only data. This feature allows data rates up to 10 Mb/s. The new high speed access is based on the new High Speed-Downlink Shared Channel (HS-DSCH) transport channel, which keeps some of the characteristics of the Release 99 DSCH. It is defined for FDD and both TDD modes. It is a time shared channel, mapped to one or more physical data channels. A new physical downlink data channels is defined (HS-PDSCH), together with an associated downlik control channel for layer 1 signalling (HSSCCH). An uplink signalling channel is also required, HS-DPCCH, based in the standard DPCCH. One of the main characteristics of HSDPA is the advanced link adaptation: the transmission scheme changes every Transmission Time Interval to adapt to the radio link conditions. HSDPA uses link adaptation AMC, adaptative modulation and coding with several predefined combinations of modulation and channel coding. The Node B selects the modulation and the coding for each TTI for each user based on an estimate of the downlink. The UE reports in the uplink signalling a measurement of the downlink. Higher order modulations (16 Quadrature Amplitude Modulation) will be used in good radio link conditions and lower schemes (Quadrature Phase Shift Keying ) are used in poor radio conditions to maintain the error rate. Automatic Retransmission Query (ARQ) is an error detection mechanism used in the link layer that brings controlled efficient retransmissons. The HSDPA benefits are: • support for services requiring high data rates in downlink, e.g. Internet browsing and video on demand. • High data rates up to 10Mbit/s
