Professional Documents
Culture Documents
TMO18315 - V7.0 SG ENLR13.3L Ed1
TMO18315 - V7.0 SG ENLR13.3L Ed1
SAFETY WARNING
Alcatel-Lucent training materials can be for products or refer to products that have both lethal and dangerous voltages
present. Always observe all safety precautions and do not work on the equipment alone. The user is strongly advised not
to wear conductive jewelry while working on the products. Equipment referred to or used during this course may be
electrostatic sensitive. Please observe correct anti-static precautions.
DISCLAIMER
ALCATEL-LUCENT DISCLAIMS ALL WARRANTIES REGARDING THE TRAINING COURSES OR THE CONTENT, EXPRESS OR IMPLIED,
INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR
PURPOSE. THE ALCATEL-LUCENT WILL NOT BE RESPONSIBLE OR LIABLE FOR ANY INJURY, LOSS, CLAIM, DAMAGE, OR ANY
SPECIAL, EXEMPLARY, PUNITIVE, INDIRECT, INCIDENTAL OR CONSEQUENTIAL DAMAGES OF ANY KIND (INCLUDING WITHOUT
LIMITATION LOSS PROFITS OR LOSS SAVINGS), WHETHER BASED IN CONTRACT, TORT, STRICT LIABILITY OR OTHERWISE, THAT
ARISES OUT OF OR IS IN ANY WAY CONNECTED WITH (A) ANY USE OR MISUSE OF THE CONTENT OR THE TRAINING COURSES
BY YOU, OR (B) ANY FAILURE OR DELAY BY ALCATEL-LUCENT, ITS OFFICERS, DIRECTORS, AGENTS OR EMPLOYEES IN
CONNECTION WITH THE CONTENT OR THE TRAINING COURSES (INCLUDING, WITHOUT LIMITATION, THE USE OF OR INABILITY
TO USE ANY COMPONENT OF THE CONTENT OR TRAINING BY YOU). SOME JURISDICTIONS LIMIT OR PROHIBIT SUCH
EXCLUSION OF WARRANTIES OR LIMITATION OF LIABILITIES AND SO THE FOREGOING EXCLUSION OF WARRANTIES OR
LIMITATION OF LIABILITY MAY NOT APPLY TO YOU.
GOVERNING LAW
These Terms of Use and Legal Notice are governed by the laws of France. The operation and use of the training course is
governed by the laws of the country that governs your employment contract, if applicable. If any provision of these Terms
of Use and Legal Notice, or the application thereto to a person or circumstance, is held invalid or unenforceable by law,
statute or a court of competent jurisdiction, for any reason, then such provision shall be modified and/or superseded by a
provision that reflects the intent of the original provision as closely as possible. All other provisions of these Terms of Use
and Legal Notice shall remain in full force and effect. You may not assign these Terms of Use or any permission granted
hereunder without Alcatel-Lucent’s prior written consent. Nothing herein shall be deemed an employment agreement or
an offer of employment or an alteration in any way of a user’s terms of employment with or within Alcatel-Lucent.
Copyright © 2013 Alcatel-Lucent. All Rights Reserved
training.feedback@alcatel-lucent.com
Thank you!
Document History
Feature Number: When possible, a reference is provided to the LTE feature number that caused the
parameter to be added or that altered the parameter’s use. When a parameter is related to several features,
multiple feature numbers may be indicated. Some parameters are not associated with a specific feature, and,
in the case of these parameters, the feature number entry will be blank.
Document History
CL-MIMO 1 layer is used in spatial multiplexing when rank = 1 in good radio conditions. It differs from
TxDiv because the eNodeB receives from the UE the PMI matrix intended to orthogonalize the 2 radio
signals and thus avoiding interference between them. The consequence is a better throughput in CL-MIMO
1 layer than in TxDiv.
CL-MIMO 2 layers is used in spatial multiplexing when rank=2 with good radio conditions. The case of a
single codeword mapped to two layers is only applicable when the number of antenna ports is 4 (not
supported in LA6.0).
In (T)LA6.0, only 2 layers at maximum can be used. In spatial multiplexing, 1 codeword is used for 1 layer
and 2 codewords for 2 layers. For transmit diversity, there is only one codeword and 2 layers.
There is one SINR per transmission scheme (for instance, in TM3, there are two SINR, one for TxDiv mode
and one for OL-MIMO)
Design Issue: The critical aspect for designing the CQI feedback mode is the trade-off between the
downlink system performance and the uplink bandwidth consumption (scare resource). Wideband CQI
optimizes the uplink bandwidth but does not allow FSS (which maximizes DL performance).
Wideband CQI is preferred for high speed UEs (channel changes rapidly and too many aperiodic CQI
reports would be required) or for VoIP (latency is more critical than the overall throughput)
The PUCCH is never transmitted simultaneously with the PUSCH from the same UE in order to retain the
single carrier property of SC-FDMA. They are time multiplexed. If there is a radio link established with the
UE, the periodic reporting is then done on PUSCH (and not on PUCCH). For TDD, the PUCCH is never
transmitted in the UpPTS field (contains the UL sounding reference signal and the random access). The
PUCCH is transmitted at the bandwidth edge, in order to provide the contiguous bandwidth in the middle
for data (as only localized resource allocation is allowed in UL, that means contiguous PRBs allocation per
UE)
There are 4ms between the PDCCH grant and the PUSCH transmission.
The UCI (CQI/PMI/H-ARQ-ACK and SR) can be multiplexed with UL-SCH data on the PUSCH channel and
there is no need to send a SR (scheduling request) since the UE BSR (buffer status report) is included. In
this case, the channel coding for H-ARQ-Ack, RI, CQI/PMI is done independently. Different coding rates can
be achieved by allocating different numbers of coded symbols, depending on the amount of allocated radio
resource. After encoding, the CQI encoded sequence is multiplexed with UL-SCH data, the output of which
is interleaved with the RI and H-ARQ-ACK encoded sequence.
Parameter cFI (case static CFI) is a key RF optimization parameter. Higher values allow for more PDCCH
robustness and/or more users served (increased capacity), but at the expense of throughput (fewer
resources for PDSCH).
The common search space: to be decoded by all UE in the cell in idle or RRC connected state. It is
subdivided into:
The non-C-RNTI common search space: for Paging, D-BCH (SIB), Mess4
The C-RNTI common search space used for Timing advance cmd, SRB1, VoIP
For SIB, the identity indicated on the PDCCH is not the UE identity but is the System Information Radio
Network Temporary Identifier (SI-RNTI).
UE-specific search spaces: to be decoded by the UE once RRC connected. The UE derives its own UE
search space based on the C-RNTI assigned during the RACH procedure (DCI CRC scrambled with UE
Identity). Several blind decoding are possible.
Note that for VoIP, in order to minimize overhead, certain frequency resources may be persistently-
scheduled, i.e. certain resource blocks in the frequency domain are allocated on a periodic basis to a
specific UE by RRC signaling rather than by dynamic PDCCH signaling. VoIP is always used with DCI1A to
improve cell coverage at cell edge. Common search space is used for VoIP.
HI is mapped onto the PHICH channel. Multiple PHICHs mapped to the same set of Resource Elements
constitute a PHICH group, where PHICHs of the same group are separated through orthogonal sequences
with a spreading factor of 4. A PHICH is identified by (Ngroup, Nseq) where Ngroup is the PHICH group
number and Nseq the orthogonal sequence index in the group.
In order to indicate which PHICH carries the ACK/NACK response for each PUSCH transmission, the PHICH
index is implicitly associated with the index of the lowest uplink resource block used for the corresponding
PUSCH transmission. This relationship is such that adjacent PUSCH resource blocks are associated with
PHICHs in different PHICH groups, to enable some degree of load balancing.
The PHICHs are transmitted on the same set of antenna ports as the PBCH, and transmit diversity is
applied if more than one antenna port is used.
Demodulation RS (DRS), associated with transmissions of uplink data on PUSCH and/or PUCCH. These
RSs are primarily used for channel estimation for coherent demodulation. Since PUSCH and PUCCH
cannot be transmitted simultaneously, there are two DRS for PUCCH and PUSCH.
Sounding RS (SRS), not associated with uplink data and/or control transmissions, and primarily used
for wideband channel quality determination to enable frequency-selective scheduling FSS on the uplink
(i.e. the frequency resource is assigned selectively on the subbands having the best CQI)
Why do we have two types of RS in UL (unlike the DL)? The DRS in UL are only transmitted on subcarriers
assigned to UE and therefore do not provide sufficient wideband channel quality information for resource
allocation (particulary for RB not allocated to UE).
SRS is not sent when there is a PUCCH format 2/2a/2b (over 10 symbols) or when there is a ACK/NACK or positive SR
in the same subframe on the last symbol
There are 4 Preamble formats in LTE. Format 0 is the ‘normal’ format. Format 1, also known as the
‘extended format’, is used for large cells (extended CP). Format 2 and format 3 use repeated preamble
sequences to compensate for increased path loss, and are used for small cells and large cells respectively;
format 4 is defined for frame structure type2 only.
ALU LR13.1 supports formats 0&2&3; ALU TLA6.0 supports formats 0&4
The MAC function starts an uplink synchronization timer (ulSyncTimer) upon receipt of a notification from
the UL scheduler that the related call has transitioned to “out of sync state”. This timer is stopped upon
receipt of a notification that the call has transitioned back to “in sync state”. If the timer expires, the UL
scheduler transitions the UE to MAC_RLF state, i.e. declares “MAC RLF” (as opposed to RLF due to the
maximum number of RLC transmissions being reached without receiving an RLC ACK). Upon expiration of
ulSyncTimer, the uplink scheduler considers the radio link associated with the UE context to be failed.
Note that TA failure notification (from DL scheduler) transitions the UE to MAC_RLF state regardless of the
state it was in. The UL scheduler ignores the TA failure event from the DL scheduler if it occurs within
100ms after RACH 3 event (for an existing UE).
Also note that in case of UE RACH-back (i.e. if a RACH message 3 is received) the UL scheduler transitions
the UE to the UL IN_SYNC state.
The Inactivity based DRX is not supported when SPS (Semi-Persistent Scheduler) is enabled for a call, or when
TTI-bundling is configured, or if Measurement Gap is configured.
DRX mode activation: (DRX mode activation: description applies to DRX used for CSFB or ANR but not for DRX under inactivity
conditions. Here DRX is activated when both onduration and drxinactivity timers are not running)
CallP sends a DRX start command to the DL scheduler to request it to send the DRX command (MAC CE) to the UE.
The DL scheduler generates a Timing Advance command to preserve UL timing alignment during the cycle and thus avoid
having to resort to a RACH procedure to re-establish data flow (in case of TA timer expiry).
The DL scheduler generates a (TA + DRX) command and considers scheduling it immediately:
- The (DRX+TA) command has the highest or the 2nd highest scheduling priority
The DL scheduler notifies the UL scheduler that it is trying to get the UE into DRX.
If the first HARQ attempt fails, the (DRX+TA) command is retransmitted until
- an ACK is received
- a new DRX On Duration starts
- A new timing correction is generated
- The new timing correction is included in a (DRX+TA) command
When the HARQ ACK is received, the DL scheduler notifies the UL scheduler, indicating that the UE is now in DRX mode.
The DRX command, when correctly received, gets the UE in the DRX mode and immediately in OFF state.
Document History
1 Schedulers 8
1.1 Introduction 9
1.1.1 Schedulers in MAC Layer 10
1.2 DL UE Context 11
1.2.1 Number of DL UE Context 12
1.2.2 Number of DL UE Context at UE 13
1.3 DL Scheduler Principles 14
1.3.1 Prebooking and Scheduling Phases 15
1.3.2 TimeFrequencyResBlockOccupancy Matrix 16
1.4 DL Static Scheduler 17
1.5 DL Semi-Static Scheduler 19
1.5.1 DL Semi-Static Scheduler Figure 20
1.5.2 D-BCH Scheduling 21
1.5.3 SIB1 Scheduling 22
1.5.4 SIB2 to 13: Scheduling Class 23
1.5.5 SIB2 to 13: Target MCS 24
1.5.6 SIB2 to 13: Periodicity 25
1.5.7 SIB2 to 13: SI Messages 26
1.5.8 SIB2 to 13: SI Windows 27
1.5.9 SIB2 to 13: Example 28
1.5.10 SIB2 to 13: Scheduling Process 29
1.5.11 PCCH Prebooking 30
1.5.12 pagingForceMCSmin 31
1.6 DL Dynamic Scheduler 39
1.6.1 FDS and FSS Modes of Dynamic Scheduler 40
1.6.2 Scheduling Types 43
1.6.3 QoS Weighted Metric 45
1.6.4 alphaFairnessFactor 46
1.6.5 Resource Allocation 47
1.6.5.1 Implementation 49
1.7 Minimum Bitrate Enforcement for Non-GBR 53
1.7.1 Downlink Scheduler Mechanism 54
1.7.2 Uplink Scheduler Mechanism 55
1.7.3 minRate QoS Weight 56
1.7.4 Downlink Parameters 57
1.7.5 Uplink Parameters 58
1.7.6 Example 59
1.7 DL Semi-Persistent Scheduler for VoIP 62
1.7.1 Semi Persistent vs Dynamic Schedulers 63
1.7.2 VoIP Support at UE and eNB levels 64
1.7.3 DL SPS Prohibit Timers 65
1.7.4 Conditions to Activate the DL SPS 66
1.7.5 DL SPS Activation Procedure 67
1.7.6 MCS Computation for DL SPS 71
1.7.7 VoIP MAC frame with RoHC 72
1.7.8 TBS Computation for DL SPS 73
1.7.9 DL SPS Release 75
1.7.10 DL SPS Deactivation Procedure 76
1.8 UL Scheduler Principles 79
1.9 UL Static Scheduler 80
1.9.1 RACH Message1 Resource Allocation 81
1.9.2 RACH Message3 Resource Allocation 82
1.10 UL Dynamic Scheduler 84
1.10.1 UL Buffer Status Reporting from UE 85
1.10.2 UL Buffer Estimation in eNodeB 86
If IE featureGroupIndicators is not sent by the UE, the eNB assumes that all the radio bearer combinations
are supported (equivalent to both bit 7 and bit 20 in IE featureGroupIndicators set to 1).
No more than two VoIP bearers per UE is supported (3GPP). If more than 2 UM is supported, then VoIP
bearer checking has to be performed to ensure this restriction. Also, requested combinations not supported
by the UE are rejected.
A prebooking stage at low rate (20ms) which reserves resources for the static and semi-static
schedulers. The static resources like RS, PSS, SSS, PBCH (static), RACH Msg2 and SIB0 are reserved
A scheduling stage at TTI rate (1ms) which assigns the resources for effective traffic. The remaining
RE are shared in time and frequency for PDCCH, PCFICH, PHICH, PDSCH.
Note that regardless of the value of parameter pagingForceMCSmin, the MCS range never goes outside the
range [0, 9] for QPSK is mandatory for the transmission of paging messages.
Restriction:
In LA6.0, to keep available bandwidth for semi-persistent VoIP and PCCH allocation, up to 1 paging sub-
frame per frame is supported: nB ≤ T. This gives a PCCH bandwidth of 1 sub-frame per Paging Cycle, so up
to 32 Paging Occasions per Paging Cycle of 320 ms.
nB represents the number of paging sub-frames per Paging Frame. Its value is {4T, 2T, T, T/2, T/4, T/8,
T/16, T/32}.
RA backoff: The RAR (Random Access Response) conveys the identity of the detected preamble, a timing
alignment instruction to synchronize subsequent uplink transmissions from the UE, an initial uplink resource
grant for transmission of the Step 3 message, and an assignment of a temporary Cell Radio Network Temporary
Identifier (C-RNTI). The RAR message can also include a ‘backoff indicator’ which the eNodeB can set to instruct
the UE to back off for a period of time before (random value between 0 and raBackoff) retrying a random access
attempt.
alphaFairnessFactor = 0.0 yields a maximum C/I scheduler. The scheduler provides more resources to UEs
in better conditions. The better the radio conditions of the UE, the more resources (and hence the higher
the data rate) it gets.
alphaFairnessFactor = 1.0 yields a fair scheduler. The scheduler attempts to provide the same number
of RBs to all the UEs (despite their different conditions).
alphaFairnessFactor = 2.0 yields an increased fairness scheduler. The scheduler attempts to allocate
the resources in such a way that all the Ues eventually get the same data rate (which is not the case
of the fair scheduler since different radio conditions result in different data rates even when the
number of resources is the same, hence the increased fairness of the scheduler, as compared to the
“regular” fair scheduler).
alphaFairnessFactor = 0.5 is a scheduler with an intermediate behavior between alphaFairnessFactor =
0.0 (maximum C/I scheduler) and alphaFairnessFactor = 1.0 ( fair scheduler).
alphaFairnessFactor = 1.5 is a scheduler with an intermediate behavior between alphaFairnessFactor =
1.0 (fair scheduler) and alphaFairnessFactor = 2.0 (increased fairness scheduler).
where Rmin(b) is the minimum throughput target dlMinThroughputTarget for the radio bearer b, and m is the
margin applied on top of the min throughput target marginForNonGbrMinRateEnforcement.
The increasing weight function is not linear as it is for the UL min-rate enforcement, but its logarithm is linear.
In fact, we can simplify the formula above by writing:
The variable u above represents the ratio of throughput in excess of Rmin(b): DL throughput measured = Ra(b)
= (1 + u) * Rmin(b), 0 ≤ u ≤ m.
To better understand the calculations, see the example a couple of slides further.
o If RoHC is used, the GBR value is mapped to one of the codecs: AMR-12.20, AMR-10.20, AMR-7.95, AMR-
7.40, AMR-6.70, AMR- 5.90, AMR-5.15, AMR-4.75, AMR-WB-23.85, AMR-WB-23.05, AMR-WB-19.85, AMR-
WB-18.25, AMR-WB-15.85, AMR-WB-14.25, AMR-WB-12.65, AMR-WB-8.85, AMR-WB-6.60. Then, the right
lookup table is used to map the VoIP packet size before the PDCP, RLC and MAC headers are appended,
associated with the nominal VoIP frame size for the considered Codec. The TBS is obtained by adding the
DL PDCP/RLC/MAC overhead, configured by parameter macRLCOverheadDl.
o If RoHC is not used, the TBS is computed as TBS = S1 DL GBR × 0.020 sec + macRLCOverheadDl.
In the UEBearerList, a bearer context is identified by the Logical Channel Identifier (LCID) of the logical
channel it is mapped onto. The bearer context also contains, among other information:
The VoIP flag (derived by CallP from QoS received from the MME over the S1-C).
The UL GBR (derived by CallP from QoS received from the MME over the S1-C).
The QCI differentiation feature has been implemented in LA4 (92095 feature): DL offers a fixed (3:1)
nonGBR QoS differentiation PDB based QoS as a temporary solution. It was possible to have 3 times more
throughput for a nonGBR bearer compared to another one.
dBm to mW conversion:
P = 10 x/10 with P in mW and x in dBm.
0dBm=1mW; 3dBm=2mW (adding 3dB means doubling the power); -3dBm=0.5mW (substracting 3dB
means dividing the power by 2) etc.
Parameter paOffsetPdsch is a key RF optimization parameter impacting the end user data rate. Higher
power levels improve the downlink throughput at the expense of the power available for other downlink
signals and channels. Smaller settings degrade the downlink throughput but contribute lower interference
to the UEs served by the neighboring cells.
The power offset assoiciated with pbOffset index (0,1,2,3) correspond to the different cases in one symbol
of the ratio between the PDSCH RE and RS RE (cell-specific and related to the antenna port). The ratio can
be:
Uplink power control in LTE can be seen as a ‘toolkit’ from which different power control strategies can be
selected and used depending on the deployment scenario, system loading and operator preference.
If above ALU parameter=50 (fixed by default), PUSCH SINR is computed from UE SRS measurement
reports sent for instance every 5ms: there will be at most one TPC command sent every 50*5 = 250ms >
closed-loop power control in LTE is a slow mechanism
Power control command is sent to the UE in DCI Format 0, 1 or 2, when dynamic grant for DL allocation is
available. Otherwise, the command is sent in DCI Format 3/3a (and addressed to TPC-RNTI associated to
the UE). DCI Formats 3/3a carry only TCP command, format 3a is more compact.
Mapping of TPC Command Field in DCI format 0/1A/1/2A/2/3 and δ PUCCH [dB]:
0 : -1; 1 : 0; 2 : 1; 3 : 3
For DCI format 3a: 0 : -1 and 1 : 1
Frequency Selective Scheduling (FSS): allocation of resources (PRB) based on channel quality
(subband CQI)
Other features like ICIC and eICIC (inter-cell interference coordination) are introduced in LR13.
Notice that I_MCS 9 and 10 are mapped on the same I_TBS 9, resulting in the same data rate. Such
modulation overlap is adopted to improve the performances around the modulation switching point, as
different combinations of modulation and coding with the same rate may provide different performance in
different scenario.
The fact to have two tables brings more flexibility in the AMC process because you can for exemple accept
less robust MCS on small packets in order to favor the throughput vs the robustness (a retransmission of a
small packet has less impact on the global UL performances than a large packet)
The HARQ process timer is configured by parmeter macHARQMaxTimerDl (per radio bearer or per signaling
bearer; fixed, range 1-500ms). It is started at the time of the first transmission of the HARQ process. At its
expiry, the HARQ process is killed (RLC ARQ process can reinitialize the sending of the message). An HARQ
process is also killed after the maximum number of transmissions reached. After the killing of a process, all
the related allocated resources are freed.
For TLA6.0 only: In an adaptive HARQ scheme, transmission attributes such as the modulation and coding
scheme, and transmission resource allocation in the frequency domain, can be changed at each
retransmission in response to variations in the radio channel conditions. In a non-adaptive HARQ scheme,
the retransmissions are performed without any change. Parameter hARQRetransmissionMode (per cell;
fixed, default non-adaptative) controls whether adaptave HARQ is applied to retransmission.
Note that HARQ process does not apply to MBSFN subframe, and special subframe in special subframe
configuration 5 in TLA6.0.
On the other hand, if the required number of PRBs for TTI Bundling is greater than the number of available PRBs
for TTI Bundling, the zone reserved for TTI Bundling will be increased by the maximum number of possible
prbGrantSizeForTTIBundling , without exceeding the zone configured by ttiBundlingPRBpriorityOrder.
Example:
prbGrantSizeForTTIBundling = 2
Max number of PRBs configured by ttiBundlingPRBpriorityOrder = 28
Current number of PRBs reserved for TTI Bundling = 14
Number of PRBs required for TTI Bundling = 17
=> The zone reserved for TTI bundling will be increased to reach 18 (=14 + 2*2)
In order to provide the best trade-off between residual BLER and VoIP packets delay when TTI Bundling is
configured, two thresholds are introduced in order to control if segmentation of the UL VoIP frames is allowed
(see also L160815 – Optimized segmentation).
ttiBundlingSegmentationBLERthresh : This parameter controls the BLER threshold measured over dynamic TTI
Bundling grants. The threshold is used to allow speech payload segmentation for dynamic grants.
ttiBundlingNoSegmentationBLERthresh : This parameter controls the minimum SINR level used for the uplink SE
metric associated to TTI Bundling users. This minimum floor is introduced to have a better control on the
fairness between TTI Bundling and non-TTI Bundling traffic.
Document History
Message 1: This message contains the random access preamble. It is randomly selected from a set of
Random Access Preambles the number of which is configured by parameter numberOfRAPreambles. Once
message 1 is transmitted, the UE starts monitoring the PDCCH for Random Access Response in the PDCCH
common space with the RA-RNTI key (the time-freq slot used for the preamble is associated with an RA-
RNTI)
Message 2 (Random Access Response): This message is generated by MAC on DL-SCH and intended for a
variable number of UEs. It conveys a Random Access preamble identifier, assignment of Temporary C-
RNTI, as well as timing advance information and initial grant for the transmission of message 3. It is
addressed to RA-RNTI on PDCCH and does not use HARQ.
Message 3 (First scheduled UL transmission on UL-SCH). The UE is UL synchronized to send data on PUSCH.
The UEs send the Scheduled Transmission with the temporary C-RNTI and its UE identity (C-RNTI for
connected UE or a 48bits random sequence). For users that do not already have a C-RNTI (but only a
temporary one), this message conveys either the RRC Connection Request (for initial access from
RRC_IDLE) or the RRC Connection Re-establishment Request (after radio link failure). After the first
transmission of message 3, the UE starts the mac-contention-resolution-timer. This timer is restarted
after each HARQ retransmission of message 3.
Message 4 (Contention Resolution on DL-SCH): eNodeB selects only one UE and sends its UE identity in
the Contention Resolution message. The UE whose identity matches the one sent by eNodeB
acknowledges. The UE Contention Resolution identity is addressed on PDCCH either to the C-RNTI (for UEs
that already have one) or to the Temporary C-RNTI (for UEs that do not already have a C-RNTI). At the
end of the Contention Resolution process, Temporary C-RNTI= C-RNTI
If the mac-contention resolution timer expires, the contention resolution is considered not successful.
Parameter macContentionResolutionTimer configures the mac-contention resolution timer.
If the Random Access Response or the Contention Resolution fails, the UE makes an immediate reattempt
SRB1 and SRB2 are secure; that is, it is integrity protected and, if a ciphering algorithm is available, then it is
also ciphered by the PDCP layer
RRC Connection Request includes the S-TMSI and the establishment cause. If accepted (radio resource
available in eNodeB), the eNodeB returns RRC Connection Setup including the initial radio resource
configuratation with SRB1.
RRC Connection Setup Complete includes the NAS message (Service Request, Initial attachment, Tracking Area
Update, etc) to be transferred to the MME. This NAS message triggers the establishment of the S1 connection,
which normally initiates a subsequent step during which E-UTRAN activates AS-security (integrity protection
and ciphering) and establishes (RRC Connection Reconfiguration message) SRB2 and one or more DRBs
(corresponding to the default bearer and optionally dedicated EPS bearers).
t302: This UE timer is started upon reception of RRCConnectionReject and is stopped upon successful RRC
establishment or cell re-selection. Sent in RRCConnectionReject (information element waitTime)
t300: This UE timer is started when sending RRCConnectionRequest and is stopped upon reception of
RRCConnectionSetup or RRCConnectionReject. Broadcasted in SystemInformationBlockType2
Any RRC Cnx Request for MO Data calls of access classes 0-9 can be barred if accessBarringInformation IE is
broadcast is SIB2. If so then some RRC Cnx Req will be barred some other will not according to the Rand value
computed dynamically by the UE of an Access Class X between 0 and 9 that does not belong to any other Access
Class or for which the Access Classes 11-15 it also belongs to are all barred
• If the UE initiates a non-emergency MO call and at least one of its ACs is barred in SIB2:
-UE draws a random number (0<=rand<1)
-For AC0-9, if rand < probability factor, access is not barred. Otherwise, access is barred during T303 (the
barring timer parameter is used to compute T303)
-For AC11-15, access is barred if all UE ACs are barred according to previous algorithm
• Case MO emergency calls. If AC10 is barred in SIB2:
-UEs with AC0-9 or without IMSI are not allowed for emergency calls
-UEs with AC11-15 are not allowed for emergency calls if the relevant AC11-15 are barred. Otherwise,
emergency calls are allowed for those Ues
The ACs [11-15] refer to high priority services and can still initiate non-emergency MO call even if their AC [0-
9] is barred (provided one of their AC [11-15] is not barred)
Before LR13.3 TAC (Time Alignment Command) transmission was unconditional , TAC was scheduled as long as
user was listening to PDCCH. LR13.3 introduces conditional TAC transmission based on user traffic.
Now UE time alignment timer can expire because of traffic inactivity. When so the eNB will not assign
PUCH/SRS resource to the UE.
Feature adds handling of RACH-back event for OOT users going back into Active state (triggered by the UL or
DL traffic arrival)
Intermediate state Pending-Active (modem state only): user is time-aligned with eNB but has no
PUCCH/SRS resource (configured by eNB). It still listens to PDCCH (in Active time if inactivity-based DRX is
configured) and receives PDSCH. Pending-Active user needs to first send SR in order to send UL traffic
Modem resets and starts TATimer
If UE is in Pending Active from RRC connection re-establishment or call setup, modem start Guard Timer
Guard Timer is hardcoded to 100ms and introduced so UE can complete PUCCH/SRS reconfiguration, after
which UL Sched Sync monitoring can start.
After Guard Timer expiry UE transits to Active State
If the TATimer expires UE transits back to OOT State
Recall (see radio module for more details): timeAlignmentTimer (for common or dedicated channels) is sent
to the UE (via RRC message) and configures the max time between two TAC (sent by eNB) before the UE
considers the UL synchronization lost. On the enB side, at UE creation, a timer
timingAlignmentCommandTimer is started. When this timer expires, a Timing Advance (TA) command is sent
and the timer is restarted. This also happens when L1 UL indicates to DL MAC that the timing advance has
changed and requests a TA command be sent to the UE. timingAlignmentCommandTimer = rounddown
(timeAlignmentTimer/2.5), in the unit of msec.
CSFB Call:
Transition from Active to OOT is disabled for (e)CSFB call establishment.
DRX parameters used to perform IRAT measurements need to be reconfigured (in the modem) when
UE returns from OOT state to active state
EC or High Priority Access (HPA)
IF detected Emergency call or HPA( RRC establishment cause or special ARP value over S1 or X2.
transition to OOT state is disabled
PUCCH/SRS pool will be reserved for EC/HPA calls to limit the possibility of blocking
Commerical CA (Carrier Aggregation) :
OOT and CA are not allowed to coexist in LR13.3.
UE supporting CA will not be eligible to OOT if CA is activated
If feature is deactivated online the criterion is only checked when the UE call context is created in eNB
CallP
Remark: When a user RACH-es back due to DL or UL TRB arrival, then if eNB finds there is no CAC capacity or
no PUCCH/SRS resource available for the user, the eNB releases the user to RRC idle.
Engineering Recommendation: isRedirectionForOotEnabled : The default value of this parameter is False. It can
be changed to True if the
operator wants the indicated behavior. In case of TRB data, the recommendation is to set to TRUE to perform
RRC Connection Release with Redirection to a specific carrier, based on the eMCTA algorithm. If set to FALSE
the UE is going to be requesting another RRC Connection after going to idle mode and if load balancing is
activated and depending on CAC filtering the UE may be accepted or redirected.
The standardized QCI label characteristics describe the packet forwarding treatment through the network
based on the following parameters:
Resource Type (GBR or non-GBR)
Priority
Packet Delay Budget (PDB)
Packet Error Loss Rate (PLR)
The service level (i.e., per SDF or per SDF aggregate) QoS parameters are QCI, ARP, GBR, and MBR.
Remark: CAD is not always involved. The PCRF may e.g. decide to remap QCI and ARP, based on information
received lately from another RAT.
HPA calls will not be eligible to OOT, due to risk of release when coming back
out of OOT.
eUTRA sharing: The priority used to identify bearers set up for High Priority Access users must be defined on a
per-PLMN basis
In case of eUTRA sharing, this parameter is under another MO. Parameter PlmnIdentity::
reactiveLoadControlActionForEcAndHpaAdmission and reactiveLoadControlActionForBearerAdmission::action
define the action (release or offload UE) for Reactive Load Control triggered by IMS VoIP EC admission and
bearer admission. The PlmnIdentity object instance is obtained from TrafficRadioBearerConf::
ReactiveLoadControlActionforBearerAdmission::plmnId.
Rationale to add a level of QoS degradation in LR13.3 to trigger a preventive load control:
Only a single or just a few UEs may use the full PRB band, in which case the eNB should not consider that the
cell is congested. The eNB is then be able to estimate the average perceived downlink RLC throughput
performance for non-GBR bearers. The downlink scheduler reports the total RLC burst size and time (activity
duration) for each non-GBR QCI, which allows calculating the average downlink RLC user throughput per QCI.
A low average throughput for active non-GBR bearers is a symptom of congestion in the cell and is used as a
criterion to trigger load balancing.
When one or multiple thresholds for load balancing features are met at the same time, the UE selection rules
will be based on the trigger that requires the higher number of UEs to be offloaded.
The differentiation between active and connected users is done with LR13.3 feature 166502 (Out of Time
Alignment OOT).
eNB will initially avoid triggering inter-frequency handover for UEs that have been inactive for an extended
period of time, since they should have not contributed to the detected congestion and are less likely to
contribute for it in the immediate future.
Note that the mechanism of offloading is fully described in the Mobility module (eMCTA).
The Criticality Diagnostics IE in X2 RESOURCE STATUS RESPONSE is sent by the eNB when parts of a received
message have not been comprehended or were missing, or if the message contained logical errors. When
applicable, it contains information about which IEs were not comprehended or were missing.
The Capacity Value IE indicates the amount of resources that are available relative to the total E-UTRAN
resources. The Capacity Value IE can be weighted according to the ratio of cell capacity class values, if
available. Value 0 shall indicate no available capacity, and 100 shall indicate maximum available capacity.
Example: if one of the cells has a capacity of 20 calls, and the overlapping neighbor cell has a capacity of 5
calls, then their cellCapacityClass parameter values might be set as follows:
Cell with 20 call capacity => cellCapacityClass value = 100 (%)
Cell with 5 call capacity => cellCapacityClass value = 25 (%)
In an operational network, cells having the same call handling capacity will be assigned the same
cellCapacityClass value.
Case LTE HeNB: Since there is no X2 connection between eNB and HeNB, cell load is not available for a LTE
HeNB open cell (134689). A HeNB open cell load status and its dedicated carrier load status are treated as ‘load
unknown’ for Serving Radio Monitoring, Reactive Load Control and Preventive Load Control.
Since UTRAN FDD small cells do not support RIM for cell load retrieving, cell load is not available for a UTRAN
FDD small cell (L115393). A UTRAN FDD small cell load status and its dedicated carrier load status are treated as
‘load unknown’ for Serving Radio Monitoring and Reactive Load Control. A UTRAN FDD small cell load status and
its dedicated carrier load status are treated as ‘loaded’ for Preventive Load Control.
Load equalization can be done between 2 cells from 2 different eNBs if they are on different frequencies.
The RadioCacCell::LoadEqualizationDeltaThreshold parameter allows triggering load equalization between
carriers, before preventive offloading is triggered. It is a negative offset to be applied on the preventive offload
threshold for static PRB usage. Value 0 means that load equalization is deactivated.
This parameter should be set to a value different than 0 (for example 20%) in order to have different triggering
thresholds for load equalization and preventive offloading (in order for load equalization to be triggered earlier
than preventive offloading).
Remark: Load equalization was requested as a means to equalize early large load imbalance between their two
5MHz PCS carriers. For instance, if loading on PCS1 is 50% whilst the loading on PCS2 is 20%, they wanted UEs
to be sent from PCS1 to PCS2 pro-actively. Their initial request was more complex then what we ended up
implementing: they wanted us to monitor the load delta between 1 cell on PCS1 and 1 or multiple cells on PCS2
and only ask UEs to measure PCS2 if the load delta was above a configurable threshold. What we ended up
implementing is a simplified version, very similar to preventive offload only with an earlier threshold. Meaning we
ask UEs to measure the other carrier as soon as the serving cell load threshold is met, and filter out
measurement reports on the target carrier if the load delta is not met.
The result in LR13.1 is that we have two similar schemes that trigger at different loading threshold on the
serving cell, and that require different loading delta between serving and target cell/carrier.
Note : the parameter cuLoadingThreshForEarlyOOTRelease, defined in RadioCacEnb, is used for both the cell
and the eNB. It is used as follows to derive the absolute threshold for early OOT release per cell and per eNB:
Per cell: Threshold for early OOT release = RadioCacCell::maxNbUser *
RadioCacEnb::cuLoadingThreshForEarlyOOTRelease/100
Per eNB: Threshold for early OOT release = RadioCacEnb::maxNbrOfRRCConnectedUsersPerEnb *
RadioCacEnb::cuLoadingThreshForEarlyOOTRelease/100
Each RoHC channel multiplexes different IP packet streams that are compressed differently using different
profiles. These different streams are identified with different compression contexts. A context is identified
with a Context ID (CID) and is a set of parameters and static/dynamic information for an IP packet stream
having the same type of header. Each context corresponds to an individual compression profile, which has
different compression algorithms. Supported RoHC profiles are defined in the Introduction slide.
Parameter rohcProfiles selects the RoHC profiles enabled for eNB. It is used for the input of PDCP RoHC profile
configuration, as well as eNB input for RRC RoHC profile configuration which also takes into account the UE capability.
This parameter is presented in a bit string format with a length of 9, with each bit representing one of the 9 profiles
defined in 3GPP. Profile 0x0000 is compulsary when RoHC is enabled so it is not included in the bitmap. The rohcProfiles
bitmap also aligns with the RRC PDCP-configuration format as follows:
Bit 1: profile 0x0001 RTP/UDP/IP
Bit 2: profile 0x0002 UDP/IP
Bit 3: profile 0x0003 ESP/IP
Bit 4: profile 0x0004 IP
Bit 5: profile 0x0006 TCP/IP
Bit 6: profile 0x0101 RTP/UDP/IP v2
Bit 7: profile 0x0102 UDP/IP v2
Bit 8: profile 0x0103 ESP/IP v2
Bit 9: profile 0x0104 IP v2
A bit set to "1" means that particular profile is enabled. A bit set to “0” disables the profile. If RoHC profiles have been set
to 0 for all bits, RoHC will be disabled for that bearer. Only Bit1 and Bit2 are supported in LA4.0. If all bits are set to 0,
this indicates ‘no compression’.
Therefore:
TrafficRadioBearerConf with qCI=1: set rohcProfiles to 0x110000000
TrafficRadioBearerConf with qCI ≠ 1: set rohcProfiles to 0x000000000 (no RoHCcompression)
Document History
1 Introduction 8
1.1 Mobility Overview 9
2 Mobility in Idle Mode 10
2.1 PLMN and Cell Selection in Idle mode 11
2.2 System Information for Mobility in Idle Mode 12
2.2.1 Neighbor Cells List and Priority 13
2.2.2 Blacklisted Neighbor Cells 14
2.3 PLMN Selection 15
2.4 What is measured by the UE for Cell (re)Selection 16
2.5 Cell Selection criterion (S-criterion) 17
2.5.1 Cell Selection Parameters 18
2.5.2 Focus on puMax 19
2.5.3 Focus on qRxLevMin 20
2.6 Cell Re-selection 23
2.6.1 Cells Measurements Occurence 24
2.6.2 Measurements Rules 25
2.6.3 Frequencies/RAT Evaluation 27
2.6.4 Cell Reselection Parameters 28
2.6.5 Cell Ranking 30
2.6.6 Cell-Reselection Algorithm 32
2.6.7 Example of Cells Measurements Period for Intra-freq 37
2.6.8 Exercise of Cell Ranking 38
2.6.9 Speed Dependant Scaling 39
2.6.10 Accessibility Verification 41
2.6.11 SIB3 Parameters Adjustment based on Cell Loading 43
2.6.11.1 Cell Loading Calculation 44
2.6.11.2 Delta Parameters Selection 46
2.7 Intra-LTE Cell Reselection Object Model 49
2.7.1 Inter-RAT Cell Reselection Object Model 50
3 RRC Measurements 51
3.1 What is measured by the UE? 52
3.2 How are the UE Measurements reported? 53
3.3 UE Measurements Usage 54
3.4 UE Triggered Events 55
3.5 eNB Mobility Software Architecture 56
3.6 UE Mobility State Diagram 57
3.7 CS Fallback Trigger 59
3.8 RRC Measurements 3GPP Object Model 60
3.9 RRC Measurements ALU Object Model 61
3.9.1 Example 62
3.9.2 measurementPurpose 63
3.10 General Diagram of an Event-triggered Measurement 64
3.10.1 reportConfig Common Parameters for all Events 65
3.10.2 measObject Common Parameters for all Events 66
3.11 Event A1 Triggering Conditions 67
3.11.1 Event A1 Diagram 68
3.12 Event A2 Triggering Conditions 69
3.12.1 Event A2 Diagram 70
3.13 Event A3 Triggering Conditions 71
3.13.1 Event A3 Diagram 72
3.14 Event A4 Triggering Conditions 73
3.14.1 Event A4 Diagram 74
3.15 Event A5 Triggering Conditions 75
3.15.1 Event A5 Diagram 76
3.16 Event B1 Triggering Conditions 77
Remark: details on the thresholds (‘poor’ and ‘good enough’) will be provided in the next slides.
PRB consumption i is the DL or UL PRB consumption per cell or per eNB in the ith period of
periodMeasForPRBConsumption (4 in total)
The E-UTRAN can configure the UE to report measurement information to support the control of UE mobility. The following
measurement configuration elements can be signalled via the RRCConnectionReconfiguration message.
1. Measurement objects. A measurement object defines on what the UE should perform the measurements – such as
a carrier frequency. The measurement object may include a list of cells to be considered (white-list or black-list) as
well as associated parameters, e.g. frequency- or cell-specific offsets.
2. Reporting configurations. A reporting configuration consists of the (periodic or event-triggered) criteria which cause
the UE to send a measurement report, as well as the details of what information the UE is expected to report (e.g.
the quantities, such as RSCP for UMTS or RSRPfor LTE, and the number of cells).
3. Measurement identities. These identify a measurement and define the applicable measurement object and
reporting configuration.
4. Quantity configurations. The quantity configuration defines the filtering to be used on each measurement.
5. Measurement gaps. Measurement gaps define time periods when no uplink or downlink transmissions will be
scheduled, so that the UE may perform the measurements. The measurement gaps are common for all gap-
assisted measurements.
The details of the above parameters depend on whether the measurement relates to an LTE, UMTS, GERAN or CDMA2000
frequency. The E-UTRAN configures only a single measurement object for a given frequency, but more than one
measurement identity may use the same measurement object. The identifiers used for the measurement object and
reporting configuration are unique across all measurement types. In LTE it is possible to configure the quantity which triggers
the report (RSCP or RSRP) for each reporting configuration. The UE may be configured to report either the trigger quantity
or both quantities.
Depending on the measurement type, the UE may measure and report any of the following:
The serving cell;
Listed cells (i.e. cells indicated as part of the measurement object);
Detected cells on a listed frequency (i.e. cells which are not listed cells but are detected by the UE).
Report-CGI:
DRX (Discontinuous Reception) can be configured for a UE in RRC connected state so that it stops monitoring
the downlink PDCCH channel during the DRX period. The main purpose of DRX is for the UEs to go the sleep (to
save battery) after a period of inactivity when no data was sent or received.
When a UE is configured to ‘Report-CGI’ (to find ECGI: E-UTRAN Cell Global Identifier of a neighbor cell),
however, eNB forces UE into DRX through the DRX Command in MAC control element. During the DRX period in
DRX cycle, the UE listens to the PBCH of the neighbor cell to receive its ECGI, TCA and PLMN ID.
When a UE is forced into DRX, it will cause transmission gap noticeable by the end user of a voice call or of any
other services required guarantee bit rate (GBR). For this reason, eNB A will direct a UE for ECGI search only if it
is capable to support the ‘reportCGI’ procedure, and it does not have on going GBR bearers. For the same
reason, DRX length defined drxCycleForReportCGI should be set just long enough for the UE to listen to PBCH of
the neighbor cell to find its ECGI.
reportAmount:
For ‘Intra-Frequency-Handover-Trigger’ and ‘Inter-Frequency-Handover-Trigger’ Event A3 trigger, the recommended value of reportAmount is ‘8’ . The main
reason to set reportAmount to ‘8’ for ‘Intra-Frequency-Handover- Trigger’ is to increase the handover successful rate when the source cell or the target cell is
congested. In this case, multiple handover triggers may increase the chance for the handover request to go through. Also, according to standards eNB can not
send HO command until SecurityMode is completed. Setting reportAmount to ‘r8’ in this case will increase the chance for HO command to be sent.
For ‘Below-Serving-Floor’ trigger, the recommended value of reportAmount is ‘8’. The reason to set to ‘8’ for ‘Below-Serving-Floor’ trigger is similar to the reason
given above for ‘Intra-Frequency-Handover-Trigger’. Also, since ‘Below-Serving- Floor’ trigger has lower priority than ‘Intra-Frequency-Handover-Trigger’, event
A2 measurement report will be dropped in case eNB is processing event A3 LTE intra-frequency handover measurement report. Setting reportAmount to ‘8’ in
this case will increase the chance for the event A2 measurement report to be processed.
For ‘Automatic-Neighbor-Relation’ trigger, the value of the corresponding reportAmount should be set to ‘4’.
The reason to set the reportAmount to ‘4’ is that with ANR feature, one PCI measurement report will trigger one ‘Report-CGI’ request. Multiple ‘Report-CGI’
requests may increase the chance for a UE to find the ECGI corresponding to a PCI reported.
For ‘Report-CGI’ trigger, the value of the corresponding reportAmount must be set to ‘1’.
timeToTrigger:
timeToTrigger is used in several process: Measurement identity removal; Measurement identity addition/ modification; Measurement object removal;
Measurement object addition/ modification; Reporting configuration removal; Reporting configuration addition/ modification; Quantity configuration; in general in
Measurement report triggering; Measurement related actions upon handover and re-establishment
For ‘Automatic-Neighbor-Relation’ trigger, timeToTrigger can only be set to the same value of the timeToTrigger (corresponding to triggerQuantity = ‘rsrp’ if it
exists; otherwise, corresponding to triggerQuantity = ‘rspq’) used for ‘Mobility-Intra- Freq’.
For ‘Report-CGI’ trigger, timeToTrigger is not used but must not be left unset.
maxReportCells:
For ‘Mobility-Intra-Freq’, the default value of the corresponding maxReportCells is set to ‘1’. Setting the parameter to a value greater than ‘1’ will not be useful in
LA3.0/TLA3.0. This is because in the current target cell selection algorithm, only the best neighbor cell will be considered as the handover target. The setting of
the parameter will need to be updated once more candidate cells are considered in handover target selection.
For ‘Automatic-Neighbor-Relation’ trigger, the value of the corresponding maxReportCells should be set to the maximum value of ‘8’. This is to ensure UE to
report as many new neighbor cells as possible in a short time.
For ‘Report-CGI’ trigger, maxReportCells is not used but must not be left unset.
triggerTypeEutran:
For measurementPurpose = ‘Leaving-Coverage-Alarm’, triggerTypeEUTRA should be set to ‘eventA1’.
For measurementPurpose = ‘Entering-Coverage-Alarm’ or ‘Below-Serving-Floor’, triggerTypeEUTRA should be set to ‘eventA2’.
For measurementPurpose = ‘Mobility-Intra-Freq’, triggerTypeEUTRA should be set to ‘eventA3’.
For measurementPurpose = ‘Mobility-Inter-Freq-to-EUTRA’, an instance of triggerTypeEUTRA should be set to ‘eventA3’. Another instance of triggerTypeEUTRA
should be set to ‘eventA5’.
For measurementPurpose = ‘Automatic-Neighbor-Relation’, the recommended setting for triggerTypeEUTRA is ‘eventA3’. triggerTypeEUTRA may also be set to
‘eventA4’ or ‘eventA5’ for testing purpose. Only one instance of triggerTypeEUTRA for measurementPurpose = ‘Automatic-Neighbor-Relation’ can exist.
For measurementPurpose = ‘Report-CGI’, triggerTypeEUTRA should be set to ‘periodicalReportCGI’
Non-GBR use the basic parameters (without offset). Other QoS class may add offset on top of the basic
parameter.
Event B1 CSFB is not concerned because in case of CSFB, the call is classified into non-GBR and not impacted by
the offset parameter
If there is no high priority carrier and no medium priority carrier, there must be low priority carrier if there is any
carrier.
When E-MCTA is triggered, it takes as an input neighboring RAT/carriers of the serving LTE cell, it applies
filters, and it provides as an output a sorted list of candidate RAT/carriers for RRC Measurements. This
functionality is part 1 of e-MTCA feature.
The RRC measurement configuration function relies on part 1 of the E-MCTA process since the list of
Measurement Objects towards which the UE performs measurements is the candidate RAT/carrier list
output of the E-MCTA process plus the mandatory intra-frequency measurements like A3 for intra-frequency
mobility or A2-floor for blind redirections.
Output ‘Candidate RAT/Carrier Sorted List’ indicates for each candidate RAT/carrier:
Target Measurement Configuration (B2/B1, A5/A3, A4 or none)
Priority (0-lowest through 7-highest)
Whether a Measurement Gap needs to be configured (yes, no)
B1/B2 are used for inter RAT measurements, A3/A5 for intra or interfreq measurements and A4 for ANR
interfreq
Example: live streaming (QCI=2) may be configured in GERAN with a low mobility priority (high eMctaPriority) in
order to favor the UTRA layer for high throughput data services
If the total number of measurements required by the eNodeB exceeds maxMeasIdForMultipleMonitoring, then
the eNodeB reduces to maxMeasIdForMultipleMonitoring this total number. The lowest priority measurements
are removed by this truncation. Prevailing measurements are the intra-frequency ones: for example with
measurement purpose set to Mobility-Intra-Freq; or Entering-Coverage-Alarm; or Leaving-Coverage-Alarm. It is
recommended that maxMeasIdForMultipleMonitoring be set to a value that will always accommodate the
essential intra-frequency measurements along with the anticipated inter-frequency/inter-RAT measurements at
the top of the eMCTA sorted list so that the eNodeB would need to remove only the lower priority inter-
frequency or inter-RAT measurements at the bottom of the eMCTA sorted list.
The algorithm also limits the number of allowed layers for multiple monitoring (requiring MG) to a maximum
called maxNbCarriersForMultipleMonitoringUsingMeasGaps which is configured by the Operator. In essence, this
parameter limits the number of RRC measurements needing a Measurement Gap (MG) which can be configured.
This could result in the lower priority RAT carriers in the ‘candidate RAT/Carriers Sorted List’ needing a MG to be
removed from the list, if the total number needing a MG in the Sorted List exceeds this parameter value.
Default=5
So, only when all TRBs fail CAC (i.e. no TRBs can be established at target cell) will lead to HO cancel.
If at least one TRB succeeds (and at least one TRB fails) in RRM in target cell, the HO execution will continue.
The TRB that was successful in RRM will be handed over to the target cell and the failed TRB will be released.
In case the eNB is requested to release a non-existent radio bearer Id or a duplicate request to delete the same
radio bearer Id, the eNB should respond back with an appropriate cause value to indicate this.
If any UL RRC messge is received with integrity verification failure, the procedure is aborted, and the eNB
initiates the S1 UE context release procedure.
In case of inter-eNB handover trigger, the Source eNB will initiate the X2-AP handover preparation providing in
X2-AP HANDOVER REQUEST the necessary information to prepare the handover in the Target eNB.
If the data forwarding is enabled in the Source eNB then the Source eNB will propose to the target eNB to
perform DL data forwarding via X2.
The eligibility to DL forwarding of each supported QoS Label (QCI) is configured via MIM. If Integrity Protection
and Confidentiality services are enabled, AS security data is also included in the X2 HANDOVER REQUEST
message.
Target eNB prepares the handover based on the received request from the Source eNB and includes in
HANDOVER REQUEST ACKNOWLEDGE the RRC CONNECTION RECONFIGURATION message to be transmitted
transparently by the Source eNB to the UE.
If the data forwarding is enabled in the Target eNB then the Target eNB will accept the proposal from the
Source eNB to perform DL data forwarding via X2 by establishing the one DL X2 tunnel for each E-RAB subject
to forwarding.
The eligibility to DL forwarding of each supported QoS Label (QCI) is configured via MIM. After this step the
target eNB is ready to receive UL transmission from the UE and DL data forwarded over X2 from the Source eNB
if configured previously.
If AS security services are enabled, the target eNB also derives keys that will be used for integrity
protection and ciphering.
This message contains the security information for the NAS signalling messages that will have integrity
protection.
During the handover completion the UL data transmission occurs normally in the Source or the Target eNB
if DL data forwarding was configured, the Source eNB continue to forward via X2 interface the received DL S1
packets until reception of GTP-U End Marker or resources are released
DL data forwarding was configured, the Target eNB shall transmit the over the radio the DL X2 received
packets until reception of X2 GTP-U End Marker or resources are released. Only after that DL S1 packets are
transmitted.
When the source eNB receives the S1 HANDOVER COMMAND from the MME, it starts timer
tS1RelocOverallForS1Handover. The timer is stopped when the source eNB receives the S1 UE CONTEXT
RELEASE COMMAND message from the MME. If the timer expires before the source eNB receives the S1 UE
CONTEXT RELEASE COMMAND message from the MME, then the source eNB sends an S1 UE CONTEXT
RELEASE REQUEST message. The expected result is that the MME will send an S1 UE CONTEXT RELEASE
COMMAND message to the source eNB, and the UE context will be released.
2. eNB will reject the incoming handover requests (including handover requests for emergency calls) to a cell
reserved for operator use in the PLMN serving the UE with cause ‘ho-target-not-allowed’ unless
ActivationService::isIncomingHoToReservedCellBasedOnSpidAllowed is set to ‘True’, and SPID (Subscriber Profile
ID) of the UE is specified in PlmnIdentity::spidAllowedInReservedCells. This parameter specifies the Subscriber
Profiles ID of the UEs allowed to be handover into a cell reserved for operator use.
Feature Benefits:
This feature enables Alcatel-Lucent's customers to perform handover testing with dedicated test UEs without
permitting commercial users to access the cells. This allows for performing parameter tuning prior to putting into
commercial service a newly deployed eNodeB or cluster of eNodeBs.
Upon receipt of the S1AP UE CONTEXT RELEASE REQUEST, the source MME sends a S1AP UE CONTEXT
RELEASE COMMAND message to the Source eNodeB.
The completion in the ENB ends upon receipt S1AP UE CONTEXT RELEASE COMMAND and sending by the ENB
to the source MME of a S1AP UE CONTEXT RELEASE COMPLETE of a or upon guard timer expiration.
When eNB receives a UE event B2 or event B1 (for CS fallback) measurement report with measurement
purpose set to ‘Mobility-Inter-RAT-to-UTRA’, and with valid reported cells (reported PhysCellIdUTRA-FDD
Corresponds to UtraFddNeighboringCellRelation::physCellIdUTRA of an instance of UTRA neighbor), LTE to
UTRAN PS handover procedure will be triggered if all of the following conditions are satisfied:
eNB will send a Handover Required message to the MME and start timer TS1relocprep with duration
PsHoToUtraTimersConf::tS1RelocPrepForPsHandoverToUtra (the PsHoToUtraTimersConf instance is
pointed to by the RncAccess::psHoToUrtaTimerConfId associated with the selected
UtraFddNeighboringCellRelation).
If timer TS1relocoverall expires, eNB considers the UE to have lost radio coverage and will trigger the
release of all UE associated resources by sending an UE Context Release Request to MME and release all UE
associated resources in eNB.
Target RNC-ID
Direct Forwarding Path Availability (RncAccess::directFwdPathAvailability)
Source to Target Transparent Container includes: Cell-ID (UtraFddNeighboringCellRelation::cId)
UTRAN FDD small cells can have their own dedicated carriers or can share the carriers with UTRAN FDD macro
cells. A UTRAN FDD small cell can only be assigned to one of the PSC (Physical Cell-ID for Small Cell) dedicated
to UTRAN FDD small cells defined under UtraFddneighboringSmallCellLayerRelation::
pciUtraFddForOpenSmallCellList. Unlike UTRAN FDD macro cells where each of them uses a unique PSC, multiple
small cells in the same coverage area are allowed to share the same PSC. For this reason, a LTE cell does not
maintain per UTRAN FDD open small cell neighbor relation. Instead, information of all small cells under a UTRAN
FDD frequency is maintained in UtraFddNeighboringSmallCellLayerRelation MO. This will not cause a problem
since PS handover is not supported for mobility from LTE to UTRAN FDD small cells. Only redirection (blind or
measurement-based) is supported with L115393.
Following is a list of mobility functionalities that are not supported by L115393 for LTE to UTRAN FDD small cell.
• PS handover to UTRAN FDD small cell is not supported. If RncAccess::isSmallCellGateway = ‘True’, RncAccess::psHandoverUtraEnabled must be set to ‘False’
under the same instance of the MO.
• SRVCC to UTRAN FDD small cell is not supported. If RncAccess::isSmallCellGateway is set to ‘True’, RncAccess::srvccType must be set to ‘disabled’ under the
same MO.
• CSFB to UTRAN FDD small cell. CSFB to UTRAN FDD is not supported for small cell dedicated or shared carriers.
For each instance of UtraFddNeighboringFreqConf that includes an UtraFddNeighboringSmallCellLayerRelation MO, ServiceTypePriorityConf:: eMctaPriority must
be set to ‘serviceOrQci-notallowed-in-RAT-carrier’ for all settings of ServiceTypePriorityConf::serviceType including ‘csfbByIdleUE’, ‘emergencyCsfbByIdleUE’,
‘csfbByConnectedUE, and emergencyCsfbByConnectedUE’.
• RIM for system information and RIM for cell load procedures through small cell gateway are not supported. If RncAccess:: isSmallCellGateway is set to ‘True’,
both RncAccess::rimForUtraSiEnabled and RncAccess::rimForCellLoadEnabled must be set to ‘False’ under the same MO.
• Preventive offloading to small cell dedicated carrier is not supported. Since PS handover or SRVCC ‘PS and CS’ are used for preventive offloading to UTRAN FDD
carriers, eMCTA will not select an UTRAN FDD small cell dedicated carrier as preventive offloading candidate for the reason if RncAccess::isSmallCellGateway is
set to ‘True’, RncAccess::psHandoverUtraEnabled is set to ‘False’ and RncAccess::srvccType is set to ‘disabled’ under the same MO.
• Mobility of VoIP calls to UTRAN FDD small cell is not supported. Since a candidate for VoIP mobility must support either VoIP or SRVCC ‘PS and CS’, eMCTA will
not select UTRAN FDD small cell dedicated carriers for VoIP call mobility because they do not support VoIP
(UtraFddNeighboringSmallCellLayerRelation::voiceOverIpCapability is set to ‘incapable’) and they do not support SRVCC (If RncAccess::isSmallCellGateway is set
to ‘True’, RncAccess::srvccType must be set to ‘disabled’ under the same MO).
• UTRAN ANR for UTRAN FDD small cell is not supported. When UTRAN FDD ANR is activated, UE will not be configured to perform PCI search for the carriers
dedicated to UTRAN FDD small cells. This can be achieved by not configuring ANR measurement for the dedicated small cell frequency (not configuring
MeasurementIdentityConf instance with MeasObjectLink which points to MeasObject instance of dedicated small cell frequency and measurementPurpose set to
‘Automatic-Neighbor-Relation’). For UTRAN FDD small cell shared frequency, if UE measurement report contains a PSC that belongs to UTRAN FDD small cells (if
the reported PSC belongs to the list of UtraFddNeighboringSmallCellLayerRelation::pciUtraFddForOpenSmallCellList), the measurement report will be ignored and
no UTRAN FDD neighbor relation will be set up for this neighbor.
Another new parameter related to the UTRAN FDD small cell neighbor, isSmallCellGateway, is added under the
RncAccess MO. This parameter indicates whether a rncAccess is a small cell gateway.
This feature provides system information of the UTRAN/GERAN cells when it is released from LTE and redirected
to UTRAN/GERAN carrier. The enhanced redirection functionality can be used for CSFB redirection or non-CSFB
redirection from LTE to UTRAN/GERAN, except for the case when the redirection is triggered by event A2 ‘Below-
Serving-Floor’ (blind redirection is triggered in this case)
Only release 9 UE that supports e-RedirectionUTRA-r9/e- RedirectionGERAN-r9 capability can make use of
system information of UTRAN/GERAN cells. Release 8 UE will ignore the system information
eNB is able to automatically retrieve system information from the neighbor UTRAN/GERAN cells through RIM
procedures
Dependency:
ePC core and 2G/3G network to support System Information exchange procedure using RIM protocol
UTRAN to support special admission handling based on CSFB and emergency indicator
When eNB is to stop the RIM procedure, it will stop timer with duration
UtraSiTimersConf::timeToWaitForEnbDirectInfoTransfer if it is running and remove the stored System
Information of the neighbor UTRAN cell.
eNB will then send a S1-eNB Direct Information Transfer to stop the RIM procedure, and start timer Trir waiting
to receive the S1-eNB Direct Information Transfer from MME in response to the request to stop the RIM
procedure
If Trir expires, eNB will retry the procedure for up to three times. If all tries fail, eNB will raise an alarm
indicating ‘UTRAN SYS info update stop failure’
The other procedures (SI Update, SI Update Stop and SI Update End) are similar to UTRAN and not described
here.
Otherwise, eNB will perform the LTE to UTRA blind redirection procedure to move the UE to UTRA for CS
fallback. CS Fallback Triggered by Redirection.
Whenever redirection from LTE to UTRA is performed for CS fallback except in the case when redirection is
triggered by UE entering ‘Below-Serving-Floor’ area, target cell System Information will be provided in the
RRCConnectionRelease message for redirection assistance if following conditions are met.
(1) Security must be activated and SRB2 and at least one DRB must be set up before triggering the HO.
It will already have been for a UE context modification, but not for an initial context setup.
(2) UTRA capabilities will have to be requested in the case of CS FB triggered at initial context setup.
This parameter is used to indicate whether a GERAN neighbor cell supports dual transfer mode (DTM). For a
GERAN cell that supports DTM, it allows simultaneous transfer of circuit switched voice and packet switched data
over the same radio channel
For CS fallback to GERAN with CCO, if both UE and target GERAN cell support DTM, in the UE-Context-Release-
Request sent to MME, eNB will include the cause ‘cs-fallback-triggered’. If either UE or the target GERAN cell
does not support DTM, in the UE Context Release Request sent to MME, eNB will include the cause ‘ue-not-
available-for-ps-service’
For CS fallback to GERAN with enhanced/basic redirection, if UE and at least one GERAN neighbor cell under the
redirected carrier support DTM, in the UE-Context-Release Request sent to MME, eNB will include the cause ‘cs-
fallback-triggered’. If either UE or all of GERAN neighbor cells under the redirected carrier do not support DTM,
in the UE Context Release Request sent to MME, eNB will include the cause ‘ue-not-available-for-ps-service’.
The big difference is that the MME identifies DTM as a CS + PS HO and sends the voice to the MSC and the PS
bearer to the SGSN. The MME does a lot more synchronization for DTM HO than for non-DTM HO.
Two parameters under Enb MO are used by eNB to generate the authentication challenge parameter (rand) sent to UE and
to 1xRTT network. The two parameters are:
• randUpdateInterval – This parameter specifies the rand update interval for 1xCSFB. This parameter must be configured
with the same value for all eNBs to allow synchronization in rand determination. This parameter must be set if
ActivationService::isRel8CsfbTo1XRttEnable is set to ‘True’.
• seed – This parameter specifies the seed value for 3G 1X authentication. The parameter must be configured to the same
value for all eNBs to allow synchronization in rand determination. This parameter must be set if
ActivationService::isRel8CsfbTo1XRttEnable is set to ‘True’. It is a write only parameter and should be stored securely.
The primary use of the reference cell is that it becomes the virtual accessed cell for setting up connections. So in
all the normal places where the accessed cell is used by the CallP (eNB software) of the legacy system for
eCSFB, the reference cell is used. For example, to map from PCI to CGI when determining handover candidates.
The 1xRTT reference cell related parameters are under OneXRttReferenceCell MO.
The parameters included in OneXRttAccessBarring MO are used to populate the fields in AC-BarringConfig1XRTT-
r9 IE of SIB8. There is no automatic method for eNB to retrieve load information from 1xRTT network, operators
will need to populate the parameters in
OneXRttAccessBarring MO in order for the access barring information to be broadcast through LTE.
If any of the above conditions are not satisfied, eNB will not be able to configure UE to perform 1xRTT
measurement. In this case, enhanced 1xCSFB will not be selected. Blind redirection Release 8 1xCSFB procedure
or dual Rx 1xCSFB procedure will be performed instead.
The UE radio access capability parameters carried in UE-EUTRA-Capability IE that are checked by eNB when selecting a
1xCSFB procedure including:
• rx-Config1XRTT – the parameter is set to ‘single’ for a single Rx UE and set ‘dual’ for a dual Rx UE
• e-CSFB-1XRTT-r9 – the parameter is set to ‘supported’ if the UE supports enhanced 1xCSFB
• FeatureGroupIndicator – FGI#16 and FGI #24 are set to ‘1’ if the UE supports non- ANR periodical measurement on 1xRTT
• interRAT-NeedForGaps – the parameter indicates whether measurement gaps are needed for inter-RAT measurement
• SupportedBandList1xRTT – list of 1xRTT bandclass the UE supports
After eNB finds out whether UE is single Rx or dual Rx, and whether the UE supports e- CSFB-1xRTT-r9, it determines
whether periodical measurement to 1xRTT is possible if enhanced 1xCSFB or measurement based R8 1xCSFB procedure is
considered. If it is possible, eNB will configure UE to perform periodical measurement on 1xRTT. After UE measurement
report is received, final decision on the selection of 1xCSFB procedure is made based on the procedure enable/disable status.
Document History
PCI Collision: occurs when in a given location, the signals from two different cells radiating the same PCI
are received by a UE. In the worst situation, the UE may be unable to access either of the two cells due to
the interference generated. The coverage gap or blackout may be quite important. At best, the UE will be
able to access one of the cells but will be highly interfered.
PCI Confusion: appears when a given cell, has two neighbors sharing the same PCI. The UE reports
measurements from two cells having the same PCI. The eNB will not know which of the two cells the report
relates to. In the best case, the eNB may request the UE to report the ECGI before HO. In the worst case it
may lead to a high rate of HO failures. The two cells having the same PCI do not interfer as much as in
case of collision because they are far from each other.
A PCI collision is considered critical enough that it shall be addressed within one hour of detection. A PCI
confusion can wait until the next maintenance window to be addressed. To resolve the difficulty of ensuring
that a newcomer eNB modifies its PCI rather than a well-established eNB, two different time periods can be
defined. For example, a newcomer eNB would be required to perform a PCI modification within 15 minutes
of detection or beginning of the maintenance period, whereas a well-established eNB would do so in the
remaining part of the hour. Once one eNB has modified a conflicting PCI and communicated the change to
its neighbours, the need for the other eNB to change its PCI disappears.
In this way, our solution guarantees that a newcomer eNB will change its PCI before a well-established eNB
attempts to do so.
The feature is limited to metro cells, as the main topic of this feature is the Interference between metro
cells and also between the macro cells and the metro cells that cover the same area. In such case, it is
better to change the metro cells PCI:
- Because just changing the macro cell PCI may not solve all the interference
- Because changing the metro cell has only local impact while change the macro cell may generate a wave
of changes in the network that would impact many users.
The feature is not compatible with PciMod3Maintained= enabled, this parameter implies that the PCI mod 3
cannot be modified, which is the opposite of the aim of this feature.
The young/mature status is taken into account, what means that the timer values related to this status are
used. Even if the main interferer of a metro is the macro, others metro may have impact. In such case,
better change first a young than a mature Metro cell.
• Note that the groups keep the ordered PCI mod 3 information meaning that information is:
• What is the best PCI mod3
• What are the PCI mod3 in the best group
The number of cells that can be included in SIB4 and Measurement Configuration messages is limited (16
cells in SIB4 and 32 in Measurement Configuration). Preference will be then given to neighbour cells with
the biggest absolute value offset, and if a second criterion is needed, cells with the lowest PCI will be
preferred to cells with the highest PCI.
The HO White List is composed of the neighbor relations with noRemove=true and noHOReselection=False
The HO Black List: noRemove=true and noHOReselection=true
In the same manner, there is also X2 White List, X2 Black List, X2 HO Black List (managed by the eNB and
not forwarded on the air)
An ANR enabled eNB is able to detect intra-freq neighbor cells. Just after the startup of a cell (cell enters in
‘unlocked, enabled’ state), if the attribute LteCell::Lte{Intra,Inter}FrequecyAnrState is set to
‘notCompleted’, the detection of neighbor cells for the started cells is performed in an ‘accelerated’ way
during a transient period. During this period, all or part of the RRC connected UEs located in the cell will be
configured to perform ANR specific PCI measurements. This transient period is entitled ‘Active phase’ here.
The Active phase ends (and the Dormant phase starts) when a preconfigured amount of PCI related
measurements (regardless whether the PCI is newly or already measured) and PCI related measurements
without any new PCI are reported. In the Dormant phase, if a new PCI is detected upon e.g. regular RRC
connected UE measurements necessary for the HO control function an ANR Wake-up phase is launched
(UEs are configured with ANR specific measurements. Should the ECGI for the PCI be reported the process
of configuration is stopped). The Dormant phase is also left for the Monitoring phase regularly every hour
to check whether the PCI/ECGI association is still valid. In both Active and Wake-up phases, if a new PCI is
detected, the UE is then requested to report the associated ECGI to be read in the SIB1 of the detected
neighbour provided that it supports the function and it is not configured with GBR RB (the UE shall be put
in DRX mode in order to perform the ECGI measurement).
3. When the message is received by the MME, SON Configuration Transfer IE is transparently copied to
MME CONFIGURATION TRANSFER message that is sent to the eNB identified through received Target eNB-
ID.
5. this one sends back ENB CONFIGURATION TRANSFER message to the MME from which it received MME
CONFIGURATION TRANSFER message. Here, SON Information Reply IE, containing X2 TNL Configuration
Info that carries the IP address, is used.
6. As previously, when receiving the message, the MME transparently copies SON Configuration Transfer IE
into MME CONFIGURATION TRANSFER message that is sent to the source eNB.
7. Source eNB receives the message and extract target eNB IP address from it.
To increase the chance to find ECGI, before the timer is expired and until the ECGI is found:
Configure a number of capable UEs selected by dispatch function to perform ANR measurement. The
maximum number of UEs selected by the ANR dispatch function is limited by
LteIntraFrequencyAnr::uEContributionTargetInActivePhase divided by four
(remark: originally the ANR dispatch period was planned to be 1 minute in length and the range of
values allowed in the MIM was targeted with that in mind, but then it was shortened to 15 seconds in
the final design; this is why the value has to be divided by four)
Request the UEs that report the same unknown PCI to search for the associated ECGI
Once the ECGI is found, eNB will start the process to setup the X2 link if it does not already exist and if
LteNeighboringFreqConf::anrInitiateX2Setup is set to ‘True’.
In the case of IRAT CGI acquisition, T321 is equal to 8s. Thus, multiple consecutive DRX cycles are
necessary. For intra-LTE ANR and CSFB, the UE could need up to 2 DRX cycles, so multi-cycle DRX can
apply to all DRX applications, just the DRX application for UTRA ANR requires many more cycles than that
for intra-LTE ANR or CSFB.
The two ‘noRemove’ flags mentioned above are used by operator to control whether a neighbor relation or
an X2 link can be deleted by the ANR garbage collection function regardless the neighbor relation or the X2
link is created by ANR or is provisioned by operator. When a neighbor relation or an X2 link is in the black
list (LteneighboringCellRelation::noHO is set to ‘true’ or X2Access::noX2HO is set to ‘true’), the
corresponding ‘noRemove’ flag should also be set to ‘true’. This is to avoid a noHO neighbor cell or noX2HO
X2 link becomes obsolete and be removed, and the neighbor cell is reported back by UE again as a new
neighbor cell or a new handover target candidate, or the X2 link is set up again.
The operator may choose to base ANR neighbor replacement on counts of ANR MRs, counts of mobility MRs, or
the sum of the two counts. This selection is controlled by the parameter nRPriCountMode.
If the parameter nRReplaceHysteresis is set to a value greater than 0, a new NR must be nRReplaceHysteresis
percent greater than an existing neighbor in order to replace that neighbor. This option is available to avoid
ping-pong of the NR replacement.
If the value of TTT is unset, the HO Optimization algorithm will not adjust TTT
If a3OffsetMin=a3OffsetMax, the HO Optimization algorithm will not adjust a3Event
If cioMin=cioMax, the HO Optimization algorithm will not adjust CIO
The value of the cellShrinkForS1LinksOutage parameter is dependent on frequency, bandwidth, downlink power,
number of antennas, and other factors. The upper bound value to be used is determined from the following
relationship:
referenceSignalPower - cellShrinkForS1LinksOutage >= minRefSigPower
Document History
1 eUTRAN Sharing 7
1.1 Introduction 8
1.2 Broadcast of PLMN ID of Operators 9
1.3 MOCN License and OAM Impact 10
1.4 Gate Core Network (GWCN) Support 11
1.5 Object Model 12
1.6 Mobility Aspects 13
2 Commercial Mobile Alert System (CMAS) 17
2.1 CMAS Principle 18
2.2 Impacts of CMAS on the DL Scheduler 19
2.3 eNB CMAS Parameters 20
2.4 CMAS Evolution 21
3 Enhanced Multimedia Broadcast Multicast System (eMBMS) 24
3.1 eMBMS Principles 25
3.2 MBSFN Area 26
3.3 MBMS Service and MBSFN Synchronization Area 27
3.4 MBSFN Architecture (3GPP) 29
3.5 Alcatel-Lucent MBSFN Architecture 30
3.6 MBMS Session 31
3.7 MBMS Radio Channels 33
3.8 PMCH Scheduling 34
3.9 SIB2 and SIB13 Usage for MBMS 35
3.10 Management of Several MCH per MbsfnArea 36
3.11 MCCH and MSI Multiplexing 37
3.12 MCE Functions 38
3.12.1 MCE Distributed Architecture 39
3.12.2 Multi Vendors MCE 40
3.12.3 MCE Configuration 41
3.12.4 MBSFN Area Block Partitioning 45
3.12.5 Interaction with SPS 46
3.12.6 Interaction with OTDOA 47
3.13 Supported Bands for eMBMS 49
3.14 eMBMS Object Model 50
4 Location Based Services (LBS) 56
4.1 Positioning based on OTDOA 57
4.2 Positioning Reference Signal (PRS) 58
• MCE (Multi-cell/Multicast Coordination Entity) distributed and eNodeB integrated => coordinates the
radio resources in MBSFN areas.
• The MCE allocates radio resources semi-statically for eMBMS services according to the M3-AP
procedures (session start and session stop).
Synchronization of cells within an MBSFN is accomplished using GPS timing and a SYNC protocol between
the eMBMS Gateway and the eNB. The GPS timing is used to time and phase align the eNB transmissions at
their respective antennas, and the SYNC protocol is used to align frame numbers for the eMBMS
transmissions.
Upon receiving the session start request, the MCE performs admission control, allocates IP multicast
addresses, and allocates and configures the radio resources. The MCE provides the relevant session
parameters (e.g. the multicast address) to the eNodeB by means of the ‘Session Start Request’ message.
Likewise, the MCE transfers the radio resource configuration within a ‘scheduling information’ message to
the eNodeB. Upon receiving these messages, the eNodeB notifies the UEs about a change of MCCH
information and subsequently provides the updated MBMS radio resource configuration information within
the MBSFNAreaConfiguration message. The eNodeB also joins the transport network IP multicast address.
The MBMS control information is mainly carried by the MCCH; the BCCH merely carries the control
information needed to acquire the MCCH and to detect MCCH information changes. Transmission of MBMS
control information on the BCCH (i.e. in SIB13) is performed according to the usual procedures for SIBs.
The MBSFNAreaConfiguration on MCCH is repeated a configurable number of times within the modification
period.
When an MBMS session is started, E-UTRAN notifies the UEs about an MCCH information change. This
change notification is provided a configurable number of times per modification period, by means of a
message on the PDCCH using Downlink Control Information (DCI) Format 1C, using a special identifier, the
MBMS Radio Network Temporary Identifier (M-RNTI). This message indicates which of the configured
MCCHs will change. UEs that are interested in receiving an MBMS session that has not yet started should
try receiving the change notification a minimum number of times during the modification period. If the UE
knows which MBSFN area(s) will be used to provide the MBMS session(s) it is interested in receiving, it has
to meet this requirement only for the corresponding MCCH(s).
As both, control and traffic channel (MCCH, MTCH) are mapped to the Multicast Channel (MCH), there is a
kind of ‘chicken-and-egg’ problem as the one channel (MCCH) contains information how the other channel
(MTCH) is organized and how to access it. This problem is solved while introducing the SIB Type 13, which
provides the following information:
MBSFN identity (MBSFN ID)
Non-MBSFN region length (1, 2 OFDM symbols)
MCCH configuration
The MCCH configuration provides information on the repetition period for the MCCH, the MCCH offset as
well as the actual subframe, where the MCCH is transmitted in as well as the MCS used for the MCCH.
Coming back to our previous example, all this information may lead to the shown scenario. Selected
repetition period of 32 means that every 32 radio frames the MCCH occurs in one of the MBSFN subframes.
Which one is also determined by information part of the SIB Type 13, including the used modulation coding
scheme. It is important to note that for the MCCH four modulation coding schemes are allowed: MCS index
2, 7 (both QPSK), 13 (16QAM), 19 (64QAM).