This is the first part in LTE KPI Optimization and more related articles will be published

soon. So, let us get started without wasting any further time.

When the UE wants to attach or connect to the network, it needs to setup a RRC
Connection as explained in my article LTE Network Entry Steps . But before that it needs
to get synchronized in uplink. This is done by sending a RACH preamble (Msg1) to the
eNB and eNB responds with a Random Access Response (RAR aka Msg2). This is
where the UE sends a Msg3 also known as the RRC Connection Request and this marks
the attempt for the RRC Success Rate KPI. This message contains the objective of the
connection and based on that it is subdivided into following major categories:

 Mo-data : Usually used for UE coming back from idle mode if it has data to send or if it
has to make a call
 Mo-signaling : Most commonly observed for TAUs and Attach messages
 Mt-access : Idle UE responds to a paging message
 Emergency
 High Priority Access

It also contains a UE identity which can be a TMSI value if the UE was already previously
attached to LTE and had a TMSI allocation or it can be a random value indicating that the
UE does not know about its TMSI or it might be coming from another RAT.

Based on this request, the eNB sends a RRC Connection Setup message which contains
the information of SRB and some basic radio parameters like power control, SRI & CQI
periodicity.

Once, the UE gets the RRC Connection Setup, it makes the changes based on the
instructions in the message and then responds with RRC Connection Setup Complete
message. This message also contains the NAS information if the UE intends to send it.
The eNB pegs RRC attempt counter when it receives the RRC Connection Request and
the process is deemed successful on the receipt of the RRC Connection Setup Complete
message.

Common Failures In RRC Setup Phase

In order to maintain and optimize the RRC KPI, one should know the major issues that
can cause a RRC setup failure.

RRC Setup Failure Due To No Response

This is the most common RRC failure which is present in every network. Most of the
failures in the RRC stage are due to no response from the UE. This means that the eNB
receives RRC Connection Request message from the UE and responds with a RRC
Connection Setup message but does not receive or is unable to decode the RRC
Connection Setup Complete message.

Now let’s understand why this happens. The RRC Connection message is usually around
7 bytes in length while the RRC Connection Setup Complete message may contain the
whole NAS information (like TAU or Attach Request) and its size can vary from as small
as 8 bytes to as big as over 100 bytes. Consider that a UE near cell edge with limited
power sends a RRC Connection Request. Since, it is only around 7 bytes, it will need a
small number of RBs so power per carrier will be high. But when it needs to send RRC
Connection Setup Complete which is around 100 bytes, it will need a bigger number of
resources even if the message is fragmented. So, the average power per carrier will be
reduced leading to a higher probability that the message may not be decoded at the eNB.
This can also happen if there is interference on the cell as it will make it further difficult
for the eNB to decode the message. It can also happen if the UE fails to decode RRC
Connection Setup message so it will never send the RRC Connection Setup Complete
message.

RRC Rejections

This is the second issue that can happen but it is usually much less observed in
commercial networks compared to the failures due to no response. In these cases, the
eNB rejects the incoming RRC Connection Request by sending a RRC Reject message.
This is mostly observed when eNB experiences congestion and there are not enough
resources left to assign to a new user requesting a RRC Connection.

If the PUCCH is congested, the RRC connection can be rejected. PUCCH carries HARQ
ACK/NACKs, CQI and SRIs. If the PUCCH resources are not available, users will not be
able to send CQIs and the eNB cannot schedule without CQI information. Usually
vendors implement PUCCH in a way that when PUCCH utilization is increased, the CQI
interval is increased. For example, users sending CQI with an interval of 10ms will be
shifted towards 40ms in order to increase the capacity of the PUCCH.

But when no further capacity is available, the eNB needs to put a limit on new incoming
connections resulting in RRC Rejections. Similarly, RRC Rejections can be seen if the
active UE count increases beyond the board limit or if the CAPS exceed the limit. The
details related to troubleshooting and optimizations for such issues is given below.

Optimization For RRC Success Rate KPI

The following procedures are usually used based on different scenarios

Conventional Method : Physical Optimization

The easiest and conventional method is the physical optimization. For instance, down-
tilting a cell will reduce the coverage and remove the far-away users. This will reduce the
probability of RRC failure due to no response. However, there will be issues that might
not be resolved by the conventional approach so I have listed down other methods that
might come in handy.

Relevant Timers

There are two relevant timers for RRC Success Rate KPI.

The first timer is maintained on the UE and it is the famous T300. UE starts it after
sending the RRC Connection Request and stops it at the receipt of RRC Connection
Setup or Reject message. If this timer is too small, the UE will stop waiting for the RRC
Connection Setup message and the RRC procedure will fail. So, increasing this timer can
help in this phase.

Secondly, eNB has an internal timer (different vendors have different names for it) which
the eNB starts after sending the RRC Connection Setup message. It stops this timer after
successfully receiving the RRC Connection Setup Complete message. So, if this timer is
small and the UE is trying to send the RRC Connection Setup Complete with
retransmissions, then the eNB will consider it a failure as soon as the timer expires. So,
increasing this timer might also help in certain scenarios.

Coverage Enhancement & Power Control

The RRC failures due to lack of response from UE can also be caused if the power
control on the PUSCH is not correct or if it is too conservative. For instance, the power
control on PUSCH depends on the P0 Nominal value as well as Alpha factor. Different
vendors use different settings here like using a low P0 Nominal value (for example -100
dBm) with a higher Alpha factor of around 0.9 or 1 or a using a high P0 Nominal value
(for example -70dBm) with a smaller Alpha factor of 0.7 or 0.8. But if both the P0 Nominal
and Alpha factor are low then the UE will use a smaller power value to send the RRC
Connection Setup Complete and therefore, the chances are that it will not be decoded
correctly.

In case there is interference on the cell, then features which mitigate interference should
be enabled. For instance, enabling Interference Rejection Combining can provide good
gains in such scenarios.

Mobile Originating Signalling RRC Success Rate

Usually Mo-Sig RRC Success Rate is lower than others. The reason is once again linked
to the size of the MSG-5 (RRC Connection Setup Complete). For a normal Mo-data or
Mt-access, the size of RRC Connection Setup Complete message is around 8 to 10
bytes but for Mo-signalling, it can vary and usually is above 50 bytes. This is because
Mo-signalling RRC Request is usually used for NAS signalling messages like Attach
Request or Tracking Area Update Requests. These messages are big in size and are
sent inside the RRC Connection Setup Complete message as NAS. So, this reduces the
RRC Success Rate of RRC Mo-Signalling compared to other RRC Request types.
 This means that if the network has a higher ratio of RRC Mo-Signalling requests then it
will have a lower RRC Success Rate. Usually, Mo-Signalling is around 20 to 25% while
Mo-data has the highest percentage. Still it can vary from network to network based on
TAC planning and mobility strategy. However, if you have very high Mo-Signalling
percentage then the chances are that RRC Success Rate will be relatively lower
compared to another similar network with lower Mo-Signalling percentage.

Incompatible UEs

It has been seen that sometimes there are users that are not compatible with the
configuration of the network. So, once they receive the RRC Connection Setup message
and they find out that they are not compatible with the configuration provided, they do not
respond with a RRC Connection Setup Complete message resulting in a RRC failure on
the eNB. However, such users keep trying again and again impacting the KPI. This kind
of issue can be seen from the traces or CHRs that verifies that it is a single user. It might
be inferred from the RRC counters as well since the number of failures are relatively
same in consecutive intervals. However, such cases usually go unsolved as it is not a
network issue but an abnormal UE problem.

PUCCH Based RRC Rejections

RRC Rejections due to PUCCH congestion can be solved by simply increasing the
PUCCH Resource Blocks. Vendors have parameters for PUCCH allocations and
minimum PUCCH Resource Block allocation is 4 per subframe. This is because each slot
has PUCCH RB allocation on both ends of the band so that means that each slot will
have atleast 2 Resource Blocks for PUCCH – one at the top of the frequency band and
the other at the bottom of the band. Since, each subframe has two slots so that means
that the subframe will have atleast 4 PUCCH Resource Blocks.

When 4 PUCCH RBs are not enough, they can be expanded to a higher value using
parameter or in some implementations, an adaptive approach can be maintained where
the eNB changes the PUCCH RB count dynamically based on the load requirement. This
approach solves the issue completely.

User Count Based or Flow Control Based RRC Rejections

Different baseband boards and vendors have different limitations on active user count
and CAPS (Call Attempts Per Second). When such limitation is reached, incoming RRC
Connection Requests are rejected by the eNB based on flow control or resource issue. In
such cases, the following basic steps can be done

 Decrease the UE Inactivity Timer to a smaller value. This will initiate early release for
the users and load due to user count will be reduced. However, this can increase the
signalling load as idle users can try to come back to network more frequently which
can increase CPU usage of the eNB. So, only use this if the issue is related to user
limitation while CPU usage is fine.
 T302 should be increased to limit the RRC signalling load. When a UE gets a RRC
Reject from eNB, it has to wait for T302 seconds before sending another RRC
 Connection Request. So, increasing T302 will increase the interval between such RRC
Connection Requests and therefore, reduce the signalling load on the eNB.
 Mobility Load Balancing is another feature that can help in such a scenario by moving
users away from the congested carrier to another less utilized carrier.

If you have any questions or feedback regarding this article, simply drop a comment
below. I will respond accordingly and also intend to write more about KPI Optimization
soon.

TURBO Receiver

The uplink receive quality ofcelledge UEs under weak coverage is poor. Specifically, the uplink SINR is lo
w, and the BLER of the decoded data transmitted using an MCS of a low index (mainly index 0) persists t
o be higher than the target value. As a result, user experience is poor in terms of uplink throughput. Wit
h Turbo Receiver, the eNodeB reconstructs data signals after decoding and uses these signals as referen
ce signals. Then, the eNodeB performs channel estimation, equalization, and decoding for a second time
. This can improve the performance in uplink reception, PKIs such as throughput, and user experience.
ERAB SSR and Optimization

The second major KPI for LTE is the LTE ERAB Success Rate which is also part of the
accessibility. After the UE has completed the RRC Connection which has been explained
in my previous article, LTE KPI Optimization: RRC Success Rate , it needs to get a
Bearer assigned to it to initiate services. The bearer can be default (usually Data QCI9)
or dedicated (VoLTE QCI1). During initial access, the default bearer is added and that
constitutes the major portion of the total ERABs.

Firstly, lets understand the definition and points where the ERAB KPI is pegged. After the
UE sends the RRC Setup Complete message to the eNB, the eNB sends a S1 Initial UE
Message to the MME indicating the purpose of the UE (Attach, TAU, CSFB, Service
Request etc) and its credentials. Once the MME receives this message and it decides
that a bearer is required, it will send an Initial Context Setup Request to the eNB. This
message is considered as the ERAB Attempt as it contains the bearers to be added
along with their QCI values. The eNB receives this message and adds the DRB (Data
Radio Bearer) based on the bearer profile in Initial Context Setup Request. But before
the eNB can add bearers, it needs to activate the security for the connection. This is
done by the Security Mode Command which carries the ciphering and integrity protection
algorithms. After this the eNB sends a RRC Connection Reconfiguration message to the
UE which adds a DRB and it includes the configuration for the DRB like bearer identity,
PDCP & RLC configuration (AM/UM etc). SRB2 is also added at this point with this
message. The UE receives these messages and reconfigures the connection. Then the
UE responds with Security Mode complete and RRC Connection Reconfiguration
Complete messages. As the eNB receives these messages, it sends an Initial Context
Setup Response to MME and this message is considered as the ERAB Success.
 Common Failures In ERAB Setup Phase

Now let’s understand the common failures that usually cause a ERAB setup failure. Most
of the times, the ERAB setup failures can be divided into two broad categories

Ø  Radio Induced ERAB Setup Failures

Ø  MME Induced ERAB Setup Failures

Let’s have an in-depth look at both of them and find ways to tackle them

Radio Induced ERAB Setup Failures

 Radio Link Failure

Consider a UE that receives Security Mode Command but fails to maintain radio
connection afterwards. This can happen in following two scenarios:

1. N310 consecutive out-of-sync events and T310 expiry

N310 indicates an interval of 200 consecutive PDCCH decoding failures. Simply put, if

the UE fails to decode PDCCH for 200ms, it will be considered one N310. However, from
here onwards, it is a sliding window with 10ms granularity. So, if the N310 value is 2 then
it means that if the UE fails to decode PDCCH for 210 ms, it will have exceeded the
configured N310 threshold. Once, N310 has been exceeded, the UE starts timer T310
and if the UE is unable to retain the connection (still unable to decode PDCCH) before
T310 expires, the UE will initiate RRC ReEstablishment. Let’s understand with an
example. Consider N310 of 11 and T310 of 500ms, then the UE will initiate RRC
 Connection ReEstablishment after 800 ms (N310 = (200 + (10*10)) = 300ms + T310 =
500ms).

2. Maximum RLC retransmission count exceeded

Consider that the UE receives both the Security Mode Command and the RRC
Connection Reconfiguration message. Now, it has to transmit the Security Mode
Complete and RRC Connection Reconfiguration Complete message in Uplink. However,
if the eNB fails to decode these responses, it will send a NACK to the UE or the eNB may
not send anything if it completely fails to even receive these messages. The RLC layer in
the UE is configured to resend the message if the message is not acknowledged. So, the
RLC layer will keep resending until a valid acknowledgement is received. But the RLC
cannot resend the same message indefinitely and it has a upper limit of retransmissions.
Once that limit is reached, the RLC will not retransmit again and the UE will consider that
the radio link is compromised. This will trigger a RRC ReEstablishment Request.

However, in both these cases, the RRC ReEstablishment Request will be rejected by the
eNB since processing this request requires to have a valid UE context at the eNB. But
since the UE did not respond to Security Mode Command, so the eNB does not consider
the context to be active yet and rejects the RRC ReEstablishment Request. At the same
instance, the eNB will send Initial Context Setup Fail to MME indicating an ERAB Setup
Failure.

 Optimization

Such issues can be reduced by increasing the N310 & T310 value. For instance, if the
value of N310 is increased from 2 to 6 and T310 is increased from 500ms to 1000ms,
 then the UE will wait longer and there is more chance that N311 will be triggered. N311 is
the In-Sync value and so it is the opposite of N310. T310 stops if N311 is triggered. If
N311 is 1 then it means that UE needs 100ms of successful PDCCH decoding to stop
T310. So, there is a higher probability of triggering N311 if the value of N310 and T310 is
big.

Similarly, if the RLC retransmission count threshold is increased from 8 to 16, then the
RLC will retransmit 16 times instead of 8 times which will increase the probability that the
eNB might be able to decode the message and prevent RLF.

 No Response From UE

In this case, the UE receives the Security Mode Command and the RRC Connection
Reconfiguration messages in downlink but does not respond to these messages in
uplink. This can result in the Inactivity Timer expiry and the eNB will send a UE Context
Release Request to the MME during ERAB setup phase which will cause the ERAB
setup failure. Let’s see why this scenario happens in live networks. Once a UE receives a
downlink message which needs a response, it will need an uplink allocation to send a
response. In order to get an uplink allocation, the UE requests the eNB by using a
Scheduling Request Indicator or SRI. The UE sends a SRI based on the SRI
Configuration shared with it in the RRC Connection Setup Message. The SRI
Configuration tells the UE about the periodicity of the SRI and it determines the subframe
where the UE will send the SRI. So, the eNB will look for that UE’s SRI in that subframe
only and based on that, the eNB allocates an uplink resource to the UE by instructing the
UE on the PDCCH. Now, the vendors have moved to adaptive SRI intervals which can
result in a new SRI configuration in the RRC Connection Reconfiguration message.
There are UEs that do not support this change of SRI configuration and they keep using
the old SRI configuration. So, once they have received the Security Mode Command and
the RRC Connection Reconfiguration messages in downlink and they want to respond in
uplink, they will have to send a SRI first. The UE will be sending SRI according to the old
SRI Configuration shared in RRC Connection Setup message while the eNB will be
looking for the UE’s SRI in the subframe defined in SRI Configuration of RRC Connection
Reconfiguration message. This will result in a scenario where the eNB will consider that
there is no response from UE and once the inactivity timer is expired, the ERAB setup
will fail.

This can also happen if the UE is in poor coverage or if the PUCCH has high
interference. The UE will keep sending SRIs in the correct location on PUCCH but the
eNB might not be able to read them resulting in a similar scenario as explained above.

 Optimization

If such a scenario is observed consistently, it will be a good idea to shift from an adaptive
SRI period to a fixed SRI period. This will avoid reconfiguring the SRI periodicity and will
prevent this issue.

Also, using PUCCH enhancements like IRC on PUCCH can help reduce the probability of
such issues.
 RLC Mode Issue

This is rarely seen in networks when a UM mode (Unacknowledged Mode of RLC) QCI is
used for UEs which do not support UM mode. A common example is the QCI7 which is a
Non-GBR QCI defined for live streaming or voice services and it usually works in UM
Mode. But there are many UEs which do not support UM mode and the eNB simply fails
to add a bearer with UM mode for them. This issue can be seen from the counters as it
will show that ERAB failures on Radio interface are happening only on QCI7 or any other
QCI which is set to UM Mode.

 Optimization

Simply changing the RLC mode for the QCI from UM to AM should solve this issue.

 Security Mode Failure

Another issue that is a bit rare is the Security Mode Failure issue. In this case, the UE
receives the Security Mode Command from the eNB but responds with a Security Mode
Failure message. Consequently, eNB sends Initial Context Setup Failure to the MME
resulting in ERAB setup failure. This happens if the security configuration on the eNB is
not supported by the UE or sometimes it can happen if the UE cannot handle both the
Security Mode Command and the RRC Connection Reconfiguration together. In most of
the cases, this turns out to be the terminal issue.

MME Induced ERAB Setup Failures

Let’s have a look at the MME induced ERAB failures. This may come as a surprise but
most of the MME induced ERAB setup failures in commercial networks are actually
caused by the radio interface and not the MME. I know it is hard to understand but those
of you who have delved themselves in RRC and S1 traces will understand it more clearly
once I explain this issue.

As explained in the section above, when the UE experiences a RLF after receiving the
Security Mode Command, it can try RRC ReEstablishment which actually tells the eNB
that there was a RLF on the UE’s side. Consider a UE experiencing a RLF before it
receives the Security Mode Command. The UE can only send a RRC ReEstablishment
after security is activated but if the UE experiences a RLF before the Security Mode
Command has been received, it cannot send a RRC ReEstablishment Request.

Now, consider that the UE experiences RLF after RRC Setup Complete message and
before Security Mode Command, this UE will go to idle and retry a new RRC connection
by sending another RRC Connection Request. Let’s say that the UE sends a RRC
Connection Request to another eNB (eNB2) and that eNB2 will start processing it. The
eNB2 does not know that the eNB1 already has a ERAB setup process going on for this
UE. The eNB2 will send a S1 Initial UE Message to MME for this UE and the MME will
see that it already has another ERAB setup process going on with eNB1. So, for MME to
initiate the new ERAB setup process by sending Initial Context Setup Request to eNB2, it
needs to first stop the process on eNB1, as it cannot have separate context of same UE
on two different eNBs. As a result, the MME will send a UE Context Release Command
 to eNB1 asking to abort the ERAB setup process. The eNB1 is trying to find the UE over
the air interface and once it receives the Context Release Command from MME, it will
consider that the MME aborted the ERAB setup and will peg it as a MME induced ERAB
setup failure. eNB1 will send an Initial Context Setup Failure to MME and the ERAB
setup on eNB1 will be pegged under MME induced failure. However, this issue was
actually caused due to radio issue but the eNB1 was not able to find that out.

This issue can also happen if the UE sends the second RRC Connection Request to the
same eNB or even to the same cell. At RRC level, the eNB does not check TMSI value
and the UE is referenced by its CRNTI. So, if the same UE sends another RRC
Connection Request to the same eNB, it will allocate a new CRNTI and will consider it a
new connection. But when the eNB will send S1 Initial UE Message to MME, the MME
will check the TMSI and will send UE Context Release Command to the previous session
resulting in ERAB setup failure on the first process.

Another scenario that can cause a MME induced ERAB Setup failure is the Initial Context
Setup Timer on the MME. If that timer is set to small value and eNB is waiting for the UE
to respond to Security Mode Command, the MME will send UE Context Release
Command due to timeout. This will also result in a MME induced ERAB Setup Failure.

 Optimization

There is no real optimization on the first scenario as it is purely a coverage issue and
coverage enhancement by physical or soft changes can be done to mitigate it. The
second scenario can be minimized by increasing the Initial Context Setup Timer on the
MME.

LTE Architecture.
The high-level network architecture of LTE is comprised of following three
main components:

 The User Equipment (UE).

 The Evolved UMTS Terrestrial Radio Access Network (E-UTRAN).

 The Evolved Packet Core (EPC).

The evolved packet core communicates with packet data networks in the
outside world such as the internet, private corporate networks or the IP
multimedia subsystem. The interfaces between the different parts of the
system are denoted Uu, S1 and SGi as shown below:

The User Equipment (UE)

The internal architecture of the user equipment for LTE is identical to the
one used by UMTS and GSM which is actually a Mobile Equipment (ME). The
mobile equipment comprised of the following important modules:

 Mobile Termination (MT) : This handles all the communication functions.

 Terminal Equipment (TE) : This terminates the data streams.

 Universal Integrated Circuit Card (UICC) : This is also known as the SIM
card for LTE equipments. It runs an application known as the Universal
Subscriber Identity Module (USIM).
A USIM stores user-specific data very similar to 3G SIM card. This keeps
information about the user's phone number, home network identity and
security keys etc.

The E-UTRAN (The access network)

The architecture of evolved UMTS Terrestrial Radio Access Network (E-
UTRAN) has been illustrated below.

The E-UTRAN handles the radio communications between the mobile and
the evolved packet core and just has one component, the evolved base
stations, called eNodeB or eNB. Each eNB is a base station that controls
the mobiles in one or more cells. The base station that is communicating
with a mobile is known as its serving eNB.

LTE Mobile communicates with just one base station and one cell at a time
and there are following two main functions supported by eNB:

 The eBN sends and receives radio transmissions to all the mobiles using the
analogue and digital signal processing functions of the LTE air interface.

 The eNB controls the low-level operation of all its mobiles, by sending them
signalling messages such as handover commands.
Each eBN connects with the EPC by means of the S1 interface and it can
also be connected to nearby base stations by the X2 interface, which is
mainly used for signalling and packet forwarding during handover.

A home eNB (HeNB) is a base station that has been purchased by a user to
provide femtocell coverage within the home. A home eNB belongs to a
closed subscriber group (CSG) and can only be accessed by mobiles with a
USIM that also belongs to the closed subscriber group.

The Evolved Packet Core (EPC) (The core

network)
The architecture of Evolved Packet Core (EPC) has been illustrated below.
There are few more components which have not been shown in the diagram
to keep it simple. These components are like the Earthquake and Tsunami
Warning System (ETWS), the Equipment Identity Register (EIR) and Policy
Control and Charging Rules Function (PCRF).

Below is a brief description of each of the components shown in the above

architecture:

 The Home Subscriber Server (HSS) component has been carried forward from
UMTS and GSM and is a central database that contains information about all the
network operator's subscribers.
  The Packet Data Network (PDN) Gateway (P-GW) communicates with the
outside world ie. packet data networks PDN, using SGi interface. Each packet
data network is identified by an access point name (APN). The PDN gateway has
the same role as the GPRS support node (GGSN) and the serving GPRS support
node (SGSN) with UMTS and GSM.

 The serving gateway (S-GW) acts as a router, and forwards data between the
base station and the PDN gateway.

 The mobility management entity (MME) controls the high-level operation of the
mobile by means of signalling messages and Home Subscriber Server (HSS).

 The Policy Control and Charging Rules Function (PCRF) is a component which is
not shown in the above diagram but it is responsible for policy control decision-
making, as well as for controlling the flow-based charging functionalities in the
Policy Control Enforcement Function (PCEF), which resides in the P-GW.

The interface between the serving and PDN gateways is known as S5/S8.
This has two slightly different implementations, namely S5 if the two
devices are in the same network, and S8 if they are in different networks.

Functional split between the E-UTRAN and the

EPC
Following diagram shows the functional split between the E-UTRAN and the
EPC for an LTE network: