3 LTE Traffic Fault Diagnosis.ppt

www.huawei.com

Copyright © 2013 Huawei Technologies Co., Ltd. All rights reserved.

LTE Traffic Fault Diagnosis


Objectives

Upon completion of this course, you will be able to: Understand the E2E data transmission upon LTE Describes a roadmap for locating traffic faults Describes the methods of identifying downlink or

uplink traffic faults, basic location and isolation methods, and common fault location process for TCP and air interface problem

Page3

Copyright © 2013 Huawei Technologies Co., Ltd. All rights reserved. Page 4

Contents1. Overview2. Methodology of Traffic Fault Diagnosis3. Air Interface Fault diagnosis4. UDP Fault Diagnosis5. TCP Fault Diagnosis


E2E Data Transmission in LTE


Protocol Stack of E2E Data Transmission

Page 6

Serving GW PDN GW

S5/S8a

GTP-U GTP-U

UDP/IP UDP/IP

L2

Relay

L2

L1 L1

PDCP

RLC

MAC

L1

IP

UDP/IP

L2

L1

GTP-U

IP

S1-U LTE-Uu

eNodeB

RLC UDP/IP

L2

PDCP GTP-U

Relay

MAC

L1 L1

UE


Overhead of Layers

Page 7

Wired transmission Wireless transmission

ETH 14 bytes PDCP 2 bytes

IP 20 bytes RLC 2 bytes

TCP 20 bytes MAC 3 bytes Note: The values of the headers listed in this table do not include the extension.

Data frame format on the PC

Data frame format of the MAC layer on the eNodeB


Key Factors for LTE Throughput

Page 8

[Key factors]: cell bandwidth, modulation & coding scheme (MCS), MIMO mode, and UE capabilities.

MCS Index Modulation Order TBS Index0 2 01 2 12 2 2

.. .. .... .. .... .. ..

19 6 1720 6 1821 6 1922 6 2023 6 2124 6 2225 6 2326 6 2427 6 2528 6 2629 2

reserved30 431 6

Cell bandwidth

Available RBs

MCS

Transport block size (TBS)

MIMO modeNumber of scheduling operations

UE capabilities

Throughput (Mbit/s)

BLER


Key Factors for LTE Throughput (cont_1)

Page 9

91 92 93 94 95 96 97 98 99 100

0 2536 2536 2600 2600 2664 2664 2728 2728 2728 2792

1 3368 3368 3368 3496 3496 3496 3496 3624 3624 3624

2 4136 4136 4136 4264 4264 4264 4392 4392 4392 4584

3 5352 5352 5352 5544 5544 5544 5736 5736 5736 5736

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .19 39232 39232 40576 40576 40576 40576 42368 42368 42368 43816

20 42368 42368 43816 43816 43816 45352 45352 45352 46888 46888

21 45352 46888 46888 46888 46888 48936 48936 48936 48936 51024

22 48936 48936 51024 51024 51024 51024 52752 52752 52752 55056

23 52752 52752 52752 55056 55056 55056 55056 57336 57336 57336

24 55056 57336 57336 57336 57336 59256 59256 59256 61664 61664

25 57336 59256 59256 59256 61664 61664 61664 61664 63776 63776

26 66592 68808 68808 68808 71112 71112 71112 73712 73712 75376

TBSI PRBN

3GPP TS 36.213: Mapping between the PRB and TBS index for the one-codeword case

MIMO mode: determines the transmission mode (one-codeword for TM2 and dual-codeword for TM3.


Key Factors for LTE Throughput (cont_2)

Page 10

3GPP TS 36.306: Physical parameters of different downlink UE capabilities

UE Category

Maximum number of DL-SCH transport block bits received within a TTI

Maximum number of bits of a DL-SCH transport block received within a TTI

Total number of soft channel bits

Maximum number of supported layers for spatial multiplexing in DL

Category 1 10296 10296 250368 1

Category 2 51024 51024 1237248 2

Category 3 102048 75376 1237248 2Category 4 150752 75376 1827072 2

Category 5 299552 149776 3667200 4

UE Category Maximum number of bits of an UL-SCH transport block transmitted within a TTI Support for 64QAM in UL

Category 1 5160 NoCategory 2 25456 No

Category 3 51024 No

Category 4 51024 No

Category 5 75376 Yes3GPP TS 36.306: Physical parameters of different uplink UE capabilities

UE capabilities: The single-UE peak rate is limited by the UE capabilities.


Theoretical Throughput of the Downlink MAC Layer (20 MHz Bandwidth)

Page 11

　 subframe0

Guard

interval

subframe1 subframe5　 slot0 slot1 slot2 slot3 slot10 slot11　 0 1 2 3 4 5 6 7 8 9 10 11 12 13 0 1 2 3 4 5 6 7 8 9 10 11 12 13 0 1 2 3 4 5 6 7 8 9 10 11 12 13

99　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　98　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　97　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　96　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　

.　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　

.　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　

.　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　52　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　51　　　　　　　　　　　　　　　　　　　　　　　　　　　　 .. 　　　　　　　　　　　　　　50　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　49　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　48　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　47　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　

.　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　

.　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　

.　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　3　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　2　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　1　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　

0　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　Legend

　 DL control channel PDCCH PHICH PCFCH 　 DL shared

channel 　 Broadcast channel 　 Secondary

sync. channel 　 Primary sync. channel


Theoretical Throughput of the Uplink MAC Layer (20 MHz Bandwidth)

Page 12

　 subframe0 (subframe5)

Guard interva

l

subframe1 　 subframe2　 slot0 slot1 slot2 slot3 　 slot12 slot13　 0 1 2 3 4 5 6 7 8 9 10 11 12 13 0 1 2 3 4 5 6 7 8 9 10 11 12 13 　 0 1 2 3 4 5 6 7 8 9 10 11 12 13

49　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　48　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　47　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　. 　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　. 　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　. 　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　11　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　10　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　

9　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　8　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　7　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　6　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　5　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　4　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　3　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　2　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　1　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　0　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　

　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　Legend

　　 PUSCH 　　 PRACH 6 RBs 　　 PUCCH 　　 SRS


Theoretical Peak Rates of Cell and UE under Different Cell Bandwidth

Page 13

Cell Theoretical Peak Throughput

Bandwidth Category Downlink Theoretical Peak Throughput Uplink Theoretical Peak Throughput

1.4M Cat3 8.784 3.24

3M Cat3 22.128 7.992

5M Cat3 36.672 13.536

10M Cat3 73.392 27.376

15M Cat3 110.112 40.576

20M Cat3 150.752 51.024

Single User Theoretical Peak ThroughputBandwidth Category Downlink Theoretical Peak Throughput Uplink Theoretical Peak Throughput

1.4M Cat3 8.784 3.243M Cat3 22.128 7.9925M Cat3 36.672 13.536

10M Cat3 73.392 27.37615M Cat3 102.048 40.57620M Cat3 102.048 51.024


Typical Symptoms of Traffic Fault

Page 15

TCP method

UDP method Up link Down link

Stable but lower than peak rate by more than 5%

Fluctuates around the peak rate

Suddenly falls to a low value

Fluctuations

Stable & sub-peak rate

The throughput is stable but lower than the baseline value.

The peak throughput is lower than the reference value.

The stationary throughput under the same path loss is lower than the reference value for more than 10%.

The throughput fluctuates sharply

Throughput fluctuation Long-time

throughput drop Intermittent

throughput drop When a UE is static,

the throughput fluctuates for more than 50%.

Low throughput The peak throughput

is low. The static

throughput is low.


Process of Traffic Fault Locating

Page 16


Air/Non-Air Interface Fault Location

Page 17

Discover abnormal throughput.

Can data be transmitted?

Diagnose the problem.

No

Yes

YesYes

Is UDP throughput

normal?

Is the UE PC throughput

normal?

Does the Uu interface

work?

NoDiagnose the

problem.

Diagnose the problem.

Is the problem solved?

Diagnose the TCP problem.

EndCollect logs & submit the problem.

Yes

Yes

No

No

YesNo

Diagnose fault based on the UDP method


Checking Alarms and Basic Parameters on the eNB

Page 18

When the throughput becomes faulty, export the alarm log file to check whether there are hardware faults, S1 alarms, or intermittent interruption.

Check basic parameters respectively on the eNB and the UE.

Check basic parameters on the eNB against the baseline parameters. If the basic parameters on the eNB are inconsistent with the baseline values, investigate the reasons.


Isolating Air-Interface or Non-Air-Interface Faults

1. Turn to simple method: UDP packet transmission If the throughput in UDP transmission is evidently greater than that in TCP

transmission, a TCP fault occurs. Check the reasons of the TCP fault.

If the throughput in UDP transmission almost equals to or is lower than that in TCP

transmission, an air-interface fault occurs. Check the reasons of the air-interface fault.

2. If UDP transmission fails, upload multiple files using multiple threads or

simultaneously. If the throughput in multi-thread-based upload is evidently greater than that in single-

thread-based TCP transmission, a TCP fault occurs. Check the reasons of the TCP

fault.

If the throughput in multi-thread-based upload is almost the same as or lower than that

in single-thread-based TCP transmission, an air-interface fault occurs. Check the

reasons of the air-interface fault.


Contents3. Fault diagnosis on Uplink and Downlink

I. Uplink Fault DiagnosisII. Downlink Fault Diagnosis


Air-Interface Fault Location and Troubleshooting

If UL transmission encounters a throughput fault, that can be identified as an air

interface problem. Identify the problem from the following four perspectives. Of

these, the problems of "insufficient UL scheduling" and "insufficient UL scheduling

RB" can be identified together in a direction from left to right.UL throughput fault

Insufficient UL scheduling

Insufficient UL scheduling RBs

Low UL MCS order

UL IBLER convergence fault

Insu

ffici

ent p

acke

ts

Insu

ffici

ent a

ggre

gate

m

axim

um b

it ra

te

(AM

BR

)

Inva

lid C

L po

wer

co

ntro

l

Traf

fic fr

om o

ther

U

Es

at th

e lo

cal

cell

Inte

rfere

nces

Unb

alan

ce b

etw

een

mai

n an

d di

vers

ity o

f a

UE Lim

ited

UE

ca

pabi

lity



The single-UE UL peak throughput is decided by the following four conditions:

1. The top MCS order is 28 for category 5 UE, is 24 for category 4 UE and category 3 UE in

most cases, and is 23 for category 4 UE and category 3 UE in some cases.

2. The UE is allocated the maximum number of RBs. Maximum number of RBs available to a

single UE = Total number of RBs – Number of PUCCH RBs. The calculation result must

meet rule 2-3-5. Rule 2-3-5 means that the number of RBs for a single UE must be a

multiple of 2, 3, or 5.

3. UL scheduling is sufficient and the UL grants approximate to 1000.

4. The IBLER is less than 1%.

Bandwidth 1.4 3 5 10 15 20

Number of RBs 6 15 25 50 75 100

PUCCH RB 2 2 2 6 8 8Number of PUCCH RBs 2 2 2 2 4 4



Small Number of RBs Small Number of UL Scheduling Times Identify a Low MCS Order Identify an Excessively High BER


Troubleshooting Methods for Insufficient UL Scheduling RBs/Insufficient UL Scheduling

Top 1: Insufficient Packets (Insufficient Data Sources)

Note that, if the number of UL Grants is smaller than 600, that is caused by

insufficient packets. However, there is a fault which can be hardly detected.

That is, when the number of UL grants is 1000, the number of RBs

scheduled at some transmission timing intervals (TTIs) drops substantially.

This fault occurs in a peak throughout test at a high bandwidth.

Solution: Perform packet transmission based on multiple threads or replace a laptop

with better performance..



Top 2: Insufficient AMBR Query the AMBR to check whether the throughput problem is caused by an

excessively low AMBR. The transmission rate for a new UE must be greater than

expected. You can consult the personnel responsible for the EPC to check the

subscriber information and use the trace function on the LMT to check the S1

interface. As shown in the following figure, the UL AMBR and DL AMBR are 20

kbit/s, which is too low to meet the LTE requirements. Observe the AMBR on the

UE.

Unit: bit/s

S1 tracing dataHuawei UE data

displayed on the OMT



Top 3: Invalid CL power control Fault DescriptionThe UE that is located at a medium or far point away from the cell center has low throughput. The number of RBs is much smaller than the theoretical value in CL power control mode. However, the throughput of a UE close to the cell center can reach the peak.

When the RSRP is –100 dBm, if the CL power control is enabled, the UE at the bandwidth of 20 MHz has more than 90 RBs, and the UE at the bandwidth of 10 MHz has about 40 RBs. In comparison, the UE in OL power control mode has a maximum of 10 RBs.



Top 4: Traffic from other UEs at the local cell Run the hidden MML

command: dsp

cellusercntstats:localcellid=4;

Monitor the number of UEs by using the User Statistics Monitoring on the M2000.


Troubleshooting Methods for a Low MCS Order

Top 1: Interference



Top 2: Unbalance between main and diversity of a UE

Main and diversity RSRP of a Huawei UE on

the Probe

Main and diversity RSRP of a

Samsung UE on the X-Cal



Top 3: UE capability restrictions

CAT 4 displayed on the OMT for Huawei

test UE CAT 3 displayed in the tracing information on the

Uu interface


RF Fault Location and Troubleshooting

The stationary RSRP in an outdoor test fluctuates for more than 6 dB. This problem is caused by deteriorating radio environment. You are advised to perform

the test in another area or adjust the UE location or antenna.

The RSRP in an outdoor test fluctuates slightly while the throughput

fluctuates sharply. Check whether the interference changes or other UEs in the cell are using

services.

Throughput performance under interference Check the interference to observe how many dBs are increased in

comparison to the back noise. This increase can be regarded as an increase

in path loss.


Contents3. Fault diagnosis on Uplink and Downlink

I. Uplink Fault DiagnosisII. Downlink Fault Diagnosis


Typical Air-Interface Fault Location

If a DL throughput fault occurs and can be identified to be an air-interface problem, investigate the reasons of the fault from the following five perspectives in a direction from left to right.

Page 34

DL throughput fault

Insufficient DL scheduling

Insufficient DL scheduling RBs

Downlink MIMO fault

DL IBLER convergence fault

Low DL MCS Order

UL Feedback Fault


Insufficient DL Scheduling (1) Insufficient DL scheduling can be detected by using the

monitoring tool on the UE or on the eNodeB. The measurement is performed every one second. The number of scheduling attempts can be calculated as follows: Number of scheduling attempts of a UE = 1000/Number of UEs in the cell. If the total number of DL grants in the monitoring period is less than 95% of the UE scheduling attempts, the DL scheduling is insufficient.

Page 35

Insufficient DL scheduling also occurs if the number of UE scheduling attempts fluctuates for 50%.


Insufficient DL Scheduling (2)

Page 36

Check the aggregate maximum bit rate (AMBR) and guaranteed bit rate (GBR). If the configured AMBR and GBR are greater than the data rate on the air interface, insufficient DL scheduling is not caused by insufficient data transmitted by the EPC

Message

Then, check whether the data received from the S1 interface is sufficient as follows:


Insufficient DL Scheduling (3)

If the problem persists, perform the following operations:

Page 37

No

Yes

Check whether the number of DL

scheduling on the eNodeB is sufficient.

In case of a UE PDCCH demodulation problem (the following solution applies to a single UE), set the CFI to 3 and increase the PDCCH TX power for workaround. collect and send TTI tracing data on the eNodeB and the UE to Huawei headquarters for analysis.

In case of a scheduling priority problem on the eNodeB, collect and send TTI data of the eNodeB and UE to Huawei headquarters for analysis.


Insufficient DL Scheduling RBs

Page 38

identification----Average number of UE scheduling RBs per second < Number of RBs in the cell at the bandwidth/Number of UEs in the cell.

(1) The figure on he left shows the query result on the Probe. Number of RBs = Total Code0RBCount/Code0Count(2) The preceding figure shows how to query the number of RBs on the Web LMT. Query the number of DL equivalent RBs under Trace Indicator(s).


Insufficient DL Scheduling RBs (Cont.)

Page 39

• Check the information about UE capability on the air interface, different UE support different number of RBs

Search for the RRC Connection Reconfiguration message containing MeasurementConfig, if the value of reportOnLeave is True, the ICIC A3 measurement is used.

• Check whether the frequency selection scheduling or DL inter-cell interference coordination (ICIC) is enabled.


DL MIMO Fault

Page 40

• To check if the MIMO mode is correctly applied or not


Diverged DL BLER Diverged BLER in a non-peak-throughput scenario can be identified based

on the following flow chart.

Page 41

No

Yes

YesNo

Is the problem solved when UL response packet transmission is stopped (for UDP

packet injection)?

Fix the PUCCH TX power and send collected data to Huawei headquarters (If the TX power cannot be fixed, report the data directly.)

Fix the PUSCH TX power and send collected data to Huawei headquarters (If the TX power cannot be fixed, report the data directly.)

Normal

NoYes

Yes

NoDoes the MCS fluctuate

for 15 orders?Send collected data to Huawei headquarters

Does the reported CQI fluctuate for more than 5

orders?

No

Yes

Diverged BLER

Is the MCS adjusted to order 0?

Normal

Does the SNR on the air interface fluctuate for more

than 5 dB?

Send collected data to Huawei headquarters (CQI measurement on the UE may encounter a fault.)


Low DL MCS Order A low DL MCS order is caused by the following three

reasons: The DL SINR after MIMO equalization is low, resulting in a low DL

throughput. The DL CRC becomes faulty, resulting in a low DL MCS order in

the adaptive modulation and coding (AMC) adjustment. In an outdoor test, if the DL MCS order is low, the problem may

be caused by inaccurate time and frequency tracing. UL feedback is faulty, resulting in abnormal AMC adjustment and

a low DL MCS order. The first two reasons cannot be separated. For

example, if the DL SNR after MIMO equalization is low, the DL CRC becomes faulty and therefore the DL MCS order is low.

Page 42


Low DL SNR and CRC Fault Reasons A low DL SNR and CRC fault may be caused by the following five

reasons. Identify the problems in a direction from left to right.

Page 43

Inap

prop

riate

ne

twor

king

Inap

prop

riate

ant

enna

co

nfigu

ratio

n

Inap

prop

riate

ne

ighb

orin

g ce

ll pl

an

Sam

e ID

bet

ween

the

lo

cal

cell

and

neig

hbor

ing

cell

Larg

e tra

nsm

issio

n ch

anne

l del

ay

Loca

l osc

illatio

n le

akag

eSp

ectru

m m

irror

ing

leak

age

Unba

lanc

ed re

ceiv

ing

chan

nels

Faul

ts in

inte

rnal

m

odul

es o

f the

RRU

High

RX

powe

r on

the UE

False

PDC

CH a

larm

an

d in

spec

tion

erro

r

PDCC

H an

d PD

SCH

not m

atch

ed

Low DL SNR and abnormal CRC

High

DL

corre

latio

n

Inte

rfere

nce

from

a

neig

hbor

ing

cell

Time/

frequ

ency

ad

just

men

t fa

ult

RRU

Sign

al

proc

essin

g fa

ult o

n th

e RR

U

PDCC

H pr

oces

sin

g fa

ult


Low DL SNR or CRC Fault Identification (1) High DL correlation

Page 44


Interference from a neighboring cell

Page 45

Low DL SNR or CRC Fault Identification (2)


Signal processing fault on the RRU

Page 46

(1) Unbalanced channels on the RRU make the UE demodulation capability deteriorate, resulting in a low MCS order. The observation method is as follows:

(2) The UE receiving power generally ranges from –50 dBm to –90 dBm. A receiving power greater than –50 dBm may cause a clipping and then a low DL SNR. A receiving power lower than – 90 dBm may lead to a low DL SNR, affecting the DL transmission performance. Observer the UE receiving power with the following methods:

(3) In cases of inconsistent TX delay on the RRU, spectrum leakage, local oscillation leakage, and faults in internal modules of the RRU, collect and send baseband data to Huawei headquarters for analysis.



A PDCCH processing fault may lead to a DL CRC fault.

The PDCCH processing fault can be caused by false inspection of DL PDCCH, inspection error, or non-matched PDCCH and PDSCH.

False PDCCH inspection may lead to a virtual DCI that does not exist. When PDSCH demodulation is performed based on signaling messages out of false inspection, CRC faults occur. In case of non-matched PDCCH and PDSCH, a PDCCH is allocated on the eNodeB but the corresponding PDSCH is not, or the PDCCH and PDSCH data is inconsistent, resulting in CRC faults.

Page 47



UL Feedback Fault The problem of DL traffic volume may persist as shown in the

following figure even if no problem is found in the number of DL scheduling attempts, number of scheduling RBs, MIMO mode, IBLER convergence, and MCS order.

Page 48

In this case, this problem can be considered as a feedback channel fault. That is, ACK/NACK messages are interpreted as NACK/ACK/DTX messages. For details about data collection of feedback channel problems, see the UL PUCCH and PUSCH fault location methods.


Typical UDP Fault Diagnosis – Insufficient Throughput at Server Egress

[Symptom]During a UDP packet injection test, the egress throughput is lower

than the specification.

[Diagnosis procedure]1. Check the server performance & operating system.2. Check whether the version & parameters of the packet injection

tool iperf are correct.


Typical UDP Fault Diagnosis– Insufficient Throughput at eNB IngressThe eNodeB ingress throughput can be queried by running the DSP ETHPORT or DSP IPPATH command, or the DRB Static Monitoring interface of the M2000 cell performance trace function.

DSP ETHPORT command output.

DRB Static Monitoring interface.


Typical UDP Fault Diagnosis – Too High Packet Loss Rate

Normally we can Start UDP Test Monitoring on the M2000 to check the packet loss rate

Packet loss rate 67%


Typical UDP Fault Diagnosis – Uu Interface ProblemCheck the air interface as chapter before


Typical UDP Fault Diagnosis - UE PC Problem[Symptom] Data packets can be injected to the server and transmitted normally. But the rate measured at the UE PC is low.

[Diagnosis procedure] A low rate measured at the UE PC can be caused by the UE or UE PC. 1. Replace the UE and check whether the problem is solved. 2. Check the UE PC.

A. Check the hardware configuration of the PC. B. Check the programs installed on the PC. You are advised to delete or close other programs except the test program, close the Windows firewall and the firewalls of other anti-virus software. C. Check the CPU usage. If the CPU usage exceeds 80%, the CPU is overloaded. Close programs and services that are unrelated to the test, or use a PC of better performance. Note: The PC must be connected to the mains power supply.


Checking TCP Parameters1. Operating system: If the PC runs in Windows XP, continue with this step. If the PC runs in Vista,

Windows 7, or other operating systems, skip this step.2. Checking and setting of TCP parametersa. Use the DrTCP tool to find corresponding network adapters under Adapter Settings on the

transmitting and receiving sides respectively. Then, set the parameters as shown in the following figure:

b. Use the TCP parameter setting.reg tool to modify parameters on the transmitting side and import the parameters.

c. After completing the preceding two operations, restart the PC to activate the modified configurations. If the parameters cannot be modified on the server, modify the configurations only on the PC connecting to the UE.

3. If the problem persists after parameter modification, go to the next step.


TTI tracing

Perform TTI trace Interpret the traced data Results analysis If the problem persists after recommended

operations are performed, collect and send related data to Huawei headquarters for further analysis.


Further TCP Fault Diagnosis

[Diagnosis flowchart]


TCP Fault Diagnosis – Possible Causes of Large RTT over the Uu InterfaceUEs generate additional overhead when processing encrypted data, affecting the RTT & throughput.The default values of BSR and SR periods are 5 ms and 20 ms, respectively, different value also affect the throughput.


TCP performance enhancer (TPE) is located on the PDCP layer of the eNB and is responsible for processing the TCP data received & transmitted by the eNB. TPE is used to preliminarily analyze the problems of packet loss, disorder, and window shrinkage.

TCP Fault Diagnosis – TPE Log Collection with M2000(1)

1. Start the M2000 and choose Monitor > Signaling Trace > Signaling Trace Management > Information Collection > IFTS Trace.

2. Select an NE.

To collect the TPE log on the M2000, perform the following operations:


TCP Fault Diagnosis –TPE Log Collection with M2000(2)3. In the Trace Layer box of the IFTS Trace dialog box, select L1, L1 DOWN, L1 UP, and L2 Performance. In the MAC Layer Trace Field box, fill in 33/49/132.

NoteCollect the TPE log in the following sequence:1. Start IFTS Trace and select the traced items according to the preceding description.2. Disconnect and then connect the UE to the network. For E398, you need to remove and then insert this UE again. 3. To avoid adding any TPE port, you are advised to set the packet injection port to 20. A monitoring port can be added by running the MOD TPEALGO command.


TCP Fault Diagnosis – TPE Log Collection with LMT1. Choose Trace > IFTS Trace to enter the IFTS Trace dialog box.2. In the Basic tab page, choose L1, L1 DOWN, L1 UP, and L2 Performance in the Trace Module box. 3. In the Other tab page, fill in 33/49/132 in the MAC Inner Data Trace Num box.

Thank youwww.huawei.com

3 LTE Traffic Fault Diagnosis.ppt

Documents

Transcript of 3 LTE Traffic Fault Diagnosis.ppt