Automatic_Iperf_Log_Result_Excel_visual_v2.pptx

IPerf3 Log Parser
How to analysis iperf log tput value
By ChenChih, 20250627

Content
• Case1: parse max throughput, filter [Sum] string
• Case3: parse bidirectional, filter [SUM][RX-C] and [SUM][TX-C]
• Case4: parse both case1 and case2 (more convenient)
• Case7: split large bidirectional logfile into small part by
• Debug: some debug setting
Link:https://github.com/chenchih/5G_automationLinux/tree/main/
Iperflog_parser

What will you gain or learn
• Read and write file using python
• Regular expression to capture or filter log
• Convert txt file into excel, create sheet in excel
• Plot graph base on excel file
• Using progress bar to show running status
• File management, move result into new folder

Understand the log file
• Case1 (Max Throughput) filter [SUM]
• Case4: Bidirectional filter
• [SUM][TX –C]
• [SUM][RX –C]
Filter Tput value
Filter date and time
Will look like this after filter
Filter Tput value
RX
TX

Case1: Unidirectional
Run one directional either UL or DL
Source:
https://github.com/chenchih/5G_automationLinux/tree/main/Iperflog_
parser/1.parse_mbit_to_gb_excel

Description
• Need to provide Iperf log file to analysis, and extract the throughput
result
• Will generate excel file for extract the datetime, and tput value
• Will not draw graph, user have to manual draw graph

Step of code
• Step1:Put your elog file example
• Step2: Enter your logname (ex: DL_dailyTpt_SFP.log)
• Step3: will start to read and analysis your log and save result into
excel file.

Output
• Run command: py parse_mbit_to_gb_excel_v3.1.py
Enter your log filename
Enter for default filename, or enter save file name

Iperf Log Structure
• Iperf command
• iperf3 -c 192.168.1.150 -i 1 -P 48 -t 600 --timestamp --logfile UL_dailyTpt.log
• Filter str: [SUM]
• Log output: will display Gbit

Analysis Log file
• Filter these string:
• Date time: datetime
• Filter String:
• [SUM]: UL or DL information
• Throughput Value: Get the throughput value

Filter and matching string
• Match [SUM]
• re.match(r"(w{3} w{3}s+d{1,2} d{2}:d{2}:d{2} d{4}) [SUM].*?([d.]+)
(bits|Mbits|Gbits)/sec",line)
• Get the result:
if match:
date_str = match.group(1)
transfer_str = match.group(2)
unit = match.group(3)
Gbit/mbit/bit
Single date

Convert the Unit type to gbit
• Solution is to covert to same unit, since gbit contain the most of it. If
match mbit or gbit then divide by 1000 which will convert to gbit unit
type.
• Code:
if unit == "Mbits":
transfer_value /= 1000.0
elif unit == "bits":

Case3:Bidirectional
Run Uplink and downlink, and get the result of both value
Source code:
parser/3.parse_bidirectional

Step of code
• Step1:Put your elog file in the same working directory
• Step2: Enter your logname (ex: iperf3_bidirectional.log)
• Step3: will start to read and analysis your log and save result into
excel file.
• Step4: create folder and move your result into folder
• Step5: calculate the datetime of your logfile for test duration

Output:
capture the iperf log tput and save excel
• bidirectional_v3.py
Type your iperf
bidirectional log

Iperf Command with iperf log file
• Iperf client command:
• iperf3 -c 192.168.1.150 -P 32 -t 2400 --bidir -b 36M --timestamp --logfile
iperf3_bidirectional.log
• Filter str: [SUM]
• Log
--timestamp --logfile: will generate log file
--bidir: bidirectional

Analysis Log file
• Filter these string:
• Date time: datetime
• Filter String:
• [SUM][TX-C]: UL information
• [SUM][RX-C]: DL information
• Throughput Value: Get the throughput value

Filter and matching string
• Match [SUM][RX-C] or [SUM][TX-C]
• re.match(r"(w{3} w{3}s+d{1,2} d{2}:d{2}:d{2} d{4}) [SUM][(RX-C|TX-C)].*?([d.]
+) (bits|Mbits|Gbits)/sec", line_str)
• Get the result:
if match:
date_str = match.group(1)
rx_tx = match.group(2)
transfer_str = match.group(3)
unit = match.group(4)

• If you notice iperf log, some throughput will occur gbit, mbit, or even
bit. If you plot graph, the unit is inconsistency will make the graph
look incorrect.

• Solution is to covert to same unit, since gbit contain the most of it. If match
mbit or gbit then divide by 1000 which will convert to gbit unit type.
• Code:
if unit == "Mbits":
elif unit == "bits":
if rx_tx == "RX-C":
rx_data.append([formatted_date, transfer_value, "gbps"])
else:
tx_data.append([formatted_date, transfer_value, "gbps"])

Extend Feature: plot data from excel
# --- Create Plot ---
fig, ax = plt.subplots(figsize=(18, 8)) # Use subplots for more control, slightly wider figsize
# Plot data
ax.plot(times, tput, label=f'{direction_label} Throughput (gbps)', marker='', linestyle='-', markersize=5)
# Set Title and Labels with increased font size
ax.set_title(f'{direction_label} Throughput', fontsize=16)
ax.set_xlabel("Time", fontsize=14)
ax.set_ylabel("Throughput (gbps)", fontsize=14)
# --- Y-axis Limits ---
ax.set_ylim(bottom=0) # Set the lower limit of the y-axis to 0
ax.xaxis.set_major_locator(mdates.MinuteLocator(interval=30)) # Minute Intervals,
ax.xaxis.set_major_formatter(mdates.DateFormatter('%Y%m%d_%H:%M:%S')) # Multi-line format
# Increase tick label size and rotate
plt.setp(ax.get_xticklabels(), rotation=90, ha='right', fontsize=12)
plt.setp(ax.get_yticklabels(), fontsize=12)
# Adjust layout to prevent labels overlapping
fig.tight_layout()
plt.savefig(output_image_file, dpi=300) # Save the plot (dpi for resolution)

Case4: Parse Log working
directory
Parse iperf log both [SUM] or [SUM][TX-C] [RX-C]
Source code:
parser/4.parse_log_workingdir

Description
• This script will help you parse both bidirectional or unidirectional iperf
log file. Essential as long as you put .log or .txt it will help you filter
the tput value
• No need to type your logfile
• It will also help you plot graph
• This is all in one script help you do everything. Unlike previous two
example will analysis individual
• Note: Excel sheet will use your iperf log file name

Feature
• Allow to check all iperf log either max tput or bidirectional (verson1)
• Iperf log file name will be your excel sheet name
• Allow user to add y axis in case future you performance is more than
5 Gb you can adjust your graph axis (version2)
• Implement average of throughput( version3.1)
• Improvement average throughput- add threshold to compare. Iperf
log name will determine which test item you test and able to compare
average threshold and display pass fail criteria (version 3.2)

Step of code
• Step1: check current directory file with .txt or .log file and read it
• Step2: read each file and search for related string
• [SUM]:UL or DL Tput
• [SUM][RX-X] and [SUM][TX-X]: bidirectional Tput
• Step3: analysis the log
• Convert data unit
• save each file into excel sheet rename as log file name
• Save excel
• Calculate duration time (only with log contain [SUM][TX-X] or [RX-C]; bidirectional log)
• Step4: plot date into line graph
• Step5 create a folder and move all your result into folder

V3.2
display_average_using_input
• Feature:
• Implement user add threshold average to compare base on different
condition base on iperflog file. Log file name must contain:
• Bidirectional tput: log file name must contain these string
• SFP or POE (max TPT)
Iperf log Filename threshold
bidirectional-poe or bi-poe >= 1.1 PASS
bidirectional-sfp or bi-sfp >= 1.7 PASS
bidirectional >= 1.1 PASS
Iperf log Filename threshold
poe >= 2.5 PASS
sfp >= 3.5 PASS
IPERF log file will check the threshold

The throughput manual test value
• SFP max throughput (unidirectional) will reach about 3.5~4GB
• POE max throughput (unidirectional) will reach about 2.5Gb
• POE Bidirectional: will reach about 1.1 GB
• SFP Bidirectional: will reach about 1.7 GB
• In this script which allow to put max throughput log, and bidirectional log file. But if
you want to compare the pass fail criterial then you need to compare your test item
with the value. So in this case I use the log file as my test item to compare the average
• POE or SFP : these keyword will be Max TPT
• Bidirectional keyword:
• bidirectional or bi-poe or bidirectional-poe: either of keyword mean bidirectional poe
• bidirectional-sfp or bi-sfp: either of keyword mean bidirectional sfp

V3.2: output
Compare with log file name
Example:
• file contain bidirectional-poe or bidirectional or bi-poe:
average threshold will be 1.1. if less than it Fail
• file contain poe: average will be 2.5 less than it fail
• file contain sfp: average will be 3.5 less than it fail

V3.2: Code condition compare threshold

V3.1 average
implement average tput
• Implement average line to check the average performance, below the
red line is the average

V3.1
flexibility of y axis and x axis tick
• Allow user to adjust flexible y axis. This will be useful in future if your
performance can reach more than 4gb and can adjust gap of graph.
Y axis , support axis range
Enter to use the default value
User enter range of y axis control TPT axis. You
cam leave empty to use default 0 as min, 5 as
max

V3.1 flexible y axis
• Original
• Change: adjust user enter y axis for max and min

V2.2
• Allow to adjust x axis time tick. There is a condition to show the time
display
duration Time x axis
<=15minute Add every 30 second
<=2 hours Add every 5minute
<=6 hours Add every 10minute
<=1days Add every 30 minute
<=3 days Add every 3 hours
<=1days Add every 30 minute
Tick Every 10 minute
Iperf run 2hrs 30 min

V2.2
• I have set some setting with the x axis datetime with this condition:
• duration <=5 minute: set interval as 5 second
• duration <=15 minute: set interval as 30 second
• duration <=30 minute: set interval as 1 minute
• duration <=2 hours: set interval as 5 minute
• duration <=6 hours: set interval as 10 minute
• duration <=1 day: set interval as 30 minute
Duration 15 minute,
Interval as 15 second

V2.2
• Code for condition y axis tick

Calculate time duration only bidirectional log

Logfile and excel sheet output
• Run command: py parse_alllog_excelSheet_plot_v3.py
All result put in this directory
each sheet place each result Output result

Case7
Split_LargeFile_smallFile
Split large logfile to multiple file for reading
Source:
parser/7.Split_LargeFile_smallFile

Description
• This feature is only if you want to split a large file into smaller. For
example if you run bidirectional for over weekend about 3-4 days
then the log file will be about more than 2 gb. Text edior might not be
able to open( notepad++ only support 2~3gb) then you need to use
my script.
• I will split into N line, into smaller file, then you can open by text edior

Split line
Enter split line,
Default 4000000
Write line into txt file

Debug tool
Some tool for debug while using manual parsing

Covert date and time to specific format
• Script Name: convertdateFormat.py
• Datedata.txt: copy your datetime into this file like below
Datedata.txt
Ouput.txt

Read your log file and capture datetime
with open(input_file, 'r', encoding='utf-8', errors='ignore') as
infile:
for line in infile:
if "[SUM]" in line:
match = re.match(r"(w{3} w{3}s+d{1,2} d{2}:
d{2}:d{2} d{4}) [SUM][(RX-C|TX-C)].*?([d.]+) (bits|Mbits|
Gbits)/sec", line)
if match:
print(match.group(1))
else:
print(f"No match found in line: {line.strip()}")
Capture log datetime, and print date time

Calculate log’s running test duration
• timeduration_calculate.py

Additional Note
• Build python to executable
• window
• pyinstaller --onefile --icon=sheet.ico .parse_alllog_excelSheet_plot_v3.1.py
• Linux :
• pyinstaller --onefile parse_alllog_excelSheet_plot_v3.1.py
• Icon download(window):
• https://www.flaticon.com/free-icon

Automatic_Iperf_Log_Result_Excel_visual_v2.pptx

More Related Content

Recently uploaded (20)

Featured (20)

Automatic_Iperf_Log_Result_Excel_visual_v2.pptx

Editor's Notes