Thanks for your reply
I got three scripts from internet and I am going to share with you
1-parallel client 2-parallel server 3- serial client
But these script doesn't support me more than 200-300 parallel instances
http://sandilands.info/sgordon/multiple-iperf-instances(scripts is
available in given link as well)
I don't have much experience for writing a script
So tell me where I am lacking now and what should i do?
Right now I have two machines
Client machine
Microsoft Windows 8 Pro(Debian install on Virtual box)
Dell Inc. Precision WorkStation 390 x64-based PC Memory: 2GB
NIC:Broadcom NetXtreme 57xx Gigabit Cont
Server Machine:
GenuineIntel
Linux version 2.6.18-194.el5
LSB Version:
:core-3.1-amd64:core-3.1-ia32:core-3.1-noarch:graphics-3.1-amd64:graphics-3.1-ia32:graphics-3.1-noarch
Distributor ID: RedHatEnterpriseServer
Description: Red Hat Enterprise Linux Server release 5.5 (Tikanga)
Release: 5.5
Note: (I access my Server Machine remotely through putty software)
So I will appreciate for your response and early reply.
On Thu, Jul 4, 2013 at 3:35 PM, Engr Muhammad Tahir Fattani <
tahirfatt...@gmail.com> wrote:
> Thanks for your reply
>
> I got three scripts from internet and I am going yo share with you
>
>
>
> On Wed, Jul 3, 2013 at 11:51 PM, Jon Dugan <jdu...@es.net> wrote:
>
>> Iperf doesn't directly support this kind of workload but it should be
>> easy enough to write a script to fire off a bunch of iperfs to do this kind
>> of thing.
>>
>>
>> On Wed, Jul 3, 2013 at 11:10 AM, Ben Greear <gree...@candelatech.com>wrote:
>>
>>> On 07/02/2013 07:20 AM, Engr Muhammad Tahir Fattani wrote:
>>> > hey
>>> >
>>> > Yeah Sure.
>>> >
>>> > Big and huge scenario is possible to implement in LANforge-FIRE
>>> software?
>>>
>>> On higher end systems (Intel E3 or E5 CPUs, 16+GB RAM, 10G Ethernet,
>>> etc) we can
>>> do at least 50,000 concurrent TCP connections, and can emulate hundreds
>>> or thousands
>>> of different machines (servers and/or clients).
>>>
>>> > What is prerequisite of this software?(means which type of platform is
>>> need to implement these type of scenario).
>>>
>>> We suggest 64-bit Fedora Linux version 17. Probably newer versions
>>> would work, but our
>>> automated install scripts might have trouble with newer releases.
>>>
>>> For hardware, I suggest Intel NICs 1G, 10G, etc. If you need high-speed
>>> 10G throughput,
>>> then you need fast modern processors and 5GT/s pci-e busses and NICs.
>>> That said, lesser
>>> systems will work fine, they just might not give you quite as many
>>> concurrent connections.
>>>
>>> If you decide you want to try LANforge, please install Fedora 17 Linux
>>> on it and
>>> contact me off the list and I'll give you a trial license. Please note
>>> that
>>> LANforge is commercial software and is not open-source. We do offer
>>> free licenses
>>> to students using it on their own hardware and to non-profit charities.
>>>
>>> > Please let me know because this is my first time to implement this
>>> type of scenario and i don't know any thing how will I implement and get
>>> the result.
>>> >
>>> > So need your advice and suggestion in this regard.
>>> >
>>> >
>>> > Please let me know
>>> >
>>> > How will I know about the Iperf software does support these type of
>>> scenario or not??
>>>
>>> I don't think it does, but I imagine if someone *does* know how to make
>>> iperf do this
>>> they will post to the mailing list and tell you :)
>>>
>>> Thanks,
>>> Ben
>>>
>>>
>>> --
>>> Ben Greear <gree...@candelatech.com>
>>> Candela Technologies Inc http://www.candelatech.com
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> This SF.net email is sponsored by Windows:
>>>
>>> Build for Windows Store.
>>>
>>> http://p.sf.net/sfu/windows-dev2dev
>>> _______________________________________________
>>> Iperf-users mailing list
>>> Iperf-users@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/iperf-users
>>>
>>
>>
>
#!/bin/bash
# Run multiple parallel instances of iperf client
# Assumes iperf servers have been started, e.g.
# iperf -s -p PORT
# Examples:
# Run 5 clients for 60 seconds to server 1.1.1.1
# iperf-multiple-clients 1.1.1.1 5 60 report
# 5 files will be created, report-1.1.1.1-5001-60.txt, ...
#
# Run 7 clients for 20 seconds with UDP
# iperf-multipleclients 1.1.1.1 7 20 report-udp -u -b 10M
# Assumes the port numbers used by the servers start at 5001 and increase
# e.g. 5001, 5002, 5003, ...
# If you want something different, then change the following parameter value
# to be: firstport - 1
base_port=5000
# Command line input: server IP address
# E.g. 1.1.1.1
server_ip=$1
shift
# Command line input: number of clients to start
# E.g. 5
num_clients=$1
shift
# Command line input: test duration
# E.g. 60
test_duration=$1
shift
# Command line input: base report file name
# E.g. report
report_base=$1
shift
# Optional command line input: other iperf options
# E.g. -u -b 10M
iperf_options="$*"
# Run iperf multiple times
for i in `seq 1 $num_clients`; do
# Set server port
server_port=$(($base_port+$i));
# Report file includes server ip, server port and test duration
report_file=${report_base}-${server_ip}-${server_port}-${test_duration}.txt
# Run iperf
iperf -c $server_ip -p $server_port -t $test_duration $iperf_options &>
$report_file &
done#!/bin/bash
# Run multiple parallel instances of iperf servers
# Assumes the port numbers used by the servers start at 5001 and increase
# e.g. 5001, 5002, 5003, ...
# If you want something different, then change the following parameter value
# to be: firstport - 1
base_port=5000
# Command line input: number of servers
# E.g. 5
num_servers=$1
shift
# Command line input: base report file name
# E.g. report
report_base=$1
shift
# Optional command line input: other iperf options
# E.g. -u
iperf_options="$*"
# Run iperf multiple times
for i in `seq 1 $num_servers`; do
# Set server port
server_port=$(($base_port+$i));
# Report file includes server port
report_file=${report_base}-${server_port}.txt
# Run iperf
iperf -s -p $server_port $iperf_options &> $report_file &
done#!/bin/bash
# Run multiple iperf clients in serial, with random wait time
# between clients
# E.g.
# ---iperf---|--sleep--|---iperf---|--sleep--|---iperf---|--...
server_ip=$1; # IP address of the iperf server
shift
server_port=$1; # Port of iperf server
shift
test_time=$1; # duration of the iperf test
shift
num_tests=$1; # Number of iperf tests
shift
ave_sleep_time=$1; # Average time to sleep
shift
iperf_options="$*" # Any other iperf options
# Calculate the range for random value
# RANDOM returns value between 0 and 2^15 (about 32000)
# If want an average sleep time of 10 seconds, then
# the range should be 0 to 20. So divide the random
# number by (2^15)/20
divider=`echo 32767/2/$ave_sleep_time | bc`
# Show the start time
date
# Repeat the iperf tests
for i in `seq 1 ${num_tests}`; do
# Select a random number between 0 and 2^15
random_num=`echo $RANDOM`
# Scale down to 0 to 2*ave_sleep_time
random_sleep=`echo ${random_num}/${divider} | bc`
# Sleep (do nothing)
echo "Sleeping for ${random_sleep} seconds"
sleep ${random_sleep}
# Start an iperf test
echo "Running iperf for ${test_time} seconds"
iperf -c ${server_ip} -p ${server_port} -t ${test_time} ${iperf_options}
done
# Show the end time (you can use it to calculate the total time)
date------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:
Build for Windows Store.
http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Iperf-users mailing list
Iperf-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/iperf-users
