Help in LENA: Throughput, Delay, Packet Loss Ratio vs. Different Velocity

[Alan Zou]

Dear all,

Thanks for reading my message!
I was testing the performance of LTE in mobile scenarios. My test
scenario was very simple: One fixed eNodeB and one moving UE. I moved
the UE to 2000 m from the eNodeB. In the beginning, the velocity of
the UE is 5 m/s and then it will be changed to a higher speed at 1000
m. Finally, I computed the downlink throughput, dealy and packet loss
ratio by using the PDCP or RLC trace files(e.g. DlPdcpStats.txt and
DlRlcStats.txt).
However, the results seem that higher performance when travels at high
speed and the worst performance when at low speed. And you can see the
graph as below.
So I was wondering that was caused by what I didn't add the fading
model...
Since last week I have tried to understand why but...

Any help will be appreciated!

Regards,
Longhao

[Longhao Zou]

Please find my graphs attached to this reply. Thanks for your time!

regards,

Longhao

--
You received this message because you are subscribed to the Google Groups "ns-3-users" group.
To post to this group, send email to ns-3-...@googlegroups.com.
To unsubscribe from this group, send email to ns-3-users+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ns-3-users?hl=en.

[Anthony Faustine]

Help me could you give me your codes

On 4/9/12, Longhao Zou <zoulo...@gmail.com> wrote:
> Sorry, I forgot to you that I used the onOffApplication and the data rate
> is 20 Mbps.
>
> Thanks!
> Longhao

[Longhao Zou]

Sorry, I forgot to you that I used the onOffApplication and the data rate is 20 Mbps.

Thanks!

Longhao

On 9 April 2012 17:52, Longhao Zou <zoulo...@gmail.com> wrote:

[Longhao Zou]

Hi Anthony,

Which codes did you need? Could you help me to solve my problems?

regards,

Longhao

[Raja Sudha]

Dear Longhao

Could you please help me with the codes for the performance testing you had done with lte. It will be really helpful for my PG finals. Its urgent...

Sudhaa

Regards

RAJA. SUDHAA :-)

[Longhao Zou]

Hi everyone, any clue?

Longhao

On 9 April 2012 17:52, Longhao Zou <zoulo...@gmail.com> wrote:

[Anthony Faustine]

The Lte code plse

[Longhao Zou]

Dear all,

I am not sure if what I used "DlPdcpStats.txt/DlRlcStats.txt" to compute the throughput, delay and packet loss ratio on PDCP/RLC layer was really correct? If not, should I need to write my own way to calculate the throughput...Or because of the high mobility of UE, I should add the fading model into the scenarios?

Many thanks in advance!

Best regards,

Longhao

[Longhao Zou]

This is my simulation code...

Thanks!

Longhao

[Longhao Zou]

Hi Nicola,

Sorry to disturb you!

Would you give me any opinion for my question?

Thanks a lot!

Longhao

[Raja Sudha]

Dear longhao

Hi this is sudhaa. Thanks for the codes. It was successful. But I have a query. What does the resulting "RESULT.xml" file consist of? I am not able to understand the output in that format....

could you please help???

Sudhaa

--
You received this message because you are subscribed to the Google Groups "ns-3-users" group.
To post to this group, send email to ns-3-...@googlegroups.com.
To unsubscribe from this group, send email to ns-3-users+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ns-3-users?hl=en.

--
Regards

RAJA. SUDHAA :-)

[Longhao Zou]

Hello again,

That file is generated by flowmonitor...You can refer to my attachments in this email.

Regards,

Longhao

[Longhao Zou]

Hi,

Actually, I want to know the highest velocity of UE and the widest coverage of eNodeB supported by LENA ?!

Thanks a lot!!!

BR,

Longhao 

[psd prasad]

Hi i am Prasad

Few parameters you have to change in cc file
there is a parameter for the coverage and the coverage in the cc file that you include in header
check it and change according to ur needs then only the value be constant
in slow movement and also fast movement of the UE.

With Regards & Thanks

M. Prasad, 

Research Scholar,

Dept of Computer Science,

PEC.

[Nicola Baldo]

Hi Longhao,

sorry for the late reply. Please find some comments below.

> However, the results seem that higher performance when travels at
> high speed and the worst performance when at low speed. And you can
> see the graph as below. So I was wondering that was caused by what I
> didn't add the fading model...

if you don't use any fast fading model, the performance will be
determined by the distance only. Hence, the actual speed of the nodes
doesn't play any role. The difference in performance that you see
between the different simulations is probably due to some other cause.

if you use the fading model, please keep in mind that the traces are
pre-generated considering a given UE speed. Hence no matter the UE
speed
in the simulation, the fading trace will not change and will be the
same
for all UEs. Still, you can have different simulations, each one with
a
different fading trace that corresponds to a different UE speed.

It would be very nice to switch to different fading traces as a
function
of the speed of the UE in the simulations, but this is currently not
supported. Patches welcome....

> I am not sure if what I used "DlPdcpStats.txt/DlRlcStats.txt" to
> compute the throughput, delay and packet loss ratio on PDCP/RLC layer
> was really correct?

They are ok, provided that you are interested in radio link
performance.
For end-to-end performance, other tools such as flow monitor are
better
suited.

> Actually, I want to know the highest velocity of UE and the widest
> coverage of eNodeB supported by LENA ?!

the LENA model does not enforce any limitation on UE speed. As for the
coverage range in meters, it depends mostly on the propagation model
that you are using, as well as other settings such as the noise figure
and the AMC settings. For example, if the Piro2011 AMC scheme is used
(see
http://lena.cttc.es/manual/lte-design.html#adaptive-modulation-and-coding
for
details), the AMC behavior is this one:
http://lena.cttc.es/manual/lte-testing.html#fig-lte-mcs-index
which means that you go out of range (MCS index 0) when SINR is approx
<= -1 dB. Based on the knowledge of the TX power, noise figure and
propagation model being used, you can translate this into a coverage
range
in meters.

Regards,

Nicola

[Longhao Zou]

Dear Nicola,

I  am very grateful to you for your detailed response! Appreciate!

I am trying the fading traces generated by the "fading_trace_generator.m". So I will let you know once I have any results...

Many thanks again!

Kind regards,

Longhao

[Longhao Zou]

Hi Prasad,

Many thanks to your reply!

I will try to add the fading traces into the scenario. So thanks to your suggestion!

Regards,

Longhao

[Longhao Zou]

Dear Nicola,

Thanks for your suggestion.

I looked at the "fading_trace_generator.m" today. But I am wondering something about that:

1) Is "traceDuration" the same to the moving time of the UE?  (e.g. now in my understanding, the "traceDuration" is 400s if I move my UE for 400s at 5m/s speed) 

2) How long does it take to finish all the generation? I have waited for a long time but nothing outputs...no error...no generated file...

Thanks a lot!

Longhao

On 12 April 2012 10:11, Nicola Baldo <nba...@cttc.es> wrote:

[Luca Costantino]

Il 12 aprile 2012 23:46, Longhao Zou <zoulo...@gmail.com> ha scritto:
> Dear Nicola,
>
> Thanks for your suggestion.
> I looked at the "fading_trace_generator.m" today. But I am wondering
> something about that:
> 1) Is "traceDuration" the same to the moving time of the UE?  (e.g. now in
> my understanding, the "traceDuration" is 400s if I move my UE for 400s at
> 5m/s speed)

have a look here
https://groups.google.com/forum/?fromgroups#!searchin/ns-3-users/about$20trace/ns-3-users/xcgv7zrdsVM/hS7EDcshbn4J

> 2) How long does it take to finish all the generation? I have waited for a
> long time but nothing outputs...no error...no generated file...

it takes LOOOONG time

--
Chiave pubblica http://luca.costantino.googlepages.com/luca.costantino.asc

Prima di tutto vennero a prendere gli zingari e fui contento, perché
rubacchiavano.
Poi vennero a prendere gli ebrei e stetti zitto, perché mi stavano antipatici.
Poi vennero a prendere gli omosessuali, e fui sollevato, perché mi
erano fastidiosi.
Poi vennero a prendere i comunisti, e io non dissi niente, perché non
ero comunista.
Un giorno vennero a prendere me, e non c’era rimasto nessuno a protestare.
(Martin Niemöller)

[Longhao Zou]

Hi,

aha, I spent the whole night to run the matlab...

Thanks!

Longhao

[Luca Costantino]

Il 13 aprile 2012 14:15, Longhao Zou <zoulo...@gmail.com> ha scritto:
> Hi,
>
> aha, I spent the whole night to run the matlab...

generating a 10seconds trace file took about a couple of hours on my
2.3GHz 2nd generation Core-i5

[Longhao Zou]

Hi Luca,

Thanks to you for giving me the link!

But I am still not really understanding how long the length of trace(traceDuration) I should define if the time of my NS-3 simulation is 100s...

So I still think the traceDuration would be 100s if my NS-3 simulation time is 100s. Could you give me a clearer answer, please?

Many thanks!

PS. I just took about half an hour if the traceDuration=10.  And I am using a desktop but sorry I am not sure its configuration...

Appreciate in advance.

Longhao  

[Marco Miozzo]

Hi Longhao,

since the traces are pretty time and space consuming, you can choose to reduce the time length of the trace and apply a windowing mechanisms (the default mode adopted in lena).
You can find more info at the following links:
http://lena.cttc.es/manual/lte-design.html#channel-and-propagation
http://lena.cttc.es/manual/lte-user.html#fading-trace-usage

Regards,
Marco.

[Longhao Zou]

Hi Marco,

Thanks for your suggestion! I will try it...

Appreciate

Longhao

[Longhao Zou]

Hi all,

I have added the fading model into the ns-3 script and run it. But every time it will suddenly stop(just liked "hold on", nothing happened) after it has run for about 10 secs. I have waited for the long time...

Any one has a similar experience?

Thanks in advance!

Longhao

[Longhao Zou]

Hi all,

I encountered the same assertion as assert failed. cond "(verIhl >> 4) ==4", file=../src/internet/model/ipv4-header.cc, line=391 again when I ran the script with the original fading trace file named "fading_trace_EVA_60kmph.fad".

I have attach my script into this email to reproduce this problem. And my velocity was 60 kmph. Please try: ./waf --run "scratch/lteepc-moving-ue-3 --distance=2000"

Thanks a lot!

Longhao

[Marco Miozzo]

Hi,

the simulation script you provided does not works with the latest version of LENA.
Please update it.

Regards,
marco.

[Longhao Zou]

Hi Marco,

Thanks for your relpy!

I have updated the LENA just now. I also ran the script and the problem didn't happen again. 

Appreciate!

Regards,

Longhao   

[Raja Sudha]

Hi Longhao

Thanks for the posts regarding the fading codes in moving users LTE program. 

But when i run the codes, it ends with a single line "Simulation begins" and the corresponding flowmonitor ".xml" file is also not generated. Could u plz help me out ?

It will be really grateful

Thanking Yo

Regards

Sudhaa

Regards

RAJA. SUDHAA :-)

[Longhao Zou]

Hi Sudhaa,

Thanks!

1. Are you using the code named 'lteepc-moving-ue-3.cc' ?

2. Have you updated your LENA to the newest version?

3. Have you deleted the comments in front of 'SetFlowMonitor();' and 'monitor->SerializeToXmlFile("result1.xml",true,true);' ?

Regards,

Longhao

[Raja Sudha]

dear longhao

Thank you so much for the kind reply. But sorry for troubling you again and again. 

1) yes i am using the same codes.... but with fading which u had posted, i suppose!

2) how do i update? plz help

3) yes deleted.

How do i proceed?

Regards

Sudhaa

[Longhao Zou]

On 25 April 2012 12:50, Raja Sudha <sudh...@gmail.com> wrote:

dear longhao

Thank you so much for the kind reply. But sorry for troubling you again and again. 

1) yes i am using the same codes.... but with fading which u had posted, i suppose!

2) how do i update? plz help

Please open your command line in your LENA folder, then using the command as below:

hg pull

hg update

./waf build

[Raja Sudha]

dear longhao

Thank you so much for the reply. I will do the needful

Will update you 

Regards 

sudhaa

[Leticia]

Hi Nicola,

In a previous message you said "They are ok, provided that you are

interested in radio link
performance.
For end-to-end performance, other tools such as flow monitor are
better

suited. " talking about PDCP traces.

So, if I want to analyze the performance of the LTE network, for
example study the QoS for Voip (modeling it with OnOff Application),to
measure Packet delay, lost packet, jitter, etc I must use the
FlowMonitor tool and not the PDCP traces?

Sorry for my question but I am starting with the simulator and I have
a lot of things to learn, I dont know yet how to use the FlowMonitor
tool.

Thank you in advanced for your reply.

Leticia

> (seehttp://lena.cttc.es/manual/lte-design.html#adaptive-modulation-and-co...

[Nicola Baldo]

Hi Leticia,

On Tuesday, May 29, 2012 6:47:37 PM UTC+2, Leticia wrote:

In a previous message you said "They are ok, provided that you are
interested in radio link
performance.
For end-to-end performance, other tools such as flow monitor are
better
suited. " talking about PDCP traces.

So, if I want to analyze the performance of the LTE network, for
example study the QoS for Voip (modeling it with OnOff Application),to
measure Packet delay, lost packet, jitter, etc I must use the
FlowMonitor  tool and not the PDCP traces?

Yes, to determine VoIP QoS it is more appropriate to use FlowMonitor and not the PDCP traces.

Nicola

 

[Leticia]

THank you very much Nicola!

Best regards

Leticia

[Agos]

Hi Alan,
I have seen your files, are very intersting, but when I tried to run lteepc-moving-ue-3 but I had different problems, can you help me to solve them?

Multiple markers at this line
    - The type 'ns3::InternetStackHelper' must implement the inherited pure virtual
     method 'ns3::AsciiTraceHelperForIpv4::EnableAsciiIpv4Internal'
    - The type 'ns3::InternetStackHelper' must implement the inherited pure virtual
     method 'ns3::PcapHelperForIpv6::EnablePcapIpv6Internal'
    - The type 'ns3::InternetStackHelper' must implement the inherited pure virtual
     method 'ns3::AsciiTraceHelperForIpv6::EnableAsciiIpv6Internal'
 
And ... undefined reference to `ns3::FlowMonitorHelper::FlowMonitorHelper()'
    src/lte/examples/lteepc-moving-ue-3.cc.19.o: In function `LteEpcSim::SetFlowMonitor()':

Thanks

Regards

[Longhao Zou]

Hi Agos,

Did you run this simulation on LENA (http://iptechwiki.cttc.es/LTE-EPC_Network_Simulator_%28LENA%29#Source_Code_Repositories) ?

If yes, have you updated the LENA to the latest one? And it would be a little different from now on because that file was uploaded five months ago. I am afraid you have to modificate that code by according to the new "lena-simp-epc.cc" which locates in the "src/lte/examples".

Good luck!

Longhao

--
You received this message because you are subscribed to the Google Groups "ns-3-users" group.

To view this discussion on the web visit https://groups.google.com/d/msg/ns-3-users/-/4U8H_RmZLLMJ.

[Srinath Gnanassegarane]

Hi Longaho,

I could just execute your both programs of lteepc-moving-ue-3 and lteepc-moving-ues but the ambiguity that arises in me is when i examined the files of DLRlcstats.txt and DLPdcpstats.txt is that the reception gets totally deteriorated after about 16 or 17s in DLPdcpstats.txt(i.e., the number received PDUs tends to 0) and in DLRlcstats.txt the performance is less in the ranges of 20 to 40s. These are the cases when i executed the lteepc-moving-ue-3 at different distances of separation of ue with enb. I could also upload the results of them for further clarification.

My question is that if this deterioration in signal is acceptable or did i do something wrong?

It would be of great help if you could sort it out.

Thx in advance.

Regards,

Srinath Gnanassegarane

[Longhao Zou]

Hi Srinath,

Actually you can change to a longer distance if you want.

Cheers

To view this discussion on the web visit https://groups.google.com/d/msg/ns-3-users/-/KiFpDL1hIO4J.

[Lawrence Chen]

Dear Longhao,

Sorry, I am a beginner in NS3.
Could you tell me how to create those graphics ? Thanks !

Regards,
Lawrence


Longhao Zou於 2012年9月22日星期六UTC+8下午7時43分48秒寫道：

[Longhao Zou]

Hi Lawrence,

I think you could generate the figures by using the results shown in the DlRlcStats.txt and DlMacStats.txt.

Cheers

To view this discussion on the web visit https://groups.google.com/d/msg/ns-3-users/-/kr_O0kC8qnIJ.

[Na]

Hi Longhao,

Did you use gnuplot to generate those graphics or another tool?

Thanks!!

[Pankaj Goyal]

hey Longhao, how are you. are you the same person i met last month in AIT ireland.?

On Wednesday, September 19, 2012 4:48:00 PM UTC+1, Longhao Zou wrote:

Hi Agos,

Did you run this simulation on LENA (http://iptechwiki.cttc.es/LTE-EPC_Network_Simulator_%28LENA%29#Source_Code_Repositories ) ?

[dee...@qmic.com]

Hello sudhaa,

How do you able to run that file? 

It is showing error like... 

../scratch/lteepc-moving-ues.cc: In member function ‘void LteEpcSim::InstallENBUEs()’:

../scratch/lteepc-moving-ues.cc:271:12: error: ‘class ns3::LteHelper’ has no member named ‘ActivateEpsBearer’

Can you please reply to me. 

Thanks,

Deepak

On Wednesday, April 11, 2012 8:00:52 PM UTC+3, Raja Sudha wrote:

Dear longhao

Hi this is sudhaa. Thanks for the codes. It was successful. But I have a query. What does the resulting "RESULT.xml" file consist of? I am not able to understand the output in that format....

could you please help???

Sudhaa

On Mon, Apr 9, 2012 at 10:20 PM, Alan Zou <zoulo...@gmail.com> wrote:

Dear all,

Thanks for reading my message!
I was testing the performance of LTE in mobile scenarios. My test
scenario was very simple: One fixed eNodeB and one moving UE. I moved
the UE to 2000 m from the eNodeB. In the beginning, the velocity of
the UE is 5 m/s and then it will be changed to a higher speed at 1000
m. Finally, I computed the downlink throughput, dealy and packet loss
ratio by using the PDCP or RLC trace files(e.g. DlPdcpStats.txt and
DlRlcStats.txt).

However, the results seem that higher performance when travels at high
speed and the worst performance when at low speed. And you can see the
graph as below.
So I was wondering that was caused by what I didn't add the fading
model...

Since last week I have tried to understand why but...

Any help will be appreciated!

Regards,
Longhao

--
You received this message because you are subscribed to the Google Groups "ns-3-users" group.

To post to this group, send email to ns-3-...@googlegroups.com.
To unsubscribe from this group, send email to ns-3-users+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ns-3-users?hl=en.

[Danial Zakaria]

hi longhao, can I know how u plot the graph with DlPdcpStats.txt/DlRlcStats.

txt? are u using some kind of software or script to plot the graph?
I hope u can show me how.

Reagrds,
Danial

[Tommaso Pecorella]

Let's recap...

- you're replying to a thread where the last posts are Jan 2013, Sep. 2013, Dec. 2014. All unanswered.

- you're specifically quoting a post from November 2012 (3 years ago).

Fine, you're trying to contact him directly. Ok, no problem, but you should double check that you're not sending the mail to the group too, DOH !

Also, don't you think that this is something that the LTE maintainers could help you with better ?

Summarizing: open a new thread.

T.

On Saturday, November 28, 2015 at 12:15:02 PM UTC+1, ALESSANDRO FIGURA wrote:

hi Longhaou,

 am alessandro and i am a new ns3 user who is trying to understand how to get performance analysis of a LTE-epc scenario with a moving UE who communicates with a remote host and , at the moment, only one enb. 

i was reading the post you discuss just 3 years ago. i know it is a long time but i have a lots of difficulties in undestanding how to calculate the performace and at which level.

you asked if it is recommanded to use the DlRlcStats and PDCP stats to calculate Throughput, if you solved this trouble can you explain me??and how to calculate it???

I am a little bit confused how to calculate the throughput using DlRlcStats and DlPdcpStats.txt, can you help me?

at the moment i used a UDP client/server application using packetsink helper and using flow monitor i have calculated the end-to-end throughput refering to the UDP application and it seems to be ok, cause i used 1024 bytes packet size and 1 ms interpacketintervall, so i sholud have a 8 Mps throughput and i reach it. 

but what i would like to understand is how to use the RLC, PDCP, MAC , Phy traces...i read the tutorial and the LTE model documentation but i am so confused. 

what i would like to undestand is how to show that LTE can reach a datarate of 100 Mbs and how to do it, or better in which level and with traces have i calculate this value???because usiing flow monitor at the application level using UDP protocol i can't. 

I hope you can help me in order to understand better the question and continue with my simulation. 

i am sorry to contact you in private and i hope am not disturbing you. 

thanks in advance. 

best regards

alessandro

Il giorno lunedì 9 aprile 2012 22:32:37 UTC+2, Longhao Zou ha scritto:

Dear all,

I am not sure if what I used "DlPdcpStats.txt/DlRlcStats.txt" to compute the throughput, delay and packet loss ratio on PDCP/RLC layer was really correct? If not, should I need to write my own way to calculate the throughput...Or because of the high mobility of UE, I should add the fading model into the scenarios?

Many thanks in advance!

Best regards,

Longhao

On 9 April 2012 20:12, Anthony Faustine <samb...@gmail.com> wrote:

The Lte code plse

On 4/9/12, Longhao Zou <zoulo...@gmail.com> wrote:
> Hi Anthony,
>
> Which codes did you need? Could you help me to solve my problems?
>
> regards,
> Longhao
>
> On 9 April 2012 18:35, Anthony Faustine <samb...@gmail.com> wrote:
>
>> Help me could you give me your codes
>>
>> On 4/9/12, Longhao Zou <zoulo...@gmail.com> wrote:
>> > Sorry, I forgot to you that I used the onOffApplication and the data
>> > rate
>> > is 20 Mbps.
>> >
>> > Thanks!
>> > Longhao
>> >
>> > On 9 April 2012 17:52, Longhao Zou <zoulo...@gmail.com> wrote:
>> >
>> >> Please find my graphs attached to this reply. Thanks for your time!
>> >>
>> >> regards,
>> >> Longhao

>> >>
>> >>
>> >> On 9 April 2012 17:50, Alan Zou <zoulo...@gmail.com> wrote:
>> >>
>> >>> Dear all,
>> >>>
>> >>> Thanks for reading my message!
>> >>> I was testing the performance of LTE in mobile scenarios. My test
>> >>> scenario was very simple: One fixed eNodeB and one moving UE. I moved
>> >>> the UE to 2000 m from the eNodeB. In the beginning, the velocity of
>> >>> the UE is 5 m/s and then it will be changed to a higher speed at 1000
>> >>> m. Finally, I computed the downlink throughput, dealy and packet loss
>> >>> ratio by using the PDCP or RLC trace files(e.g. DlPdcpStats.txt and
>> >>> DlRlcStats.txt).
>> >>> However, the results seem that higher performance when travels at high
>> >>> speed and the worst performance when at low speed. And you can see the
>> >>> graph as below.
>> >>> So I was wondering that was caused by what I didn't add the fading
>> >>> model...
>> >>> Since last week I have tried to understand why but...
>> >>>
>> >>> Any help will be appreciated!
>> >>>
>> >>> Regards,
>> >>> Longhao
>> >>>

[Shamma Nikhat]

Hello,
 I am a new user of ns-3. can u please suggest me how to find the throughput from the trace files ( DlPdcpStats.txt and
DlRlcStats.txt)? I want to plot throughput vs time graph.

Thanks
Shamma

[Mojgan Chakeri]

Do you find answer for this question? if yes, can you help me?

[Biljana Bojovic]

Hi Mojgan,

please check https://www.nsnam.org/docs/models/html/lte-user.html#simulation-output documentation to see what are available simulation outputs, and then you can figure out if they can be used for the calculation that you need. If not you can look in ns-3 user documentation how to create your own traces.

Cheers,

Biljana

--
Posting to this group should follow these guidelines https://www.nsnam.org/wiki/Ns-3-users-guidelines-for-posting
---

You received this message because you are subscribed to the Google Groups "ns-3-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to ns-3-users+unsubscribe@googlegroups.com.

To post to this group, send email to ns-3-...@googlegroups.com.

Visit this group at https://groups.google.com/group/ns-3-users.
For more options, visit https://groups.google.com/d/optout.