Editing endtoenddelay parameter in INET

[mohit shah]

Hello..

I wanted to know how is it possible and which program file has code to edit the 'endtoenddelay' statistic in INET source package. I can see that the vector endToendDelay can be seen in all results that use TCPSessionApp program. But where is this endtoenddelay vector stored? Any help will be largely appreciated. This is very urgent so kindly reply if you know of this.

Mohit

[Rudolf Hornig]

TCPSessionApp.ned:

@statistic[endToEndDelay](title="end-to-end delay"; source="messageAge(rcvdPk)"; unit=s; record=histogram,vector; interpolationmode=none);

This is the statistic's defining line. Means: whenever a packet arrives (scvdPk signal emitted), the age of the received message should be written out.

Rudolf

[mohit shah]

Hi Rudolf,

Thanks for letting me know of this. Kindly answer this one question of mine related to the same parameter:

When we say endToendDelay of the application/message, does it measure the difference between the time the message was created at the application layer of the source minus the time the message reaches the application layer of the destination i.e. endToendDelay is sum of [processing delay at source + network transmission delay]

OR

it its the difference between time the message was sent from the physical layer of source and time it reaches the lowest layer of the destination i.e. endToendDelay is just the network transmission delay?

This is a fundamental doubt I had kindly correct me with either/both of above assumptions. Your help is greatly appreciated you've always answered all my questions till date. :) Awaiting your reply.

Thank you,

Mohit

--
--
Sent from the OMNeT++ mailing list. To configure your membership,
visit http://groups.google.com/group/omnetpp
 
---
You received this message because you are subscribed to the Google Groups "omnetpp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to omnetpp+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
Best Regards,

Mohit Shah

Graduate Research Assistant
Dept. of Electrical & Computer Engineering
North Carolina State University
Profile: https://sites.google.com/site/mohitshah0316/

[Rudolf Hornig]

On Friday, 15 February 2013 14:16:18 UTC+1, mohit shah wrote:

Hi Rudolf,

Thanks for letting me know of this. Kindly answer this one question of mine related to the same parameter:

When we say endToendDelay of the application/message, does it measure the difference between the time the message was created at the application layer of the source minus the time the message reaches the application layer of the destination i.e. endToendDelay is sum of [processing delay at source + network transmission delay]

Yes except, the other way: currentTime - messageCreationTime (messageAge(rcvdPk)  is actually doing this calculation. See src/base/ResultFilters.cc). It is calculated on the received packet emitted. Obviously this is an application level packet if we are speaking about the endToEdnDelay of the application module.

OR

it its the difference between time the message was sent from the physical layer of source and time it reaches the lowest layer of the destination i.e. endToendDelay is just the network transmission delay?

No, the creation time of the physical layer message (i.e. when the message was put out to the physical medium) is available only inside the physical layer message. Once you decapsulate the content of the message (i.e. pass the content up to the network layer) that information is lost. When the packet is delivered to the TcpApplication it is excatly the same packet that was originally created in the source at the application level.

 

Be careful with the messageAge() function in statistics, though: some applications (notably the ICMP echo service) tend to send back the same message that was received. This means that the endToEndDelay measured in a ping application is actually the WHOLE roundtrip time.

[mohit shah]

Hi Rudolf,

Thanks for clarifying this. So to summarize based on my understanding:

endToendDelay statistic will summarize the time difference between packet transmitted at application layer of the source and received at the application layer of the destination. Also, the endtoendDelay statistic in the case of a TCPSessionApp in sync with a TCPSinkApp will actually be the one way delay and not RTT. Is my understanding correct? 

Thanks so much.

mohit

[Rudolf Hornig]

Yes.

[meriem...@gmail.com]

hi, I am a student of networks I need your help, I beg you, I want to collect statistics (End-to-End Delay,Packet Loss Ratio,Hop-to-Hop Delay,Per Flow Throughput,Convergence Latency,Scalability) on the ad hoc networks Implemented in inetmanet framework and displayed as graphs.Please help me as soon as possible.and thank you in advance.

[Bizzat Zaidi]

I want to know how is it read .. like what are the parameters on X and Y axis in case of histogram ?

I guess it is the msg count on Y and delay on X ... am i right 

? 

[Karthick Subramanian]

Hi Rudolf,

I am trying to get the endToEndDelay of a data sent from the application layer of TCPSessionApp and recieved at TCPSinkApp. When I look at the endToEndDelay it always zero. Dont know the reason can you please help me.


I have tried with insertion of min,max and mean, all these value appears to be zero.

@statistic[endToEndDelay](title="end-to-end delay"; source="messageAge(rcvdPk)"; unit=s; record=min,max,meanhistogram,vector; interpolationmode=none);

Thanks,
Karthick

[Michael Kirsche]

Did you check that your network connection actually has any delay (propagation delay whatsoever) and that the delay is diverse for different transmissions?

[Karthick Subramanian]

HI Michael,

I have delays in the system. I feel the problem is that in TCP layer we are deleting the segment from client app and create a new cMessage which whose age is calculated at the server application which is actually zero. Can you please confirm this.

Thanks,
Karthick