search

adjacentlink / emane

Distributed wireless network emulation framework
Other
127 stars 37 forks source link

RFPipe PCR curve issue #238

Open ijamc756 opened 1 year ago

ijamc756 commented 1 year ago

I used RF Pipe model and transmitted two radio links simultaneously to see how they interfere with each other. Unfortunately, they did not interfere with each other. The rx success rate was 100%.

For this test, all radios used similar parameters such as Bandwidth 8M, NoiseFigure 4dB, tx power 0 dBm using emane-tutorial/demo 5.

I also realised the PCR curve for RF Pipe is a straight line slope and based on packet size 0. It makes it difficult to estimate POR using POR0^(s1/s0) formula.

What changes should I make to successfully test the interference between two RF Pipe links and resolve the PCR issue?

patelkb commented 1 year ago

The RF Pipe model is a very simplistic model that does not account for interference due to in-band collisions from multiple transmitters. You can update the model to add that feature or develop a new model that supports the specific channel access effects and features you are looking to emulate. Feel free to review existing opensource models (TDMA, IEEE802.11abg, LTE) which account for in-band collisions with their own unique implementations.

The linear PCR curve provided is simply as a place holder and can be replaced with a curve which represents your waveform's behavior.

Also, the documentation for PCR states the following: "Specifying a packet size (

attribute pktsize) in the curve file will adjust the POR based on received packet size. Specifying a pktsize of 0 disregards received packet size when computing the POR."

So, if you want completion rates to change based on packet size, you have to set the packet size to a non-zero value and make sure the curve is representative of that non-zero packet size you set.

kb

Kaushik B. Patel Adjacent Link LLC

On 6/13/23 02:41, ijamc756 wrote:

I used RF Pipe model and transmitted two radio links simultaneously to see how they interfere with each other. Unfortunately, they did not interfere with each other. The rx success rate was 100%.

For this test, all radios used similar parameters such as Bandwidth 8M, NoiseFigure 4dB, tx power 0 dBm using emane-tutorial/demo 5.

I also realised the PCR curve for RF Pipe is a straight line slope and based on packet size 0. It makes it difficult to estimate POR using POR0^(s1/s0) formula.

What changes should I make to successfully test the interference between two RF Pipe links and resolve the PCR issue?

— Reply to this email directly, view it on GitHub https://github.com/adjacentlink/emane/issues/238, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABPLXMY4BMZ3CNYZWEOUNWDXLADRLANCNFSM6AAAAAAZELYX6A. You are receiving this because you are subscribed to this thread.Message ID: @.***>

dax0922 commented 1 year ago

Hi, I am new to EMANE, please correct me if I am wrong. You say, 'Specifying a pktsize of 0 disregards received packet size when computing the POR.' Why does tutorial 1 run with success rate of 100% with default pktsize 0 under /usr/share/emane/xml/models/mac/rfpipe/rfpipepcr.xml? or is the pktsize defined elsewhere? Grateful if you could clarify please?

patelkb commented 1 year ago

Disregarding packet size in this reference means the same PCR curve is used regardless of the actual packet size that is received. So, when "pktsize" is set to 0 in the curve file, you are basically telling the receiving node to compute the PCR based on the received SINR only.

Again, if you look at the documentation, it states the following: "The POR is obtained using the following calculation when a non-zero pktsize is specified: POR = POR0^(S1/S0)"

So what that means is, if you specify a packet size of 0 in your curve file, your POR will always be POR0 obtained from your curve file for the received/computed SINR. Whether you send a 100 byte packet or a 1000 byte packet, the POR will be the same if the SINR is the same.

It might be worthwhile for you to try the following simple tests using 2 nodes: Test-1:

  1. Use the default curve file with "pktsize" set to 0.
  2. Adjust the pathloss between two nodes so the SINR = 10.
  3. Send a stream of 100 byte packets and see what your completion rate looks like.
  4. Send a stream of 1000 byte packets and see if the completion rate changes. Hint: It should not.

Test-2:

  1. Change the "pktsize" parameter in your curve file to 128 Bytes.
  2. Keep the pathloss from above so the SINR = 10.
  3. Send a stream of 100 byte packets and see what your completion rate looks like.
  4. Send a stream of 1000 byte packets and see if the completion rate changes. Hint: It should based on the above POR formula.

kb

Kaushik B. Patel Adjacent Link LLC CEO, Director of Engineering 908.655.5079 @.***

On 6/13/23 16:43, dax0922 wrote:

Hi, I am new to EMANE, please correct me if I am wrong. You say, 'Specifying a pktsize of 0 disregards received packet size when computing the POR.' Why does tutorial 1 run with success rate of 100% with default pktsize 0 under /usr/share/emane/xml/models/mac/rfpipe/rfpipepcr.xml? or is the pktsize defined elsewhere? Grateful if you could clarify please?

— Reply to this email directly, view it on GitHub https://github.com/adjacentlink/emane/issues/238#issuecomment-1589995008, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABPLXM2XCQ3ZOODST6E2JA3XLDGFNANCNFSM6AAAAAAZELYX6A. You are receiving this because you commented.Message ID: @.***>

dax0922 commented 1 year ago

Thank you very much for the quick response, appreciate it. I will give it a try.

dax0922 commented 1 year ago

I have another question please if I may- the PCR curve xml files for TDMA and 802.11 have exactly same numbers (POR and SINR), are they supposed to be same or is there a reason for similar numbers for SINR and POR values?

  • © Githubissues.
  • Githubissues is a development platform for aggregating issues.