TCP throughput unfairness

[Mahesh]

Hi all,

So I am simulating a scenario where node 0 (10.1.1.5) in the topology is sending TCP traffic to nodes 1 to 4. When I observe the throughput stats, I find a huge unfairness in the throughput received by the nodes.

Flow 1 (10.1.1.5 -> 10.1.1.1)
simulationTime: 24
Tx Packets: 7331
Tx Bytes: 10971720
TxOffered: 3.65724 Mbps
Rx Packets: 7241
Rx Bytes: 10836720
Throughput: 3.61224 Mbps
******

Flow 2 (10.1.1.5 -> 10.1.1.2)
simulationTime: 24
Tx Packets: 42327
Tx Bytes: 63484712
TxOffered: 21.1616 Mbps
Rx Packets: 42327
Rx Bytes: 63484712
Throughput: 21.1616 Mbps
******

Flow 3 (10.1.1.5 -> 10.1.1.3)
simulationTime: 24
Tx Packets: 127
Tx Bytes: 181780
TxOffered: 0.0605933 Mbps
Rx Packets: 115
Rx Bytes: 163780
Throughput: 0.0545933 Mbps
******

Flow 4 (10.1.1.5 -> 10.1.1.4)
simulationTime: 24
Tx Packets: 34509
Tx Bytes: 51757712
TxOffered: 17.2526 Mbps
Rx Packets: 34509
Rx Bytes: 51757712
Throughput: 17.2526 Mbps
******

Please find attached the code and topology. Why is this happening ? Kindly reply.

Thanks,
Mahesh

[Mahesh]

I don't know why, but all my posts on ns3-users are getting deleted immediately after I post them. Also, I am not able to create a new post. Why so ?

Anyway, please find attached my code and topology. Looking forward to your kind reply.

Thanks,

Mahesh

--
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 a topic in the Google Groups "ns-3-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ns-3-users/sENvJEQu404/unsubscribe.
To unsubscribe from this group and all its topics, 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.

[Mahesh]

Hi all,

So I have recreated the same topology using yans-wifi-phy instead of spectrum-wifi-phy. There is no unfairness in yans-wifi-phy case, unlike with spectrum-wifi-phy.

With yans-wifi-phy, I am getting the following stats:

Flow 1 (10.1.1.1 -> 10.1.1.2)
  Simulation Time: 24
  Tx Packets: 23969
  Tx Bytes:   35947712
  TxOffered:  11.9826 Mbps
  Rx Packets: 23969
  Rx Bytes:   35947712
  Throughput: 11.9826 Mbps
******

Flow 2 (10.1.1.1 -> 10.1.1.3)
  Simulation Time: 24
  Tx Packets: 25531
  Tx Bytes:   38290712
  TxOffered:  12.7636 Mbps
  Rx Packets: 25531
  Rx Bytes:   38290712
  Throughput: 12.7636 Mbps
******

Flow 3 (10.1.1.1 -> 10.1.1.4)
  Simulation Time: 24
  Tx Packets: 15857
  Tx Bytes:   23779712
  TxOffered:  7.92657 Mbps
  Rx Packets: 15857
  Rx Bytes:   23779712
  Throughput: 7.92657 Mbps
******

Flow 4 (10.1.1.1 -> 10.1.1.5)
  Simulation Time: 24
  Tx Packets: 25485
  Tx Bytes:   38221712
  TxOffered:  12.7406 Mbps
  Rx Packets: 25485
  Rx Bytes:   38221712
  Throughput: 12.7406 Mbps
******

Why is this so ? Kindly clarify.

[Mahesh]

Why is SpectrumWifiPhy causing unfairness among the different flows ? Has it got to do with the multi-model spectrum channel ? Can someone please explain.

On 27 September 2017 at 11:57, Mahesh <mahes...@gmail.com> wrote:

Why is SpectrumWifiPhy causing unfairness among the different flows ? Has it got to do with the multi-model spectrum channel ? Can someone please explain.

[Mahesh]

Can someone please help me ? I am stuck on this. 

Thanks.

[Mohammad Abu Shattal]

Hi Mahesh,
One quick thing, did you tried to replicate the experiment multiple times and average the throughput for each node. It might be due to scheduling of the packets for different nodes. Just a thought.
Best,
Mohammad

To unsubscribe from this group and all its topics, send an email to ns-3-users+...@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.

--
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 a topic in the Google Groups "ns-3-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ns-3-users/sENvJEQu404/unsubscribe.

To unsubscribe from this group and all its topics, send an email to ns-3-users+...@googlegroups.com.

[Mahesh]

Hi Mohammad,

I did try that. But still, the throughput difference among the various flows is too much in the case of spectrum-wifi-phy. Yans-wifi-phy doesn't have such differences.

- Mahesh

[Mahesh]

Hi all,

I solved the issue. Apparently, a bug resolution was raised on ns-3 (https://www.nsnam.org/bugzilla/show_bug.cgi?id=2460) in which the "CCAMode1Threshold" had to be set to around -62 dBm when using spectrum-wifi-phy. The default values are different.

So if anyone is using spectrum-wifi-phy with ns-3.26, make sure that CCAMode1Threshold is -62 dBm.

//Bug 2460: CcaMode1Threshold default should be set to -62 dBm when using Spectrum
Config::SetDefault ("ns3::WifiPhy::CcaMode1Threshold", DoubleValue (-62.0));

- Mahesh