[ndnSIM] Confusion about consumer settings

Mon Apr 29 20:25:41 PDT 2019

Hi Klaus, 

Thanks for your advice! 

I checked this way, and run the simulation by NS_GLOBAL_VALUE="RngRun=3" ./waf --run=ndn-line, the number of packets did change a little bit for scenario 1:
   A1Packet: 3203
   A2Packet: 1320
   A3Packet: 1235
   totalPacket: 5758

but the node5 still received the most, and node 7 received the least, that's strange.

Best regards,
Chris Chen

> -----Original Messages-----
> From: "Klaus Schneider" <klaus at cs.arizona.edu>
> Sent Time: 2019-04-30 05:31:06 (Tuesday)
> To: yao930 at mail.ustc.edu.cn, ndnsim <ndnsim at lists.cs.ucla.edu>
> Cc: 
> Subject: Re: [ndnSIM] Confusion about consumer settings
> 
> Hi again,
> 
> Yeah it just looks like randomness to me.
> 
> On 4/29/19 2:00 AM, yao930 at mail.ustc.edu.cn wrote:
> > Hello Klaus，
> > 
> > Thanks for your reply.
> > 
> > The link delay is the same on all links, and consumers starts at the same time. I also disable cache on every node.
> > 
> > I ran 3 other scenarios and the result is:
> > 
> > 1. Install consumer on node5, then node6, then node7, the rate-trace count is as follows. I repeated that 3 times, and the result is the same.
> 
> The result will always be the same, since ns-3 uses the same seed by 
> default.
> 
> You can change it via using different seeds or runs:
> 
> > RngSeedManager::SetSeed (3);  // Changes seed from default of 1 to 3
> > RngSeedManager::SetRun (7);   // Changes run number from default of 1 to 7
> 
> https://www.nsnam.org/docs/manual/html/random-variables.html
> 
> I'd expect for different seeds you'll get different results.
> 
> Best regards,
> Klaus
> 
> 
> >     A1Packet: 3253
> >     A2Packet: 1280
> >     A3Packet: 1212
> >     totalPacket: 5745
> > 
> > 2. Install consumer on node7, then node6, then node5, the rate-trace count is as follows, which seems exactly a reverse of scenario 1.
> >     A1Packet: 1212
> >     A2Packet: 1280
> >     A3Packet: 3253
> >     totalPacket: 5745
> > 
> > 3. Increase the queue size to be 600 packets, and the rate-trace count is as follows, the total packet is much less than above scenarios.
> >     A1Packet: 325
> >     A2Packet: 225
> >     A3Packet: 198
> >     totalPacket: 748
> > 
> > Topology files and simulation files are attached.  I put the topology file in ../ndnSIM/examples/topologies, and simulation file in ../ndnSIM/examples, then run ./waf --run=ndn-line to start the simulation.
> > 
> > Best regards,
> > Chris Chen
> > 
> > 
> > 
> >> -----Original Messages-----
> >> From: "Klaus Schneider" <klaus at cs.arizona.edu>
> >> Sent Time: 2019-04-29 09:42:06 (Monday)
> >> To: yao930 at mail.ustc.edu.cn, ndnsim <ndnsim at lists.cs.ucla.edu>
> >> Cc:
> >> Subject: Re: [ndnSIM] Confusion about consumer settings
> >>
> >> Hey Chris,
> >>
> >> Can you attach your full simulation file and your topology file?
> >>
> >> Did you set different link delays for your consumers? Do they all start
> >> at the same time?
> >>
> >> Maybe it's just randomness. Did you try changing the parameters slightly
> >> and see a hugely different result?
> >>
> >> Best regards,
> >> Klaus
> >>
> >>
> >>
> >> On 4/28/19 9:00 AM, yao930 at mail.ustc.edu.cn wrote:
> >>> Hello all，
> >>>
> >>> I am doing experiment with ndnSIM2.7 recently, and encountered some
> >>> strange problems.
> >>>
> >>> The topology I used has 8 nodes,whick looks like the follows:
> >>> 5
> >>>    \
> >>>     \
> >>> 6---4--3--2--1--0
> >>>     /
> >>>    /
> >>> 7
> >>> Node 5,6,7 is consumer, and node 0 is producer. Bandwidth in 1-2 and 2-3
> >>> is set to be 1Mbps, and others is set to be 5Mbps. Max packet size is
> >>> set to be 10 packets. Packet payload is set to be 1024.
> >>>
> >>>
> >>> I use 3 consumerHelper to make node 5,6,7 generate interests with
> >>> different prefix, like:
> >>>    NodeContainer consumerNodes1;
> >>>     NodeContainer consumerNodes2;
> >>>     NodeContainer consumerNodes3;
> >>>     NodeContainer producerNodes;
> >>>     NodeContainer routerNodes;
> >>>
> >>>
> >>>     consumerNodes1.Add(Names::Find<Node>("5"));
> >>>     consumerNodes2.Add(Names::Find<Node>("6"));
> >>>
> >>>     consumerNodes3.Add(Names::Find<Node>("7"));
> >>>
> >>>     std::string prefix1 = "/A1";
> >>>     std::string prefix2 = "/A2";
> >>>     std::string prefix3 = "/A3";
> >>>
> >>>     // Install NDN applications
> >>>     ndn::AppHelper consumerHelper1("ns3::ndn::ConsumerZipfMandelbrot");
> >>>     consumerHelper1.SetPrefix(prefix1);
> >>>     consumerHelper1.SetAttribute("Frequency", StringValue("100")); // 100
> >>> interests a second
> >>>     consumerHelper1.SetAttribute("NumberOfContents", StringValue("5000"));
> >>>     consumerHelper1.Install(consumerNodes1);
> >>>
> >>>     ndn::AppHelper consumerHelper2("ns3::ndn::ConsumerZipfMandelbrot");
> >>>     consumerHelper2.SetPrefix(prefix2);
> >>>     consumerHelper2.SetAttribute("Frequency", StringValue("100")); // 100
> >>> interests a second
> >>>     consumerHelper2.SetAttribute("NumberOfContents", StringValue("5000"));
> >>>     consumerHelper2.Install(consumerNodes2);
> >>>
> >>>     ndn::AppHelper consumerHelper3("ns3::ndn::ConsumerZipfMandelbrot");
> >>>     consumerHelper3.SetPrefix(prefix3);
> >>>     consumerHelper3.SetAttribute("Frequency", StringValue("100")); // 10
> >>> interests a second
> >>>     consumerHelper3.SetAttribute("NumberOfContents", StringValue("5000"));
> >>>     consumerHelper3.Install(consumerNodes3);
> >>>
> >>> what I expected is that node5,6,7 should receive data packets equally,
> >>> but the actual result is node5 receive most, and node7 receive the
> >>> least. I counted the rate-trace.txt, and the result is:
> >>>
> >>>      A1Packet: 4460
> >>>      A2Packet: 1386
> >>>      A3Packet: 382
> >>>      totalPacket: 6228
> >>>
> >>>
> >>> My questions are：
> >>> 1、why the data packets received by the 3 consumer is so different?
> >>> 2、I observed that in this scenario 1Mpbs bandwidth can transmit at most
> >>> 120 or so，how is this result calculated?
> >>>
> >>> Thanks in advance!
> >>>
> >>> Sincerely,
> >>> Chris Chen.
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> _______________________________________________
> >>> ndnSIM mailing list
> >>> ndnSIM at lists.cs.ucla.edu
> >>> http://www.lists.cs.ucla.edu/mailman/listinfo/ndnsim
> >>>