Setup virtual nodes (containers) to simulate 2 mobiles and one base station.

[Deu82]

Hello @roymx ,

First of all, thanks for helping me out to fix the meepctl deploy issues last week. Everything is running fine and I can now load the Demo 1 example and play around with it. I would like to ask for your help once more to help me setup a test use case, where:

I would like to setup 3 virtual nodes (containers) to simulate 2 mobiles and one base station.

After the setup is done, I would then perform the simulations on this setup.

Since I'm a beginner with AdvanrEDGE, I don't know how to, or where can I go on the AdvantEDGE GUI to setup my test setup.

It would really be helpful if you could point me to the right direction.

Thanks in advance!

[roymx]

Hello @Deu82

Regarding the steps to get started,

• this post may be helpful in providing higher level overview of getting started - Study material on how to use & apply AdvantEDGE in use cases #216 (reply in thread)
• this documentation from the Wiki will help in scenario creation - https://github.com/InterDigitalInc/AdvantEDGE/wiki/first-scenario
• finally when creating the scenario, use this button
  for specific help filling the fields

Regarding the pre-requisites mentioned in the 1st link, it would help if you could provide details on the vertical you are trying to run (what is the application, is it running in containers, etc.)

Thanks,
Mike

[Deu82]

Hello @roymx ,

Thanks a lot for your detailed answer. I really appreciate it. I am currently trying the first approach,i.e., topological distance. However, while trying it, I found that on the 'Execute' tab, the Event and other options are disabled. Due to this I could not send a Mobility event.Although the round button is green (meaning all systems are running!).

A few weeks ago, when I was working with the demo1 app, these options were enabled. Could you please inform me what could I do to enable them.

Once these options are enables, I would then proceed with the first suggestions you made.

A note: Every time I start working with AdvantEDGE, I execute these commands:
meepctl deploy dep
meepctl dockerize all
meepctl deploy core

and then
cd ~/AdvantEDGE/examples/demo1/bin/iperf-proxy
./iperf-proxy

And, after the work is done, I delete the dep and the core

Thanks again for your help!

[roymx]

Hello @Deu82 ,

No problem, glad you are experimenting with the platform.

Regarding the Note, I'll refer you to the wiki cheat sheet

• when you update the AdvantEDGE repo, you need to re-build/dockerize - do 4-5-6-7
• when you start to work without updating/changing AdvantEDGE (e.g. you already built it), you just need to do 4-7

Regarding the green LED - it means that the AdvantEDGE containers are all up and running
Regarding the event buttons not enabling, it means that something in your scenario did not deploy correctly.

• If you are not using it, I recommend using this CLI tool K9s
• Just download one of the pre-built executable and execute on the CLI.
• it will help you identify which pods are running/failing and allows to quickly look at logs and open a shell.

My guess is that one of the POD in your scenario is not starting. I recommend doing kubectl describe pod <name-of-failing-pod> to identify what is wrong.

[Deu82]

Hello @roymx ,

The event tabs are now working. I created a new scenario by following the instruction here ). My previous scenario was not working properly for reasons I could not figure out. All the pods were running, still the event options was not enabled. Please see my further questions below:

Hello @Deu82 ,

There are two ways to represent "distance":
1- Topological distance (e.g. number of hops & network characteristics of hops between point A to B)
2- Geographical distance

For Topological distance, the screenshots you provided show the topological view of the network and therefore will show the topological distance (e.g. you can count the hops) - the network characteristics will be influenced by the number of hops and the net char settings of each hop.

Since I'd like to go for the first approach (Topological distance), this is my current setup. Here I've added another PoA (named bs2) on Zone1 and added another Zone with another PoA and edge server. I've a few followup questions, please check below the image.

I've three followup questions:

1. If I send a Mobility event to allow faruser (in the above image) to move to bs2 or bs3, would that mean that the hops are increased now?

2. you previously mentioned that I can set different network characteristics for the PoAs "To increase the topological distance you can add another PoA to your zone and set different network characteristics for the new PoA (number of hops will remain the same, network characteristics will differ)"

-Could you please inform me where do I change the network characteristics? And, how do I change the number of hops.

3. I'd like to execute these actions:

i) nearuser sends x1 bits to the bs1 with an error rate e1.
ii) faruser sends x2 bits to the bs1 with error rate e2.
I would then compare the response time.

Could you please inform me, how do I set the bits (x1 and x2) and error rate in the GUI? And, I suppose the response time is actually the latency, please correct me if that's wrong.

Thanks a lot again for your help and suggestions!

[roymx]

Hello @Deu82

Regarding 1

• Purely speaking about "number" of hops - number will not necessarily change (it depends on network topology)
• The hops will definitively be different.
• If connected to bs1 PoA, hops will be ue -> bs1 -> fog1 so that's 2 hops
• If connected to bs2 PoA, hops will be ue -> bs2 -> zone1 -> edge1 so that's 3 hops
• If connected to bs3 PoA, hops will be ue -> bs3 -> zone2 -> edge2 so that's 3 hops but not the same hops as bs2
• So as you can see, the "number" of hops might be the same, but the hops themselves are different and may have different network characteristics.

Regarding 2

• You can statically configure these in your scenario or sent a NET.CHAR event to change the values at run time
• Changing the number of hops requires changing the topology as you did in Publish Initial Release #1

Regarding 3

• Response time is the latency
• error rate is also configured in the scenario - see "packet loss" parameter
• the data (e.g. "bits") are sent by the application running in the UE... so assuming you are using iperf, you would need to change the parameters of the iperf command.

[Deu82]

@roymx ,

Thanks a lot! These explanations really help to clarify some of the things :-) 👍