[Measurement-wg] Mobile traffic measurement project

Amreesh Phokeer amreesh at afrinic.net
Fri Jul 24 19:23:30 UTC 2020


Hello Hugo,

Thanks for sharing that would be very useful.

It’s good to see that both the client and the server are open source, which means that we can control both ends.

I’m happy to hear about the passive measurement mode, that is definitely one way to go. We would definitely be glad to hear more.

We are planning to recruit a few members of the community than can host a client. The ideal situation would be to work with regulators and/or ISPs and get a subsidized mobile line.

We intend to use USB dongles attached to RPis that we will deploy at strategic locations. Instead of creating our own measurement platform, we would use existing ones (RIPE Atlas, NDT, etc), for which we can easily access the data.

Thanks for this very useful head-up.

Cheers,
Amreesh





> On 24 Jul 2020, at 20:33, Hugo Salgado <hsalgado at nic.cl> wrote:

>

> Dear Measurement WG.

> My name is Hugo Salgado, from .CL, the ccTLD for Chile,

> South America.

>

> I attended your great webinar about measurements in Africa

> this week. Let me congratulate you for the excellent content,

> and especially for the interesting projects you have done

> and plan to carry out.

>

> In part of the webinar you discussed measurement systems for

> cellphones on mobile networks, and the lack of awareness of

> such projects. Please let me share a project in which we work

> with the NIC Chile laboratory, which belongs to the University

> of Chile.

>

> Its name is "Adkintun mobile". It consists of a client that is

> installed on android phones, and a server that receives

> measurements periodically. Both parts are open source, and you

> can try it yourselves:

> https://github.com/niclabs/adkintunmobile-server

> https://github.com/niclabs/adkintunmobile-androidclient

>

> The website of the project is in spanish, but you can see some

> research papers in:

> http://adkintunmobile.cl/#research

> with the results of studies from 2013-2015.

>

> As discussed in the webinar, one of the main problems with this

> type of measurement is getting volunteers to install the client,

> mainly due to the traffic used in active testing. This is why

> since last year we're working on a new project that performs only

> passive measurements, without generating any extra traffic from

> that is made normally by the user. This project is still in

> development, unfortunately its source is not available yet, but

> it will be as soon as it's in a mature state.

>

> Hope you'll find this information useful. Anyone can install the

> client just to try it from Google Play, install a local

> Adkintun architecture, or just use the code as a basis for a

> new tool. And let me know anything I could help!

>

> Best regards,

>

> Hugo Salgado

>

> _______________________________________________

> Measurement-wg mailing list

> Measurement-wg at afrinic.net

> https://lists.afrinic.net/mailman/listinfo/measurement-wg





More information about the Measurement-wg mailing list