Internet Society Virtual Peering Series - Central Asia, 7 Dec 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: VPS-CA#4 - Mapping Connectivity and the Digital Space in Central Asia

Description

Mapping Connectivity and the Digital Space in Central Asia
Louis Pétiniaud, Institut Français de Géopolitique
Kavé Salamatian, Université de Savoie

Virtual Peering Series – Central Asia #4: Where the Traffic Flows
7 December 2021

https://www.internetsociety.org/events/virtual-peering-series-central-asia/vpsca-4/

A

Hello everyone, I'm here with my colleague, kave salamtain.. We have started a research program.

A

Called mapping connectivity into the digital space using ripe data and tools in central asia.

A

So I'm going to try and keep it as far as much as I can from methodological issues regarding.

A

What we are doing,, but just as a contextual thing,, we are a team of geographers, kevin.

A

Limonier and I, and computer science, researchers such as kavé, and loqman in mathematics,, and we.

A

Are trying to put all these areas together to understand how the network and geography.

A

Interact, and also the network and geopolitics. this research stemmed from the idea that.

A

Internet and network mapping are not usually combined with methods of cartography that are.

A

Common in my field, and kevin's., we had two main research questions at the beginning of our work.

A

The questions are,. To what extent does the geopolitical context influence the shape of the.

A

Internet? and,. To what extent does the shape of the internet influence the geopolitical.

A

Situation of countries.

A

central asia as a case study, is interesting to improve our measurements and technologies,.

A

Because it is a kind of an enclave network, and is undergoing changes.

A

We are using, and we will be using, four different methods of data, collection,, analysis,.

A

And visualization. what we do is as-level graphs, and we use a geopolitical analysis of.

A

Graphs, we use latencies to map connectivity, and we use traceroute and geolocation to map the.

A

Geography of the network, and finally, we'll be doing our fieldwork in kyrgyzstan. If covid.

A

Allows it. the first one is as-level graph.. What we do is use the data of the as path to reveal the.

A

Existence of links between autonomous systems, and we infer the as topology, and we updated.

A

Following the updates., we know that the inferred topology is not fully complete,. However, we use.

A

Different feeds of data,, including ripe, routing information system,, we take one hour to get an.

A

Almost full, as graph of the whole world, with about 70,000 autonomous systems.

A

What we get is basically graphs like that.. This is one of the post soviet countries, and our.

A

Methodology consists in trying to uncover what are the geopolitical constraints that produce.

A

This graph. here you have all the autonomous systems in the pacific countries with the colors.

A

Depending on the country,, and the idea is to do the same on central asia, to try and understand.

A

What are the geopolitical constraints that produce the connectivity in central asia.

A

This is an example from ukraine to let you see what we have been able to do. it's a lengthy.

A

Work to understand why and how crimea and the donbass are slowly drifting away from the.

A

Ukrainian network,, and for that we use both usual sources of data from ripe and whois.

A

Databases,, as well as open source, intelligence sources, and the idea is to apply this kind of.

A

Methodology to central asia again., we have other works on the way.. This is actually the work of loqman salamatian.. I don't know if.

A

Kavé, you want to...

B

The issue that we have, from a methodological perspective, is that we want to map as graph.

B

That is a logical entity to the geographical space.

C

So, there's a question,: why have you decided to do it in kyrgyzstan?.

A

It was kevin's and my decision, actually,, because we think it will be the easiest place to.

A

Do fieldwork. some countries are cautious regarding research on the network, and we just.

A

Want to be able to talk as much as possible with stakeholders in the field., so basically, that.

A

Was the main reason. actually,? Can, I add, also one element to what you just said. in fact,, that the issue is that,.

B

Myself, I have been trying to make some contact with some other academic institution in the.

B

Region, and it was not an easy task,, because most of institution they have not so many.

B

Researcher working on internet issues, and also sometimes there are reluctant to work with.

B

Foreign researchers., one of the reasons that we also have this discussion today in this meeting.

B

Is, in fact, to ask for collaboration from the region. if academic institution, operator, and.

B

The stakeholder, are interested about gathering more measurement about the topology about the.

B

Infrastructure, to make a kind of analysis of it,, we are very interested to collaborate with.

B

Them,, but we will need to have some partners in this kind of they will be another. Ask that I'm.

B

Going to do at the end of the technical presentation, for concrete element, because we.

B

Will not be able to do everything online and just looking at the bgp feed, and we need to.

B

Have a more concrete view about some infrastructural elements.. This is also a call.

B

For collaboration that we are launching here., I had some contact at some point of time with.

B

Uzbekistan with the university of tashkent,, I have not been able to be in touch with them in.

B

The past three years,, and also similarly with kazakhstan, and other institutions,, we are.

B

Searching for contact and collaboration.

B

for everything to do with the academic research institution in central asia,. I believe this is.

D

Not exactly the right way to go,, because the academic institutions do not have the holistic.

D

Picture, and primarily the technical one., that's number one, and everything to do with.

D

Regulation,, including among others., as far as I know,, the academic institution, everywhere --, I.

D

Know in tashkent,, in bishkek,, in kazakhstan, in tajikistan,, are outside of the scope of this.

D

Activity. we tried to do the internet service provider research, and the only information we.

D

Could get, we could only get from the businesses, from the providers with the national.

D

Operator,, so that's all we could get, and the central asian market is special,, because it's.

D

Only one or two companies that have the monopoly in the market. for instance, take kazakhstan,.

D

95% / 96% belong to kazakh telecom.. I say, again,. This is trunk network,, then several.

D

Percent belong to two to three companies,. One of them also is state owned, and kazakhstan.

D

Delivered close to 100% of the traffic, say, to kyrgyzia., they have a smaller channel to.

D

Afghanistan, but I guess they get less than 1% there. all main traffic goes through kazakhstan,.

D

And so my question is,, how are you going into kyrgyzia to consider what you have been showing.

D

Those maps and everything you're told today?. How are you going to match it to kyrgyzia,.

A

Actually,, I'm not sure about the relationship between the maps nda, the fieldwork,. The maps are.

A

Not done using that. however, having a foot in kyrgyzstan will help us make measurements.

A

Again,, these are mostly the work of logman salamatian.. I can give you his email. on the.

A

Other hand,, the idea also is to map the discrepancies between what we can see from.

A

Outside of the country, and what we may find inside,, because you always have discrepancies.

A

Between when you are inside the network, to when you observe it from afar.

B

Something that I wanted also to add is that,. Thank you for the comments that was given in.

B

Russian, and we are academic people., our natural tendency is to first go to academic institution.

B

If you can help us in finding contact with operator,, it will be also interesting for us.

B

That's just also the same kind of data that we have, produced with a strong methodological.

A

Approach, and this one is a very clear, cut, and, let's say, qualitative example.. That was just a.

A

Traceroute, that was part of an experience to understand how borders are built in the digital.

A

Space., what we also aim to do in central asia is using traceroutes, like that,, multiply them of.

A

Course, trying to find again the geopolitical constraints that could drive the path of data.

A

This map just shows how the donbass border is being created as well in the digital space,.

A

Because data from dnipro will go to moscow through europe,, even though there are more.

A

Direct links that could be used.

A

also,. One of the end goals of that is actually to generate ideas and methodologies to create a.

A

General framework to think about internet geography.. This is also why we chose central.

A

Asia, because it provides an interesting case as a small and more or less enclave network, to.

A

Develop what we are doing. in this case, fieldwork would help us to actually not only.

A

Find data,, but also to find what data is missing.

A

And I guess that's it.

C

I have a question myself., where from have you taken this map? so,? This data for the map, is a.

C

Itu one or something else?.

A

No,, this is a map that is, the work,, is developing right now in colombia.

A

D

It's do not see any direction that go from china direct through russia to europe and other.

D

Cities., I cannot see them here.. I do not see the caspian hub. That from aktau goes to baku and.

D

Direct channel to italy.

A

No,, this is not a complete map of cables.. It's a work in progress, aiming at uncovering the.

A

Location of cables,, including terrestrial cables.,.

C

We'll follow that,, but we're quite short in time and let's go for the next presentation.