This idea came to me when I first heard about The Central Asia Peering and Interconnection Forum (CAPIF) — an annual event where Central Asian economies discuss peering, Internet operational technologies, and networking.
Mongolians are very familiar with Kazakhstan and its people, sharing a strong historical connection not only with Kazakhstan but also with other Central Asian economies like Uzbekistan, Turkmenistan, Kyrgyzstan, and Tajikistan. As business collaborations between Mongolia and Central Asia continue to grow, I believe the demand for interconnection is also increasing, making connection quality an interesting topic.
In this post, I wanted to take a closer look at how Internet connectivity performs between Ulaanbaatar (Mongolia’s capital and business hub) and Almaty (Kazakhstan’s former capital and current economic centre). This would also give me a chance to experiment with Internet measurement using a few of the free tools available online.
Traceroute methodology
I performed a few traceroutes between these cities using a simple three-step process:
Step 1: Define the destination IP address
First, I needed to identify a popular upstream provider in Kazakhstan as my target. I checked the IPv4 interconnection map for Kazakhstan using APNIC’s REx (Figure 1). The dark dots in the centre of the graph indicate the points with the most local peers in Kazakhstan. Based on this, the most connected providers appeared to be AS9198 or AS41798.
To verify this, I checked the CAIDA rankings website but found no feature to filter Autonomous System Numbers (ASNs) by economy. Next, I used APNIC Labs Statistics, navigating to ‘Users per ASNs’ and sorting by economy by clicking on the CC button at the top of the columns. Due to the volume of data processed, it took some time to display the results. Eventually, KAZ Telecom AS9198 ranked first, confirming it as my target ASN.
To find the IP addresses announced by this ASN, I used Hurricane Electric’s BGP Toolkit — a tool I use daily. I searched for AS9198 to view all the IPv4 prefixes it announced.
I examined prefixes tagged with Almaty and found that the 37.150.8.0/21 block was sub-assigned to customers in Almaty. The RIPE whois lookup confirmed this:
- inetnum: 37.150.8.0 – 37.150.15.255
- netname: GPON_FTTH_SERVICES_KT
- descr: JSC Kazakhtelecom, Almaty Affiliate
- CC: KZ
Step 2: Find the source IP addresses
Being Mongolian myself, I found it straightforward to identify well-connected ISPs and their IP resources. For a different economy, I would repeat the previous step to find well-connected ASNs and IP addresses. I chose AS45204 (Gemnet) and AS55805 (Mobicom Networks) as my source networks.
Step 3: Perform traceroutes
The most crucial part was conducting traceroutes from Mongolian ASNs to Kazakhstan. I used the RIPE Atlas portal for this, which offers various measurement options. See the RIPE Atlas measurement results.
After watching some RIPE tutorials, I learned that measurements require credits (60 credits per measurement in my case). Since I didn’t have enough credits, I reached out to Lia Hestina from RIPE, who quickly transferred credits to my account. Thanks, Lia!
I logged into the Atlas portal, set up a measurement with the definition “Traceroute measurement to Kaz Telecom 37.150.8.100 from Mobicom Networks”, and obtained the results. You can verify the trace using the probe ID.
I also created another measurement from AS45204 (Gemnet Mongolia) to the same destination.
Initially, my traceroute yielded no results with the default settings. After consulting with ChatGPT, I changed the protocol from UDP to ICMP, which resolved the issue. Thanks, ChatGPT!
Observations and next steps
These results focus on measuring data reachability between upstream ISPs in Ulaanbaatar and Almaty. Real-world traffic between these cities may vary due to factors such as interconnection bandwidth, router configurations, and other technical details. However, we can still gauge basic connectivity using the available public tools.
Bayar Batjargal is an Internet Resource Analyst at APNIC.
Based on the original post at LinkedIn.
The views expressed by the authors of this blog are their own and do not necessarily reflect the views of APNIC. Please note a Code of Conduct applies to this blog.