The Invisible Backbone: Why Bangladesh is falling behind in the Internet Revolution

An Exclusive Interview with Geoff Huston, Chief Scientist at APNIC

Recently, Senior Journalist Mohammad Kawsar Uddin visited Kuala Lumpur, Malaysia to attend the APRICOT 2025 & APNIC 59 conference. During the conference Mr. Kawsar talked with Geoff Huston, Chief Scientist at APNIC, where he conducts research on Internet infrastructure, IP technologies, and address distribution policies. He is widely regarded as the foremost expert on IPv4 exhaustion and is frequently referenced by international agencies and quoted by the ICT media. He is also an active member of the Internet Engineering Task Force (IETF).

Question: What is your general observation about Bangladesh’s IPv6 uptake compared to the rest of the world?

Geoff Huston: This is a data-driven question, so I’ll need to refer to our measurements at APNIC. We use a unique method to measure IPv6 adoption through Google’s advertising network. Every day, we run about 30-35 million ads across the internet, each containing a script that checks whether users can access IPv6-only resources. This allows us to obtain an accurate, up-to-date view of global IPv6 adoption.

Looking at Bangladesh specifically, we see that the country began its IPv6 journey in late 2021 and early 2022. On March 23, 2022, one of the major providers rolled it out. Currently, Bangladesh has an overall IPv6 adoption rate of about 2%, which is significantly lower than neighbouring India’s 78%.

The three major providers in Bangladesh- Grameenphone, Axiata, and Banglalink, show relatively strong IPv6 deployment rates, ranging between 35% and 60%. However, their market share is relatively small. The numerous smaller ISPs in Bangladesh have yet to deploy IPv6. This is a common challenge: large companies with sufficient resources can afford to hire expertise for deployment, whereas smaller operators with limited technical staff struggle to do so.

Question: What initiatives can the government take to expedite IPv6 deployment in the country?

Geoff Huston: Bangladesh has a unique market structure compared to most countries. In many nations, the telecommunications market has consolidated around three or four major ISPs that collectively control about 90% of the market, with only a few niche providers serving specialized segments.

However, Bangladesh has not followed this pattern. The market here is highly fragmented, with numerous small operators. This fragmentation is important because IT operations benefit from economies of scale- larger providers can operate more cost-effectively than smaller ones. A single large ISP serving 100 million users will have a lower cost per user than 100 small ISPs each serving a fraction of that market. Bangladesh has three larger providers, but they are not dominant enough, and the market remains divided into many small segments.

Governments typically avoid intervening in markets, as telecommunications industries have been deliberately deregulated to encourage competition and consumer choice. In most countries, market forces lead to natural consolidation, where larger providers thrive and smaller ones are acquired. This type of consolidation has not yet occurred in Bangladesh, which is unusual from a global perspective.

Question: What advice would you give to ISPs and telecom operators in Bangladesh regarding IPv6 adoption?

Geoff Huston: In a market-driven system, service providers must offer services that customers need at prices they are willing to pay. If your services don’t meet consumer demand, your business will struggle.

Interestingly, IPv6 itself is not a selling point for customers- they won’t pay extra for it. What they care about is reliable and affordable service, particularly the ability to stream high-quality video without interruptions. If an ISP can provide that at a competitive price, it will thrive. If it cannot maintain service quality at reasonable rates, customers will switch to providers who can. In a competitive market, consumers make rational choices based on price and quality, which ultimately drives market evolution.

Question: Bangladesh has strong RPKI ROA uptake, but RPKI Origin Validation remains low. What is your observation on this?

Geoff Huston: Security is a complex field where decisions are often driven by mandates rather than independent risk assessments. RPKI does not necessarily make networks more secure in a broader sense- it primarily helps prevent accidental misconfigurations.

I understand the hesitation to fully implement RPKI Origin Validation. Handing over control of route filtering to an automated system is a significant step, and engineers are naturally cautious about allowing automated systems to control critical infrastructure. This is a standard conservative engineering approach- avoid changes that might cause unexpected failures in the middle of the night. That caution is both professional and appropriate.

Question: What can be done to improve RPKI Origin Validation?

Geoff Huston: This question touches on a deeper issue. RPKI provides less protection than many people assume. It primarily prevents accidental misconfigurations, but it is not very effective against deliberate attacks.

A major challenge is that the industry implemented partial security measures before completing the full security design. Currently, RPKI focuses on validating the origin of route announcements- ensuring that the entity creating the route is legitimate. However, routing security also requires protecting the path that routing information takes across the network. Existing technologies do not secure this path, allowing attackers to create deceptive yet seemingly legitimate routes.

As a result, while RPKI can catch accidental mistakes, it does little to stop sophisticated attackers from manipulating routes. Efforts to secure routing paths have been ongoing, but early solutions were too complex for widespread adoption, and newer proposals have been stuck in development for over a decade. Since there is no imminent solution, there is little urgency in deploying the current partial approach.

Question: Bangladesh has very low participation in the IETF. What challenges do you see?

Geoff Huston: I don’t view this as a major issue. The IETF plays a specific role in the industry- it primarily ensures that network equipment from different vendors adheres to interoperability standards, much like ensuring that all cars can drive on the same roads.

Since Bangladesh does not manufacture network equipment, there is limited need for local participation in IETF activities. Equipment vendors must engage with the IETF, but network operators generally do not.

For professional development in network engineering, organizations like APRICOT and regional network operator groups are more relevant than the IETF. These forums focus on practical engineering expertise rather than equipment standards. Thus, Bangladesh’s low participation in the IETF is not necessarily a concern- it may not be the most relevant platform for the country’s needs.

Question: What is your view on policymakers and non-technical participants engaging in the IETF?

Geoff Huston: IETF is not the right place for policymakers. The Internet has raised many important regulatory and policy questions regarding industry structure, digital infrastructure resilience, and more. These are critical topics, but IETF is not designed for such discussions.

In OECD member countries, regulatory and policy discussions occur in forums like the OECD itself. However, this organization includes only about 40 countries, and I am unaware of equivalent venues for non-member states. Regardless, the IETF is not a substitute for policymaking discussions.

Question: So, would you say that IETF is meant exclusively for technical professionals?

Geoff Huston: The IETF is most effective when it stays focused on technical matters. Some have attempted to bring broader discussions into the IETF, thinking they were missing an important aspect, but they quickly realize that the discussions there are highly technical.

For instance, topics like human rights and digital policy do not fit well within the IETF’s framework. The organization exists primarily to ensure that network equipment is safe, reliable, and interoperable for telecommunications providers. That is its core mandate, and expecting it to serve a broader role is unrealistic. If it successfully fulfils this mission, then it is doing its job.