{"id":42327,"date":"2026-04-29T15:06:52","date_gmt":"2026-04-29T12:06:52","guid":{"rendered":"https:\/\/web-dev.iptp.net\/?p=42327"},"modified":"2026-04-29T15:26:15","modified_gmt":"2026-04-29T12:26:15","slug":"ipv6-asia-pacific-region-report","status":"publish","type":"post","link":"https:\/\/web.ams.iptp.net\/vi\/blog\/ipv6-asia-pacific-region-report\/","title":{"rendered":"IPv6-first as the New Normal: What the APNIC Region Data Shows"},"content":{"rendered":"

IPv6 in the Asia-Pacific Region: How the Market Is Moving from Deployment to Operational Maturity<\/h2>

\"\"<\/p>

A major new phase is beginning for the IPv6 market in the Asia-Pacific region. The issue is no longer simply whether the new protocol has been deployed, but how deeply it has been embedded into real-world operations: user traffic, access architecture, routing security, and network management practices. That is why the latest APNIC data and the related discussions at APRICOT 2026 should not be seen as a set of separate news items, but as signs of the industry’s systemic transition into a more mature phase.<\/p>

For infrastructure providers and enterprise teams building their long-term IPv6 strategy<\/a>, this shift matters at both the architectural and operational levels.<\/p>

A Milestone That Matters in Practical, Not Symbolic, Terms<\/h2>

In April 2025, APNIC Labs reported that the 30-day average IPv6 capability had exceeded 50% across all 56 economies in the APNIC region. For the industry, this is an important milestone: it was reached 25 years after the region began its IPv6 journey and reflects not isolated pilots, but the large-scale presence of IPv6 in the real user environment. APNIC Labs builds this metric on approximately 25 million end-user tests per day, comparing the reachability of IPv4-only, IPv6-only, and dual-stack addresses, and then aggregating the results by economy and origin AS using World Bank Internet user statistics.<\/p>

This is especially important because APNIC measures not only the presence of IPv6 as such, but also the actual ability of end systems to work over it. The methodology distinguishes between two states: IPv6 capable<\/strong>, when a device can fetch a URL over IPv6, and IPv6 preferred<\/strong>, when it actually chooses IPv6 even when the service is available over dual stack. For operators and enterprise networks, this distinction is critical: having IPv6 enabled does not yet mean that the network is truly operating according to an IPv6-first model.<\/p>

\"APNIC
\nAPNIC infographic: IPv6 capability has exceeded 50% across all 56 economies in the region, 25 years after the start of the region’s IPv6 journey, based on approximately 25 million end-user tests per day.<\/em><\/sub><\/p>

Why the APNIC Region Is Especially Indicative Today<\/h2>

From a global perspective, the role of the Asia-Pacific region has long outgrown the status of a regional case study. According to APNIC, the region’s economies account for 64% of all IPv6 users worldwide, while aggregate IPv6 capability stands at 50%. This has a significant impact on the global picture, because the global average in the same view is estimated at around 40%. In other words, APNIC is not merely catching up with the global Internet on IPv6; it is to a considerable extent shaping the global trend itself.<\/p>

APNIC also points out separately that leadership among individual economies in the region is determined by a combination of three factors: the scale of the Internet user base, government support for IPv6, and capital investment in infrastructure. The largest deployments are found in China, India, Japan, Viet Nam, Thailand, Indonesia, South Korea, the Philippines, and Pakistan. This is an important point: IPv6 becomes mainstream not only where there is technological readiness, but also where institutional drivers and long-term investment in access networks are present.<\/p>

The Largest APNIC Economies by IPv6 Deployment Volume<\/h2>\n\n\n\n\n\n\n\n\n\n\n
Economy<\/strong><\/td>\nIPv6 capability<\/strong><\/td>\nUsers<\/strong><\/td>\n<\/tr>\n
China<\/td>\n45.28%<\/td>\n810,519,875<\/td>\n<\/tr>\n
India<\/td>\n78.16%<\/td>\n600,671,221<\/td>\n<\/tr>\n
Indonesia<\/td>\n14.93%<\/td>\n115,444,083<\/td>\n<\/tr>\n
Japan<\/td>\n58.2%<\/td>\n109,438,612<\/td>\n<\/tr>\n
Philippines<\/td>\n17.66%<\/td>\n79,160,582<\/td>\n<\/tr>\n
Vietnam<\/td>\n60.18%<\/td>\n54,269,706<\/td>\n<\/tr>\n
South Korea<\/td>\n23.75%<\/td>\n51,142,259<\/td>\n<\/tr>\n
Thailand<\/td>\n50.62%<\/td>\n45,014,196<\/td>\n<\/tr>\n
Pakistan<\/td>\n20.59%<\/td>\n40,485,679<\/td>\n<\/tr>\n<\/table>

Source: APNIC Labs \/ APNIC Blog, data as of 22 April 2025.<\/em><\/sub><\/p>

What the 50% Threshold Means in Practice<\/h2>

The “50%+” figure is easy to perceive as a nice milestone, but its practical significance runs deeper. It means that IPv6 in the region can no longer be viewed as a peripheral or optional technology. For a substantial share of networks, it has already become a mainstream transport layer that affects architectural decisions, equipment requirements, access models, and service development logic. This is no longer a matter of preparing for the future, but part of current operational reality.<\/p>

At the same time, APNIC explicitly notes that the global Internet will remain dual-stack for the foreseeable future. In their assessment, the transition to IPv6 is continuing at a linear rate and stands at 34%, while further growth is being driven in particular by the mobile segment amid the rollout of 5G and future 6G networks. In fixed broadband, the pace may be slower because of longer equipment refresh cycles, but where upgradeable CPE is used and national modernization plans exist, deployment is accelerating.<\/p>

This leads to an important operational conclusion: at this new stage, the market is not moving toward an immediate abandonment of IPv4, but toward an IPv6-first model with controlled and gradually shrinking dependence on IPv4. That is also why understanding the operational and security implications of dual-stack risks<\/a> becomes essential rather than optional.<\/p>

\"From
\nFrom dual-stack coexistence to IPv6-first operations with controlled residual IPv4<\/em><\/sub><\/p>

Sign of Maturity No. 1: IPv6 Is Becoming Part of the Institutional Agenda<\/h2>

Characteristically, at APRICOT 2026 \/ APNIC 61, discussions during the NIR SIG session were no longer focused on whether IPv6 is needed, but on more practical issues: the current state of deployment, RPKI growth, new data analysis tools, and even a review of coordination formats between national Internet registries. The framing itself is revealing: when the market is discussing not whether to deploy, but how best to manage a mature ecosystem, that already reflects a different level of development.<\/p>

Following the APNIC AGM in March 2026, the organization separately stated that advancing IPv6 and RPKI remains a strategic priority, and that training and hands-on capacity building will be scaled through a train-the-trainer model and updated labs focused on IPv6, RPKI, routing, and network security. This confirms that, for the regional ecosystem, IPv6 is no longer treated as a standalone initiative, but as part of a broader task: developing a resilient and secure Internet infrastructure.<\/p>

At the infrastructure layer, this broader task also includes how organizations source and manage IPv4\/IPv6 services <\/a>as part of a production-ready network stack.<\/p>

Sign of Maturity No. 2: Economies in the Region Are Moving at Different Speeds, but in the Same Direction<\/h2>

The NIR SIG materials clearly show that, within the region, the pace varies, but the overall vector is shared: IPv6 growth is accompanied by greater attention to routing security and to a more systematic model of resource governance. In some economies, the emphasis is on mass deployment; in others, on RPKI; in others still, on preparation for IPv6-only scenarios.<\/p>

Some Illustrative Signals from the APNIC Region<\/h2>\n\n\n\n\n\n\n\n\n\n\n
Economy \/ NIR<\/th>\nConfirmed indicators<\/th>\n<\/tr>\n<\/thead>\n
Taiwan \/ TWNIC<\/td>\n336 members; IPv6 user availability at 61.36%; over 97% valid IPv4 prefixes and 98% valid IPv6 prefixes by RPKI validity<\/td>\n<\/tr>\n
Japan \/ JPNIC<\/td>\n518 IP members; 5,686 allocated IPv6 \/32 blocks; among 549 LIRs, ROA adoption grew by 17% year-on-year<\/td>\n<\/tr>\n
China \/ CNNIC<\/td>\n1,000 members; about 332,000 IPv4 \/24 blocks and more than 32,000 IPv6 \/32 blocks; around 60% of members hold IPv6 resources; 869 million IPv6 users by the end of 2025<\/td>\n<\/tr>\n
Viet Nam \/ VNNIC<\/td>\nmore than 1,000 members; IPv6 usage at around 65%; a national roadmap toward IPv6-only is being developed with a horizon through 2030<\/td>\n<\/tr>\n
Indonesia \/ IDNIC<\/td>\nmore than 5,000 members, including over 1,000 ISPs; IPv6 growth has accelerated since 2020; ROA coverage for IPv4 at around 89%<\/td>\n<\/tr>\n
India \/ NIXI<\/td>\n4,649 affiliates as of January 2026; IPv6 adoption at around 77.6%; notable growth in IPv6 allocations while IPv4 has stabilized<\/td>\n<\/tr>\n
Republic of Korea \/ KRNIC<\/td>\n1,093 members; RPKI adoption for IPv4 remains low at around 2%; production RPKI rollout is planned in phases through 2027<\/td>\n<\/tr>\n<\/tbody>\n<\/table>

Source: APNIC NIR SIG, APRICOT 2026 \/ APNIC 61.<\/em><\/sub><\/p>

What makes this table important is not just the numbers themselves, but the broader logic behind them. In China and India, the effect of scale and national programs is evident; in Viet Nam, there is visible movement toward IPv6-only planning; in Japan and Taiwan, mature deployment is combined with a strong culture of routing security; in Indonesia, fast growth is clearly supported by regulation; and in South Korea, IPv6 is progressing while RPKI still lags behind. The region is moving forward unevenly, but it is no longer in the phase of fragmented experiments.<\/p>

Sign of Maturity No. 3: The Market Is Beginning to Measure Not Only Address Space, but Real Network Behavior<\/h2>

Another important signal is the transformation of the analytical tools themselves. In March 2026, Internet Society Pulse described APNIC Resource Explorer (REx) not simply as a statistical portal, but as a platform that combines data on IPv4\/IPv6\/ASN allocations, historical trends, transfer records, IPv6 deployment indicators, RPKI coverage, DNSSEC, QUIC, and interconnection topology. For the industry, this is an important shift: assessing network maturity is becoming less about how many resources have been allocated and more about how the network actually operates.<\/p>

In this context, APNIC Labs measurements are especially important. They make it possible to identify the gap between IPv6 availability and its actual preference in traffic, and therefore to detect hidden fallback mechanisms, unsuitable application settings, legacy segments of infrastructure, and the points where dual stack still effectively drags a network back toward IPv4. This is no longer just an addressing issue, but an issue of observability and operational discipline.<\/p>

From a transport perspective, that discipline increasingly affects how traffic is engineered across IP Transit<\/a> environments and how consistently IPv6 is treated as the preferred production path.<\/p>

A Practical Case: Why Even a Well-Configured Network Still Generates Too Much IPv4<\/h2>

The practical side of the issue is particularly revealing. At APRICOT 2026, a case study was presented on a small but technologically advanced SOHO network in which the initial share of IPv6 traffic was 67.7%, based on a 16-day NetFlow analysis. At the same time, the network itself was already very well prepared: it used native IPv6 from the provider, an IPv6-oriented DNS infrastructure, SLAAC with RDNSS, a MikroTik gateway, and an IPv6-first approach for internal services. In other words, this was not a raw network, but an environment many would already describe as well optimized.<\/p>

The methodology of the study was highly practical: first, a baseline traffic picture was established; then applications with a low IPv6 share and high IPv4 bandwidth consumption were identified; after that, targeted configuration changes were applied; then the effect was measured; and finally, the inevitable IPv4 floor was isolated — the residual IPv4 that could no longer be removed through simple tuning. This approach is important in itself, because it shows that the road to IPv6-first rarely runs through abstract declarations; it almost always requires measurement and step-by-step optimization.<\/p>

The main lagging application in this case turned out to be the BitTorrent client. Despite a working dual-stack environment, its IPv6 share was only 44.3%, and a significant portion of traffic continued to go over IPv4 because of default behavior and fallback to IPv4 peers. After one targeted change — binding qBittorrent exclusively to the IPv6 interface — the picture changed radically: the IPv6 share for this application rose to 92.6%, while the average for the entire network increased to 79.2%.<\/p>

Change in the Share of IPv6 in Application and Overall Traffic<\/h2>\n\n\n\n\n\n\n\n\n\n
Metric<\/th>\nBefore the change<\/th>\nAfter the change<\/th>\n<\/tr>\n<\/thead>\n
BitTorrent: IPv6 share<\/td>\n44.3%<\/td>\n92.6%<\/td>\n<\/tr>\n
BitTorrent: average IPv6 traffic<\/td>\n0.86 Mbps<\/td>\n4.72 Mbps<\/td>\n<\/tr>\n
BitTorrent: average IPv4 traffic<\/td>\n1.08 Mbps<\/td>\n0.38 Mbps<\/td>\n<\/tr>\n
Entire network: IPv6 share<\/td>\n67.7%<\/td>\n79.2%<\/td>\n<\/tr>\n
Entire network: average IPv6 traffic<\/td>\n4.82 Mbps<\/td>\n7.08 Mbps<\/td>\n<\/tr>\n
Entire network: average IPv4 traffic<\/td>\n2.30 Mbps<\/td>\n1.86 Mbps<\/td>\n<\/tr>\n<\/tbody>\n<\/table>

Source: Project IPv6-first case study, APRICOT 2026.<\/em><\/sub><\/p>

The key takeaway here is that a dual-stack environment can create an illusion of well-being. The application seems to work, but in fact continues to use a suboptimal IPv4 path. In the case under discussion, this was because the client was receiving IPv4 peer addresses and initiating outbound connections through NAT, even though a full native IPv6 path was already available. In other words, the problem was not the lack of IPv6, but the logic of transport path selection at the application level.<\/p>

Why IPv4 Does Not Disappear Even After Optimization<\/h2>

After optimizing BitTorrent, the author of the case study separately analyzed the remaining IPv4 traffic and concluded that most of it was already caused by external limitations. These included IPv4-only IoT hardware, certain legacy web services, and features of the provider’s infrastructure. And this is perhaps one of the most important lessons of the whole story: in a mature network, the bottleneck is increasingly found not inside the network itself, but outside it.<\/p>

This conclusion aligns well with Internet Society Pulse data. On its Enabling Technologies panel, the current share of the world’s top 1000 websites that support IPv6 is listed at 50%. In other words, even if the access network, internal services, and a large share of applications are ready for IPv6-first, the external web ecosystem still relies on the old model by roughly half. That is why the path to full IPv6-only scenarios in most cases requires either transition mechanisms or very careful traffic and service segmentation.<\/p>

For distributed enterprises, this also has implications for how managed connectivity services<\/a> are designed across sites, providers, and regions when IPv6 and residual IPv4 must coexist predictably.<\/p>

What the IETF Says About This: IPv6-first Is Not an Ideology, but a Working Architecture<\/h2>

It is important that these scenarios are described not only in blogs and measurement projects, but also in IETF documents. RFC 6877 defines 464XLAT<\/strong> as an architecture that enables limited IPv4 connectivity across IPv6-only edge networks without encapsulation. RFC 8683 separately addresses additional deployment recommendations for NAT64\/464XLAT in mobile operator networks, broadband provider environments, and enterprise networks, including DNS64, applications with literal IPv4 addresses, and devices that are not IPv6-ready. This means that the model of IPv6-first with controlled residual IPv4 is not merely theoretical, but a recognized operational approach.<\/p>

For the market, this is an important signal. At the mature stage, the question is less and less “Are we ready for IPv6?” More often, it is framed differently: “What residual IPv4 is actually needed?”, “Which network segments can already be run as IPv6-first?”, “Where are translation scenarios required?”, and “Which applications still prevent us from moving away from IPv4?” That is what the practical agenda of the transition period looks like.<\/p>

This becomes even more relevant in hybrid and multi-cloud environments where direct cloud connectivity<\/a> can influence path stability, latency, and the consistency of IPv6-first design choices.<\/p>

\"IPv6-first
\nIPv6-first networks can preserve limited IPv4 reachability through translation mechanisms such as DNS64, NAT64, and 464XLAT<\/em><\/sub><\/p>

The Second Half of Maturity: IPv6 Without RPKI Is No Longer Enough<\/h2>

At the current stage, it is impossible to talk about IPv6 maturity in isolation from routing security. Internet Society Pulse notes that, although RPKI has existed for more than ten years and was standardized by the IETF back in 2012, in 2024 only around half of the IP addresses announced in BGP were covered by RPKI records. Pulse also reminds readers of the basic mechanics: it is not enough to simply know about RPKI; networks need to publish ROAs for the address space they use so that other networks can validate routing data.<\/p>

This is exactly why the APNIC region’s story is especially important. Here, IPv6 growth is increasingly accompanied by discussion of RPKI, ROA publication, and route validation. This can be seen both in the NIR SIG materials and in APNIC’s strategic priorities for 2026. Taiwan, Japan, and Indonesia demonstrate a fairly mature logic: IPv6 is not considered separately, but in connection with routing security and with resource management practices. It is in this combination that the transition to true operational maturity becomes visible.<\/p>

What All This Means for Operators, Clouds, and Enterprise Networks<\/h2>

If we put these signals together, the picture becomes fairly clear. The IPv6 market in the Asia-Pacific region has already passed the stage where the main issue was simply the existence of the new protocol. The next stage is about:<\/p>