DNS query duplication reaches 80% in global measurements

APNIC Labs has identified a widespread pattern of DNS query duplication, with nearly every observed query appearing at least twice. The finding emerges from an advertising-based measurement system that tracks DNS requests and web page fetches across approximately 30 million browsers, devices, and applications daily. While the system records around 150 million unique DNS labels each day, it logs approximately 270 million incoming queries—an 80% duplication rate that prompted further investigation into resolver behavior and transport reliability.

What the data shows

The measurement system relies on unique DNS queries to generate corresponding web page serves, allowing APNIC Labs to monitor internet-wide behavior. However, intermediary systems such as caches, proxies, and even persistent browser tabs can reissue requests, leading to repeated observations of the same DNS names or URLs. Unlike web fetches, where subsequent requests can often be discarded as redundant, DNS operates under unreliable transport conditions. Features like DNSSEC may generate SERVFAIL responses, necessitating repeated attempts to resolve a name-to-address mapping. These retries are not merely redundant but reflect the system's inherent design to ensure resolution under failure conditions.

Geoff Huston, APNIC Chief Scientist, noted that the duplication occurs within very short intervals, suggesting that resolvers or client stacks are rapidly retrying queries. The pattern raises questions about how many resolvers lie behind a single user's system and whether those resolvers vary in configuration—such as DNSSEC support—which could influence retry behavior. The findings indicate that duplication is not an anomaly but a systemic characteristic of modern DNS resolution.

Why duplication matters

For network operators and DNS service providers, the prevalence of query duplication has practical implications. Repeated queries increase load on resolvers and authoritative servers, potentially affecting performance and capacity planning. Operators may need to account for higher-than-expected query volumes when sizing infrastructure or evaluating caching strategies. The data also highlights the role of transport reliability in DNS design, where retries are a fundamental mechanism rather than an edge case.

The APNIC Labs findings also provide a benchmark for evaluating resolver behavior across different networks and regions. If duplication rates vary significantly between providers or geographies, it could indicate differences in resolver implementation, client software, or network conditions that warrant further investigation.

Next steps for the industry

APNIC Labs plans to continue analyzing the data to identify patterns in resolver behavior, such as whether certain types of resolvers or client stacks exhibit higher duplication rates. The organization has invited community contributions to expand the dataset and refine the understanding of how DNS retries impact global infrastructure. Operators and researchers can share insights or data by contacting APNIC Labs directly.

The podcast episode discussing these findings is available on major platforms, including Apple Podcasts, Spotify, and YouTube, with additional details published in a companion blog post on the APNIC Labs website.