Image: blog.apnic.net · rights & removal

[Podcast] ‘Just say no’ isn’t as simple as you think

Reporting by APNIC BlogRead the original at blog.apnic.net

Executive Summary

Geoff Huston and the interviewer discuss repeat queries observed in the APNIC Labs experiment related to DNS protocol usage changes over time. The number of these repeated queries has shifted due to changes in DNS protocol usage, resulting in records that may have multiple queries regarding different topics. Geoff investigates two areas: the effect of different ways of responding with 'no' on query volume and the workload for authoritative servers, intermediate resolvers, and clients. He also explores the impact of alternative DNS transport protocols, specifically User Datagram Protocol (UDP), on query repetition. While some argue TCP-like protocols have high connection state overhead unsuitable for DNS at scale, the growing volume of repeated queries suggests that end-to-end state maintenance might be less costly than repeatedly handling identical queries.

Facts Only

* Geoff Huston and an interviewer discussed repeat queries on the APNIC Labs experiment.
* The number of repeated queries has changed over time due to shifts in DNS protocol usage.
* DNS records now have multiple queries concerning different subjects.
* Geoff examined how different ways of answering 'no' affect query volume and work for servers, resolvers, and clients.
* Different responses, such as not responding at all, lead to varied outcomes regarding query handling.
* Geoff explored the impact of alternative DNS transport protocols on query repetition and load.
* The discussion touched upon the overhead of maintaining connection state versus repeatedly handling queries.

Full Take

The investigation centers on the mechanics of protocol communication, specifically the semantic meaning of negation in DNS responses and the performance trade-offs between connection management and repeated processing under network stress. The contrast between the perceived inefficiency of TCP/UDP state maintenance and the observed persistence of query repetition suggests a deeper structural tension regarding abstraction versus implementation overhead in distributed systems. The implication is that optimizing for protocol consistency (maintaining state) might be less computationally demanding than managing the repetitive workload generated by ambiguous or inefficient denial mechanisms, challenging conventional assumptions about protocol efficiency at scale. This prompts a necessary inquiry into whether system resilience should prioritize minimizing latency through connection persistence or reducing processing redundancy through clearer communication semantics.

From the original · APNIC Blog

In this episode of PING, APNIC Chief Scientist Geoff Huston and I continue a discussion about the repeat queries Geoff sees on the APNIC Labs experiment. Over time, this number has changed, as DNS protocol usage has changed.

Read the full story at blog.apnic.net

Sentinel — provisional

Raw score, 0 to 1 (uncalibrated)0.15

No strong signs of machine writing were found in the source article. Provisional estimate, not a finding that a person wrote it.

The model's own note (unverified)

The text reads like an excerpt from a technical podcast summary or blog post, demonstrating organic integration of complex ideas rather than synthetic optimization.

This looks only at the wording of the original source article, not at this page's AI-written sections. A small local AI model made this estimate. It has not been checked against known human and machine texts, so treat it as provisional. It cannot show who wrote an article.

What the model noticed in the source article

Varied sentence structure; conversational flow mixed with technical exposition.

Natural narrative flow connecting experimental observations to broader protocol implications without excessive hedging or mechanical transition usage.

Direct presentation of research findings tied to specific explorations (answering 'no', transport protocols) suggesting a direct source citation.

Content appears to be drawn directly from an academic/industry blog post, using domain-specific terminology appropriately.

Signs the model read as against machine writing

Direct citation/link reference to APNIC Blog and specific named individuals; inclusion of a direct quote attribution notice at the end.