Thank you so much, Dear Paul. As you said, RFC4786 is a very good document for the operation and configuration of Anycast service. However, how to decide the location and network of the node (DNS root mirror sites) is somewhat general: In Section 4.2: In general, node placement decisions should be made with consideration of likely traffic requirements, the potential for flash crowds or denial-of-service traffic, the stability of the local routing system, and the failure modes with respect to node failure or local routing system failure. In addition, it seems that the RSO needs to have more communications and understandings with the local operator to schedule the location and network of the deployed mirror site in order to optimize the overall performance of the root service and also fulfill the local requirements. From another aspect about the root service, I totally agree with you that the community can widely use some localization schemes (such as RFC7706, RFC8198...) and these schemes are more scalable than just add the mirror sites or new letters. YAN Zhiwei From: Paul Vixie Date: 2019-09-17 17:50 To: rssac-caucus Subject: Re: [RSSAC Caucus] Geolotation and BGP influence on the mirror site On Tuesday, 17 September 2019 03:23:00 UTC YAN Zhiwei wrote:
... For example, even a mirror site is deployed, the nearby resolution may reach other faraway node due to the decision of BGP (although yes, the geolocation is different from the network topology). Maybe the “local” mirror site can be deployed to attract more resolution traffic in the local network and the performance of overall DNS root service can still be improved (no matter of the deployed locations) from the global perspective, some metrics and measurements are still needed to optimize the deployment locations of the mirror sites. ...
zhiwei, the best current practice for anycast operations is RFC 4768. sections 4.3 and 4.4 are especially pointful. does your operational experience or laboratory prototyping/testing reveal any gaps or outdated concepts? my own observations of global cache miss traffic show me that the root zone is not statistically important, and i predict that more good can be done at least total cost of complexity by getting DNSSEC validation and QNAME minimization deployed in more recursive servers ("full resolvers") than by adding more nodes or more letters to the RSS, or by improving their local reachability. -- Paul _______________________________________________ rssac-caucus mailing list rssac-caucus@icann.org https://mm.icann.org/mailman/listinfo/rssac-caucus _______________________________________________ By submitting your personal data, you consent to the processing of your personal data for purposes of subscribing to this mailing list accordance with the ICANN Privacy Policy (https://www.icann.org/privacy/policy) and the website Terms of Service (https://www.icann.org/privacy/tos). You can visit the Mailman link above to change your membership status or configuration, including unsubscribing, setting digest-style delivery or disabling delivery altogether (e.g., for a vacation), and so on.