If you have a question about DNS for Mr. DNS, he’d love to hear it. Drop him a line at firstname.lastname@example.org.
We’re back again, scraping the bottom of the mailbag for questions. Erik Radde helped us out with a question on the interaction of wildcards and the search list, and Lenny Tropiano tweeted a question at Mr. DNS about Dyn’s support for a feature that provides CNAME-like semantics at a zone apex. Along the way there were detours into the three laws of thermodynamics and, more importantly as the AI revolution grows ever closer, the three laws of robotics.
Well, we said we’d try to keep to a monthly schedule, and we arguably just made it! This episode, number 44, features a special guest: Andrew Sullivan, Matt’s colleague at Dyn and Chair of the Internet Architecture Board. Now, if we’d planned ahead and let you know Andrew was going to be on the show, we could have let you know so that you could have submitted lots of thoughtful questions for him to answer, but by now you know not to expect that kind of forethought from us. Instead, we asked him about stuff we’re interested in, including the IANA transition and ARCING, an IETF effort to identify alternative resolution contexts. We also answer a question from Sheridan West about some suspicious-looking log messages from his name server and one from Jeff Helman about the right DNS configuration for handling multiple back-end web servers.
In this, our holiday episode, we’re joined by returning special guest, Duane Wessels, who discusses a recent event involving the root name servers and a lot of obviously spoofed traffic, as well as his ongoing work in the IETF around DNS privacy. We reach into the mail bag and find a question from our friend, Rob Fleischman, musing about possible additional metadata that recursive servers could send to authoritative servers. As it happens, Duane’s also working on a DNS protocol extension directly related to Rob’s question, which he tells us about. Finally, we end with a brief and spoiler-free discussion about The Force Awakens.
In Episode 42, we discuss the meaning of life, the universe and everything with a very special guest, @dnsreactions, creator of the hit DNS Reactions Tumblr. “DR”, as we call him or her (or it?), prefers to stay anonymous, so we have obscured his/her/its voice using the magic of technology. Our long-suffering listeners submitted questions for DR, who was very accommodating. Enjoy!
Welcome to our special Halloween episode! Okay, not really, but we are recording in late October… This time we answer a record-breaking three questions from the same listener, Grant Taylor, who single-handedly supplied the material for all our tangents in this show. We remind everyone of the dangers of cache poisoning in a discussion about CNAMEs, we strain our memories back to the early days of DNSSEC to discuss SIG(0), and we explain and opine on EDNS Client Subnet, a recent and increasingly popular DNS protocol extension. Considering the time of year, we also lapse into a discussion of candy, specifically peanut M&Ms.
In this 40th episode–a milestone!–Matt and Cricket answer long-suffering listener Grant Taylor’s question about sorting replies by type and wander into the Land of Happy Eyeballs, then explore an answer Joe Abley received from Mark Andrews of ISC. Meanwhile, a discussion of the term G-job causes Matt to recount accidentally insulting a group of public servants, and both Matt and Cricket discuss their hope that the new AppleTV will lead to the end of their paying for streaming content they could have watched for free.
In this star-studded episode, Matt and Cricket take advantage of a meeting of the DNS Cabal–that is, the annual “Inside Baseball” event–to answer Donald Rudder’s question about whether synthesizing NXDOMAIN responses to avoid random subdomain attacks would work with NSEC3 as well as NSEC records. This is followed by a wildly entertaining (by DNS standards, anyway) discussion of the future of DNS, new TLDs, communication in the event of attacks, and more.
Guest-starring some of the brightest lights in DNS, including Kris Beevers, Brian Brady, David Dagon, Casey Deccio, Rob Fleischman, Olafur Gudmundsson, Shumon Huque, David “Tale” Lawrence, and Duane Wessels.
In this episode, long-time (and likely now sole) listener Yiorgos Adamopoulos asks about the the process of signing the root zone, which Mr. DNS has some experience with. Matt also recaps some of the goings-on at the latest DNS-OARC meeting in Amsterdam, omitting that which must stay in Amsterdam, but revealing some lapses from his DNSSEC RFC-editing days.
Back after a long absence they try to avoid talking about, Cricket and Matt tackle some meat-and-potatoes questions: Why can’t one have a CNAME with other records at a domain name? Are registrars buying up expired domain names? How can one make a name server generate answers dynamically? Listen as Matt embarrasses himself by forgetting the name of the Registry-Registrar Protocol (RRP), the predecessor to the Extensible Provisioning Protocol (EPP), used today between registrars and registries. Cricket’s memory is working fine, though, as he dredges up a reference to lbnamed, a simple, Perl-based name server now remembered only by Google and DNS geeks. And as usual, there are tangents: the episode winds up with an impromptu discussion of standing desks and how Matt is an effective but not-at-all-subtle choral conductor.
In this episode, Matt and Cricket respond to Tommi Nikkilä’s followup to his original question about the legality of multiple CNAME records in a DNS answer, and then react to (to claim they “answer” it is a reach) dedicated listener Yiorgos Adamopoulos’s question about registering domain names with underscores in them. On the way, Matt describes his quest to set a personal record in his commute from his home in Bethesda to Dyn’s headquarters in Manchester, New Hampshire, and then (inadvertently?) disses Cricket’s manhood by suggesting that Real Men Drive with Standard Transmissions. Finally, the guys bemoan their lack of questions, implying that this is somehow responsible for their sporadic production, when we know in fact it’s their own damn fault.