Disclaimer: This post does not reflect the views or policies of SoylentNews Public Benefit Corporation (SN PBC), its staff, or my role as president. The opinions and statements within are my own, Michael Casadevall, and neither I nor SN PBC were financially compensated for this post.
There are times in life where you simply don't know where you will end up. For me, a chance encounter in Puerto Rico lead to a rather interesting series of events. I have spent the previous week (October 20th-26th) attending the ICANN 63rd International Public Meeting. For those who aren't familiar with the Internet Corporation for Assigned Names and Numbers (ICANN), it is essentially the not-for-profit organization that administrates the Internet root zone which forms the linchpin of the modern internet, and allows domain names such as soylentnews.org to exist.
As a fellow, I have been working to help advance policy from the perspective of Internet end-users, as well as improving access to the Internet in the form of Internationalized Domain Names. For those less familiar with the technical underpinnings of the Internet, I'll also talk a bit about DNS, and more of the work I am currently in the process of handling at ICANN.
In This Issue
- DNS - What is it?
- The Internet Root Zone, Top Level Domains, and Second Level Domains
- What Is ICANN?
- The Fellowship Program
- New Generic Top Level Domains
- String Contention and Name Collisions
- Internationalized Domain Names
- In Closing
- If You Want To Get Involved
DNS - What is it?
Before we can talk about ICANN in any fashion, we need to talk a bit about the Domain Name System, or DNS. Every computer on the internet is assigned one or more numbers known as IP addresses. IP addresses take the form of 18.104.22.168 or 2600:3c00::f03c:91ff:fe98:90b, and are used as ways of uniquely identifying every device. Every site and service has an IP address, as does your phone and computer to allow two way communication; they can be best thought of telephone numbers for computers.
Just like phone numbers, there needs to be a method to look up information based on name. In the days of old, we would use the Yellow Pages for this type of information. For computers, our version of the yellow pages is DNS, specifically what we call A and AAAA records. For example, here are two types of lookup requests for SoylentNews:
$ nslookup > server 22.214.171.124 Default server: 126.96.36.199 Address: 188.8.131.52#53 > soylentnews.org Server: 184.108.40.206 Address: 220.127.116.11#53 Non-authoritative answer: Name: soylentnews.org Address: 18.104.22.168 Name: soylentnews.org Address: 2600:3c00::f03c:91ff:fe98:90b
Besides basic name lookup information, DNS often contains information such as mail routing in the form of MX records, or even user authentication data in the form of Hesiod TXT records. It would be fair to call DNS the worlds largest distributed dynamic database. At its core, DNS is comprised of a network of worldwide servers that provide name lookup services from the internet, starting from the root of a tree.
The Internet Root Zone, Top Level Domains, and Second Level Domains
When I said root of a tree, I wasn't being very metaphorical. Let's look a little closer at an actual domain name, and break it down into its component parts, in this case, our development site at dev.soylentnews.org. For the purposes of demonstration, we'll walk the domain from the top down. Each section of a domain name is divided into levels separated by a period. So dev is a third level domain, soylentnews is a second level domain, and org is a top level domain.
Likewise, each of these levels points to the one higher than it. Let's ask the soylentnews domain about dev; our hosting provider (and DNS servers) are hosted by Linode so we'll query them directly.
$ nslookup > server ns1.linode.com Default server: ns1.linode.com Address: 2400:cb00:2049:1::a29f:1a63#53 Default server: ns1.linode.com Address: 22.214.171.124#53 > dev.soylentnews.org Server: ns1.linode.com Address: 2400:cb00:2049:1::a29f:1a63#53 Name: dev.soylentnews.org Address: 126.96.36.199 Name: dev.soylentnews.org Address: 2600:3c00::f03c:91ff:fe6e:d0a3
Notice that the "Non-authoritative answer" is missing. This is because soylentnews.org directly controls the level above it. We can see the same effect if we query the .org nameservers for SoylentNews; note that we need to ask for the NS record type which acts as a pointer to the next level of domain.
$ nslookup > server a0.org.afilias-nst.info. [...] > set type=ns > soylentnews.org [...] Authoritative answers can be found from: soylentnews.org nameserver = ns1.linode.com. soylentnews.org nameserver = ns3.linode.com. soylentnews.org nameserver = ns2.linode.com. soylentnews.org nameserver = ns4.linode.com. soylentnews.org nameserver = ns5.linode.com.
Cool! We can see the five nameservers that have authoritative data for soylentnews.org. But how did I find the nameserver for .org? Well, I queried the internet root zone for it.
nslookup > server a.root-servers.net Default server: a.root-servers.net Address: 2001:503:ba3e::2:30#53 Default server: a.root-servers.net Address: 188.8.131.52#53 > set type=ns > org Server: a.root-servers.net Address: 2001:503:ba3e::2:30#53 Non-authoritative answer: *** Can't find org: No answer Authoritative answers can be found from: org nameserver = a0.org.afilias-nst.info. org nameserver = a2.org.afilias-nst.info. org nameserver = b0.org.afilias-nst.org. org nameserver = b2.org.afilias-nst.org. org nameserver = c0.org.afilias-nst.info. org nameserver = d0.org.afilias-nst.org. a0.org.afilias-nst.info internet address = 184.108.40.206 a2.org.afilias-nst.info internet address = 220.127.116.11 b0.org.afilias-nst.org internet address = 18.104.22.168 b2.org.afilias-nst.org internet address = 22.214.171.124 c0.org.afilias-nst.info internet address = 126.96.36.199 d0.org.afilias-nst.org internet address = 188.8.131.52 a0.org.afilias-nst.info has AAAA address 2001:500:e::1 a2.org.afilias-nst.info has AAAA address 2001:500:40::1 b0.org.afilias-nst.org has AAAA address 2001:500:c::1 b2.org.afilias-nst.org has AAAA address 2001:500:48::1 c0.org.afilias-nst.info has AAAA address 2001:500:b::1 d0.org.afilias-nst.org has AAAA address 2001:500:f::1
The root zone contains all information on all the top level domains, as well as the special KSK (Key Signing Keys) keys that underpin the DNSSEC system. In domain names, the root zone is represented as a final '.' at the end of the domain which is typically implied although there are rare technical reasons where it has to be referred to directly. Now that we've discussed and slightly explored the root zone, let's talk about the organization that administrates it, and the policy and rules related to the root, and the top-level domains referenced within.
What is ICANN?
The full history of ICANN is too long to recap here, but in short, ICANN is a multi-stakeholder community that represents various stakeholder groups and their interests and needs. In no specific order, these groups are as follows:
- GAC - Government Advisory Committee
- SSAC - Security and Stability Advisory Committee
- RSSAC - Root Server System Advisory Committee
- GNSO - Generic Names Supporting Organization
- ccNSO - Country Code Names Supporting Organization
- CSG - Commercial Stakeholder Group
- ASO - Address Supporting Organization
- At-Large/ALAC - At Large Advisory Community (Internet Endusers)
That's a LOT of acronyms, groups, and organizations, and this isn't even a complete list. Each of these groups (known as stakeholders) are essentially cross-sections of all internet users and work to drive policy that meet the goals of their interests and charters. Other groups primarily act in an advisory role such as SSAC in evaluating impact of policy changes to the ICANN board. ICANN stakeholder groups create working groups (many of which are open to the public) to accomplish goals and draft policy, respond to public policy comments, and create a final report. These are then followed by implementation.
As you can plainly see, ICANN is a massive multi-headed hydra that at first is not the most user-friendly beast to approach. At least from my perspective, getting involved was rather difficult. For this purpose, ICANN offers two programs to help get people involved: a fellowship program to bring both those with diversity or unique skills in and the NextGen program. I can't speak on NextGen, but I can speak to the fellowship program, and my personal story in how I both got involved and the topics and work I was involved in at ICANN63.
ICANN Fellowship Program
To talk about my experience as a fellow, we need to go back to February 2018 when I was in San Juan, Puerto Rico, visiting with a friend. While I was there, I saw large banners with the ICANN name and logo and some sort of conference. While I did not know the specifics at the time, what I was seeing was the ICANN61 General Policy Forum. As such, I walked in off the street, registered for a badge, and sat down at a high level meeting regarding an issue known as name collision hosted by the SSAC. This, and a few other meetings convinced me that becoming involved with ICANN was something I was personally interested in.
Unfortunately, getting your foot in the door with ICANN from the outside is something of a tall order. To help solve this problem, ICANN offers a fellowship program to help bring both diversity and talent within the community. As such, I was selected to attend ICANN63 on the basis of my position as an independent freelancer combined with strong technical skills. The fellowship program, currently managed by Siranush Vardanyan, is meant to help bring people into the ICANN community and guide them into position and niches where their skillsets can help. Many within ICANN bring technical, legal, policy, and activism talents to the table, and it is an extremely inclusive community to say the least. As was oft-repeated, 'Once a fellow, always a fellow'. Through the fellowship program, I was assigned a coach, Alfredo Calderon who helped me get involved with the gTLD working groups, and help decode the maze that I described above.
The intent of the fellowship is to prepare those to attend a face-to-face meeting (in this case, ICANN63), and help the fellow become active within the ICANN community. In my case, I managed to hit the ground running as in the intervening months between ICANN61 and 63, I had gotten involved with the Internet Society, and several working groups within the IETF (albeit it on a semi-active basis). That combined with closely following the news allowed me to be productive from the start. What follows are issues that I was primarily involved with — it doesn't cover some of the larger discussions such as the GDPR/WHOIS policy development sessions.
New Generic Top Level Domains (gTLD)
Generic top level domains are generally the most common type of TLD most people encounter, comprising .com, .net, .org, etc. compared to the two letter country code TLDs (ccTLDs) such as .us or .io. Back in 2005, ICANN began developing policies to allow for the creation of new gTLDs, and in 2013, these new gTLDs began being added to the root zone as part of the New gTLD Program. Since the initial land rush and additions, ICANN has been developing new rules relating to this process in the form of the New gTLD Subsequent Procedures PDP (Policy Development Process) working group (known as the SubPro), which I'm a member of. I've primarily worked to ensure that not for profit and smaller communities can't be outbid or driven out of the process of obtaining their own gTLDs.
Expansion of the generic TLDs help relieve strain on the already crowded .com/.org/.net registries and pave the way for full internationalization of the internet (a topic I'll cover below). While there have those who've felt that expanding gTLDs was a mistake, the ability to have domains such as .nyc for sites relating to New York City has shown that the new gTLD program has real world benefits that we're already experiencing today. However, creating and expanding gTLDs also has opened a paradox's box of sorts which involves the SubPro, specifically in the the realm of string contention and name collisions.
String Contention and Name Collisions
In a perfect world, everyone would have one unique name and registering a new gTLD would be an easy and straightforward process. Unfortunately, we don't live in that sort of world; we live in a world where the Government of Brazil, and Amazon both want the .amazon TLD. This is what's known as a string contention; when multiple parties want the same domain string, and part of my work within the SubPro is building and designing mechanisms for handling contentions, as well as a last resort process which is fair for all parties. In the last round of gTLD additions, many string contentions were solved either by private party, or through a last resort auction process. At the direction of the ICANN board, the SubPro has been reviewing the results of this last round, ensuring that all actors, especially smaller community-based ones have an equal chance of being given a gTLD, and making sure no one can be strong-armed out of the process. I (and others within the SubPro) have been working on creating and streamlining the new gTLD process, and making sure that no single party can monopolize a string by simply outspending everyone. Of course, social issues aren't the only hangup when creating a new top-level domain; you can have a name collision.
Name collisions are a closely related problem dealing with the technical issues of what happens when you add a name to the root zone that's already in use in other contexts. For example, the Tor network could be entirely shafted if .onion was added to the root zone as it's used as a pseudo-TLD. Unfortunately, because of literally decades of bad practice, poor device coding, and similar historical artifacts, it means that the root zones get thousands of requests per second for bogus top level domains. As part of adding any new TLD, a review is done to determine the technical impact — research by SSAC into the name collision issue as a whole is ongoing. While I'm not personally involved in this work as of yet, I am interested in joining it in the near future
Internationalized Domain Names (IDN)
Last, but not least, the final major activity I worked on was discussions related to the internationalization of domain names, and email address internationalization (EAI) with the goal of making ensuring the web is available for everyone. Due to the fact that DNS was designed in 1987, it was never designed with internationalization in mind and has required some arcane hacks to make it work. Let's take the string тест which is Russian for test; it displays properly because we support UTF-8. However, DNS was never designed to work with 8-bit characters. Instead, a system was created known as punycode. This system represents unicode in ASCII in a method that DNS can handle; so the domain тест.example becomes xn--e1aybc.example which can be handled by existing tools.
This however creates a disconnect between the displayed name (known as the U-label) and the ASCII representation (A-label) of a domain name, which is known to break software that either renders domain names, or in the cases of email, must amend information to its log files. It also leads to issues with SSL certificates, and other confusion within the ecosystem due to poor support. While IDNs have been around for awhile, new codepoints including right-to-left ones are being added that require more testing and development. I've started one of two projects to help study and test IDNs, and an active participant of the Universal Acceptance mailing lists on the subject.
dnscatcher and idn-root-zone
As part of the meetings and other work, I've started work on two projects to help raise awareness and study ongoing problems with the world of DNS by creating tools to help monitor the health of the ecosystem as a whole. The first of these is a project that I'm tentatively calling DNS Catcher, and its intent is to study the perspective of the domain name system from the viewpoint of the end user.
As we know from study from data related to authoritative name servers, and the root zone, a lot of recursive revolvers and end-user devices send bogus data, such as catching all missing domains with a wildcard, or sending bogus requests to the root. DNS Catcher is an attempt to quantify the problem from the last mile and understand what data devices are sending out. While it's still in very early proof of concept, the catcher's end goal is to compare known good authorize zone data to data collected from various locations such as public access points and more so as to identify bad actors within the DNS community. It's still in the early pre-alpha stage, but my initial coding efforts have left me optimistic I may have an alpha version ready to go by the end of the year which will be subject to its own blog post.
The other is what I'm calling tentatively calling Root Zone in a Box, a series of shell scripts, instructions and docker containers to automatically recreate a simulation of the DNS root zone, and other core internet functionality to allow testing of potential changes to DNS, as well as help study and debug various issues related to Internationalized Domain Names. Compared to dnscatacher, I've gotten further on this project as it's somewhat higher priority. While likely not of interest to most as of yet, RZiaB is basically designed to help validate and ensure that internationalized domain names and email address internationalization works smoothly and that issues can be quickly identified and fixed using an easy-to-host environment that can be quickly set up.
I'll likely talk more of these projects in separate posts at later dates, but I invite people to comment and review my work.
Other Odds and Ends
As with any conference, there was various interesting conversations, discussions, and round tables that you really don't experience in a purely electronic environment. One of these (which was the direct inspiration for DNS catcher) was discussing why some devices send bogus data (in the form of random hex strings) to the internet root zone. I postulated that the answer was it was the one more-or-less sure fire way to know if you have anyone tampering with your DNS data such as captive portals, restrictive firewalls, or ISPs who don't like to return NXDomain.
Another big part are social dinners and gatherings. One personal highlight is I also had a fairly decent conversation with the appointed representative to the GAC from the Holy See, dealing with domain name issues relating to the Vatican. We primarily talked about working at the Vatican, the papacy's interest in ICANN, and life within the city. As far as unique individuals go, this easily makes a spot on my top ten list!
Although my time in Barcelona has come to an end, my involvement within ICANN is higher than ever. We're doing strong work to try and keep the internet open and accessible to all, and we're always looking for anyone with an interest to get involved. The Fellowship experience helped me connect with individuals that let me reach my personal goals of working on the SubPro, as well as connected me to the IDN working group folks in a way that I hope to pave a new cornerstone of the internet for non-English speaking individuals. There's a lot of work ahead, but I can say with certainty that my work with ICANN will continue, and I look forward to what the future will hold. If you're interested in my projects, comment below, or follow me on Twitter: @fossfirefighter where I post about my work on DNS catcher, RZiaB, and other things that don't make SoylentNews... like a retroBBS hacking project.
If You Want To Get Involved
If what you've read interests you, and you want to get involved in ICANN yourself, a good starting place is the alac-announce mailing list which posts which working groups are in progress, have meetings, and other good information, as well as joining your regional At-Large community. Most working groups (WGs) don't require membership in a stakeholder group, so you can just dive in; you're simply expected to familiarize yourself with the WG's previous history up to that point for the most part. There is also a set of learning resources at learn.icann.org, and I'm happy to take questions here or on Twitter.
Before signing off, I want to personally thank several individuals who helped me get here. First, Alfredo Calderon, my ICANN coach and Siranush Vardanyan, manager of the fellowship program. Both were very understanding and helpful with some personal difficulties I ran into during the fellowship program and both of them contributed greatly to a successful face-to-face meeting. Next, I'd like to thank Martin Pablo Silva, who continuously encouraged me to apply for the Fellowship, and helped make sure my application was in tip-top shape, and last, but not least, Dina Solveig Jalkanen (who prefers to go by Thomas), who introduced me to ICANN and is a close personal friend and who is was instrumental in making this possible.
Until the next time ...
73 de NCommander