[HN Gopher] The Internet runs on free and open source software a...
       ___________________________________________________________________
        
       The Internet runs on free and open source software and so does the
       DNS
        
       Report: _The Domain Name System Runs on Free and Open Source
       Software (FOSS) [pdf]_ https://itp.cdn.icann.org/en/files/security-
       and-stability-ad...
        
       Author : ChrisArchitect
       Score  : 256 points
       Date   : 2025-10-29 18:16 UTC (1 days ago)
        
 (HTM) web link (www.icann.org)
 (TXT) w3m dump (www.icann.org)
        
       | jongjong wrote:
       | But the infrastructure is highly centralized and only certain
       | chosen entities can operate gTLDs and certificate authorities.
       | It's extremely misleading to call it 'free software'. Why can't
       | there be multiple competing systems. There should be a push for
       | Blockchain-based alternatives. I still don't understand why
       | projects like Unstoppable Domains aren't getting more traction.
       | The idea of a domain name that you actually own is appealing.
        
         | lelandbatey wrote:
         | I think you may be applying a very weird definition of "free
         | software", even compared to the usual gratis vs libre axis.
         | 
         | Also, I really don't think controlling a domain name NFT in a
         | system that's mostly computers you neither own nor control
         | constitutes "more ownership" than the IRL law and contract
         | bound rental world we currently live in. Especially if all the
         | requirements and outcomes (payments for control resulting in
         | land grabs of valuable names) are the same as our current
         | system.
        
           | pessimizer wrote:
           | Maybe not more ownership by the owner, but certainly more
           | ownership by the _user,_ which is what 's important. If there
           | are multiple blockchain-based alternatives, I can choose
           | which one to resolve with; it's also essentially built-in
           | namespacing (with each name marked with its resolver.) And
           | although I'm personally very crypto-negative, a distributed
           | ledger is exactly what I would want to make sure that any
           | nodes that I use to resolve a name on a particular registry
           | are trustworthy.
           | 
           | The throughput problem that poisons cryptocurrency becomes
           | irrelevant when we're talking about something that's as
           | naturally long-lived as domain names. Every domain blockchain
           | can have its own gatekeeping process; one can sell names for
           | thousands of $ each, and another can give away thousands for
           | a $. They can require that domain owners have a camera
           | pointed at them personally for 24-hours a day or be revoked,
           | or they can hand out infinite names through a onion-routed
           | API.
        
         | denkmoon wrote:
         | There can be and are multiple competing systems. There are
         | alternative dns roots, opennic.org for example, and entirely
         | separate protocols like ipfs and i2p with their own methods of
         | mapping names to numbers.
         | 
         | You can go make your entire own alternative DNS system, with
         | your own governance and policy. Free as you like. You just have
         | to convince people to resolve against you.
        
         | zzo38computer wrote:
         | I think that permanent identifiers might be helpful, which
         | could include a timestamp and other stuff. This can be combined
         | with web of trust, that it can then be verified in this way.
        
         | renewiltord wrote:
         | They just run one root. You can run a different root. E.g. Some
         | people run an ENS bridge.
        
       | charcircuit wrote:
       | >In the cloud, hyperscale computing platforms such as Microsoft
       | Azure, Google Cloud, and Amazon Web Services all operate
       | significant resolver infrastructure to support their services. At
       | least four of the biggest hyperscalers rely on FOSS for DNS
       | resolving, while others have built proprietary solutions based on
       | FOSS DNS libraries.
       | 
       | This is surprising. I would have expected them to have custom
       | needs with so many customers that using an off the shelf service
       | would be sufficient.
        
         | lofties wrote:
         | It's all cURL.
        
           | giantrobot wrote:
           | It's cURL, SQLite, and one weird awk script no one
           | understands but works so no one touches it.
           | 
           | https://xkcd.com/2347/
        
       | mumber_typhoon wrote:
       | Imagine a scenario where you want to start gardening. Go to
       | gardening clubs and you'll find a lot of free information there
       | and people to guide you. Public libraries exist if you want to
       | join a book club and start reading. Again free. Agriculture,
       | irrigation, building homes, woodworking, stitching clothes, etc.
       | everything essential has been free to learn and do.
       | 
       | Apply this to the internet and essentials are FOSS. Linux, DNS
       | and maybe RISCV someday will mean you can build computers and
       | internet on essentials that are free to learn and use.
        
         | staplers wrote:
         | Raspberry Pi's obviously trying to make this a reality.
         | 
         | Learning to self-host and get off cloud services might be one
         | of the most personally freeing feelings I've had in a long
         | time.
         | 
         | Rent-seeking is obviously growing out of control and one of the
         | most powerful ways to combat it is personal ownership (if
         | possible).
        
         | cobertos wrote:
         | In the same analogy, doesn't that mean that vendor-locked
         | software like iOS or ChromeOS would be akin to vendor-locked
         | seeds from Monsanto?
        
           | mrarjen wrote:
           | Bayer these days and yes, avoid like the plague for nothing
           | good will come of it.
        
         | throw844958585i wrote:
         | Land for garden in my town costs like 1000 eur per square
         | meter. Gardening clubs are full of old dudes, who want to have
         | a sex with me. Public libraries are homeless shelters now. If I
         | actually plant some vegetables (in pots or front yard), it will
         | be full of dog/cat excrements the next day!
         | 
         | You are living in imaginary land, nothing is free in todays
         | society!
        
           | 1718627440 wrote:
           | Just because you live in some shitty, doesn't mean everyone
           | has to or wants to be...
           | 
           | (I do realise you used euro, I just don't think we need to
           | adjust our standards down, when our locality sucks.)
        
       | seydor wrote:
       | We should tax cloudflare, aws etc. for using public
       | infrastructure
        
         | renewiltord wrote:
         | We do tax them, my dude.
        
           | tellarin wrote:
           | Just maybe not enough.
        
         | LeonM wrote:
         | I don't understand your sentiment against Cloudflare here.
         | 
         | Cloudflare also delivers a rather large portion of said public
         | infrastructure free of charge. They also released a few of
         | their own projects as FOSS, and regularly contribute.
         | 
         | Granted, the centralisation part worries me too, but it feels
         | like a bit of a cheap shot against CF just because they are a
         | large player.
        
       | neom wrote:
       | And don't forget about this awesome report, Roads and Bridges:
       | The Unseen Labor Behind Our Digital Infrastructure - by Nadia
       | Asparouhova.
       | https://www.fordfoundation.org/learning/library/research-rep... -
       | great research into what goes into OSS.
        
         | matheusmoreira wrote:
         | Too often unpaid too. Certainly taken for granted and exploited
         | by the corporations. We should have more copyleft than
         | permissive software.
         | 
         | https://zedshaw.com/blog/2022-02-05-the-beggar-barons/
        
       | evertedsphere wrote:
       | Perhaps it doesn't even matter anymore, but I'm not yet past the
       | point where it's disheartening every time I click on a link and
       | it's clear that it came out of an LLM. Hopefully this doesn't
       | extend to the actual report.
        
       | hliyan wrote:
       | It's interesting that there is a generation of developers now who
       | seem to believe that the Internet is an achievement of pure
       | commercial, market dynamics and are surprised to learn about
       | ARPANET and its early development within academia (we were taught
       | this history in the first year of university). If the foundations
       | of the Internet (particularly the protocol suites) had not been
       | open, government-funded and not-for-profit, we would probably
       | have a number of competing closed platforms instead of a single
       | Internet, with paid services to perform protocol translations
       | between them.
        
         | n3storm wrote:
         | So true. I find another prove that altruist collaboration wins
         | any other model although users may not perceived as such or
         | there is no interest spreading these facts.
        
           | cousin_it wrote:
           | Altruist? DARPA is a military agency, ARPANET was a prototype
           | network designed to survive a nuclear strike. I think the
           | grandparent comment's point is that the innovation was
           | government-funded and made available openly; none of which
           | depends on the slightest on its being altruist.
        
             | tremon wrote:
             | The resilience of ARPANET was influenced by CYCLADES, which
             | was developed in French Academia:
             | https://en.wikipedia.org/wiki/CYCLADES
             | 
             | > The CYCLADES network was the first to make the hosts
             | responsible for the reliable delivery of data, rather than
             | this being a centralized service of the network itself.
             | Datagrams were exchanged on the network using transport
             | protocols that do not guarantee reliable delivery, but only
             | attempt best-effort [..] The experience with these concepts
             | led to the design of key features of the Internet Protocol
             | in the ARPANET project
             | 
             | Keeping with the theme of the thread, CYCLADES was
             | destroyed because of greed:
             | 
             | > Data transmission was a state monopoly in France at the
             | time, and IRIA needed a special dispensation to run the
             | CYCLADES network. The PTT did not agree to funding by the
             | government of a competitor to their Transpac network, and
             | insisted that the permission and funding be rescinded. By
             | 1981, Cyclades was forced to shut down.
        
             | msla wrote:
             | https://siliconfolklore.com/internet-history/
             | 
             | > Rumors had persisted for years that the ARPANET had been
             | built to protect national security in the face of a nuclear
             | attack. It was a myth that had gone unchallenged long
             | enough to become widely accepted as fact.
             | 
             | No, the Internet (inclusive of ARPANET, NSFNet, and so on)
             | was not designed to survive a nuclear war. It's the worst
             | kind of myth: One you can cite legitimate sources for,
             | because it's been repeated long enough even semi-experts
             | believe it.
             | 
             | The ARPANET was made to help researchers and to justify the
             | cost of a mainframe computer:
             | 
             | > It's understandable how it could spread. Military
             | communications during Nuclear War makes a more memorable
             | story than designing a way to remote access what would
             | become the first massively parallel computer, the ILLIAC
             | IV. The funding and motivation for building ARPANET was
             | partially to get this computer, once built, to be "online"
             | in order to justify the cost of building it. This way more
             | scientists could use the expensive machine.
        
               | cousin_it wrote:
               | That's a valiant attempt at myth-fighting, but it doesn't
               | fully convince me. For example, one hop to Wikipedia
               | gives this:
               | 
               | > _Later, in the 1970s, ARPA did emphasize the goal of
               | "command and control". According to Stephen J. Lukasik,
               | who was deputy director (1967-1970) and Director of DARPA
               | (1970-1975):_
               | 
               | > _" The goal was to exploit new computer technologies to
               | meet the needs of military command and control against
               | nuclear threats, achieve survivable control of US nuclear
               | forces, and improve military tactical and management
               | decision making."_
        
             | fellowniusmonk wrote:
             | Any technology that is developed to be federated and
             | resilient in the face of apocalyptic events is
             | definitionaly altruistic toward humanity.
        
         | pmontra wrote:
         | We would have had telcos (or satellite TVs) creating and
         | competing on siloed services and eventually some big company
         | would buy a telco to have access to their customers. Some
         | multinational telcos would provide the same services around the
         | world but most countries would have a wall at the border.
        
           | foobarian wrote:
           | The thing is, we did have telcos and others provide siloed
           | services. AOL being the big one, but also things like
           | Prodigy, Compuserve, French Minitel, etc. What I wonder about
           | is, would an Internet-like system have arisen anyway just
           | because it's such a better ux? You saw a development like
           | that with Fidonet for example but I think the Internet stole
           | its oxygen.
        
             | pmontra wrote:
             | We did have that well into Internet years because some
             | services where not there yet. I give you an example: I
             | worked for one of the early 3G providers (2003-). YouTube
             | is from 2005. We developed and run streaming video services
             | because our customers did not have any meaningful source of
             | videos to watch, so we had to build that and other services
             | to drive the sales of our phone contracts. Mobile web sites
             | were not a thing for a few more years so we had news and
             | meteo and many other things. Our little private internet.
             | Then YouTube became big, other services were born and
             | became big too, all phone companies basically became
             | internet providers.
             | 
             | Without IP protocols we probably would not even be born as
             | a company. BBSes on analog modems or GSM data would be all
             | we had.
        
             | dredmorbius wrote:
             | Minitel was offered by France Telecom, then a government-
             | operated service, under what was originally the Ministry of
             | Postes, Telegraphes et Telephones. (The name "Minitel" is
             | _not_ short for  "ministry of telecommunications", but
             | rather _Medium interactif par numerisation d 'information
             | telephonique_. Though I suspect some advantageous ambiguity
             | was exploited.
             | 
             | France Telecom was privatised during the 1990s with the
             | aquisition of the Orange Group, its current (commercial)
             | branding.
        
         | no_wizard wrote:
         | Even among the open protocols there was healthy competition
         | like Gopher and BBS networks
         | 
         | It was truly a time to be alive, the 1980s and 1990s, if you
         | were into technology. The change was interesting and quite
         | rapid, lot of variety and experiments
        
         | giantrobot wrote:
         | > we would probably have a number of competing closed platforms
         | instead of a single Internet, with paid services to perform
         | protocol translations between them.
         | 
         | That _did_ exist with the likes of Tymnet and the various
         | Online Services (AOL, CompuServe, etc). The Internet won out
         | over those because it was open, as you alluded to. Internet
         | adoption really exploded with unlimited services rather than
         | ISPs that billed hourly.
        
         | mhurron wrote:
         | Not really surprising when you have political parties and large
         | tech companies both either outright minimizing the role public
         | institutions had in developing anything or outright lying that
         | the public sector can't possibly do anything at all.
        
         | Sophira wrote:
         | Yep, services like AOL, CompuServe, Prodigy and so on, would
         | absolutely be separate. As it was, each of these services
         | basically had to allow access to the greater Internet.
        
         | fellowniusmonk wrote:
         | Open source and hacker culture is basically a massive ethical
         | triumph that has catapulted humanity forward.
         | 
         | There isn't a single person living on the planet that isn't
         | touched and benefitted by this in some way, even remote island
         | tribes we consider untouched, there is hardly a single ounce of
         | space payload that doesn't have open source in it's causal
         | chain.
         | 
         | There is no more pragmatically altruistic culture that has ever
         | existed or impacted more people.
         | 
         | In the U.S. no culture has helped lift the impoverished out of
         | poverty into the upper middle class through empowerment (ask
         | any of us adults who were once kids going to bed hungry.)
         | 
         | People have their pet beliefs and metaphysics and ideological
         | communities but "information deserves to be free" is the goat.
        
           | AnimalMuppet wrote:
           | > There isn't a single person living on the planet that isn't
           | touched and benefitted by this in some way, even remote
           | island tribes we consider untouched...
           | 
           | That seems a bit over the top. I'd love to see your causal
           | chain for that claim.
        
             | fellowniusmonk wrote:
             | The Sentinelese monitoring program run by the government in
             | India the works to enforce no contact rules relies on open
             | source in it's causal chain.
             | 
             | My mother did remote tribal work in the brasilian interior,
             | she grew up on the Amazon, I personally know well off Ruby
             | devs who grew up unable to even read.
             | 
             | Whether you are against intervention or pro intervention,
             | all across that spectrum, whatever your luxury belief may
             | be, everyone is impacted by open-source.
             | 
             | Open source is an insane force multiplier, not just for
             | production but for the entire training and r&d pipeline as
             | well.
        
           | _DeadFred_ wrote:
           | Thanks for pointing this out. We're so surrounded by the
           | dystopian tech future, it's nice to see some positivity to
           | tech, it's been a while.
        
         | Dracophoenix wrote:
         | The protocols were made open by necessity, not by design. The
         | motive was to connect academic , government, and commercial
         | institutions across the country, all of which operated on
         | incompatible operating systems and data networks. However, the
         | common man would not have benefited from this before 1993, as
         | the government effectively operated as a semi-competent
         | firewall against commercial content and the broader public.
         | They even sued ISPs that permitted legitimate accounts from
         | remotely accessing the net through PPP or SLIP protocols. Not
         | even commercial news feeds were permitted until the late 80s.
         | 
         | The only Internet the common man interacted with is the one
         | that began to flourish as the government relinquished control.
         | The Internet since the mid-90s is and has been a purely
         | commercial achievement.
        
           | icedchai wrote:
           | There were some early ISPs, like The World (Boston), that had
           | IP access around 1991 or so. I believe they were connected
           | through UUNet. I don't know if they were routeable on the
           | NSFNet? "Commercial" traffic was supposedly prohibited.
        
           | _DeadFred_ wrote:
           | I remember a few BBSs that offered email (done via nightly
           | UUCP so slower than people today are used to) to the common
           | person.
           | 
           | Because it was new there weren't really thought out mechanism
           | for people connecting early on, but some did.
           | 
           | http://www.armory.com/~spcecdt/deepthought.html
        
             | icedchai wrote:
             | I had a UUCP feed, by either late 1992 or early '93.
             | Several other local BBSes had one, as well.
        
       | LeonM wrote:
       | For what I remember, most of the DNS root servers used to run
       | Bind9 exclusively. I'm glad to see that this is now more diverse
       | with NSD and Knot also being used (see table 4 in the report).
       | 
       | Nothing against Bind9, but it is almost exclusively maintained by
       | the ISC, so the DNS's future used to depend heavily on the ISC
       | getting the funding needed to continue operating.
        
         | embedding-shape wrote:
         | Not to mention how much better it is for
         | standards/protocols/standarization to have multiple
         | implementations of the same protocol in real-world usage so we
         | can nail down the protocol in all situations. Bind9 almost
         | ended up being "DNS" itself, which would have been bad overall,
         | and instead we're seeing more diversity which means we'll be
         | able to make better use of the specifications.
        
         | akvadrako wrote:
         | That hasn't been true for at least 15 years. I was a k-root DNS
         | operator then, and we ran several software stacks on each
         | cluster in case one had a bug.
        
       | whitehexagon wrote:
       | I was wondering what it would take to fork the internet. Or more
       | seriously, what it would take to create an open source / hardware
       | alternative. Maybe something true to it's routes of academia, ie
       | no commercial activity permitted, especially adverts! Maybe even
       | limited to some of the original simple protocols like HTTP, FTP
       | and maybe RSS?
       | 
       | When I was in Holland many years ago there was a university
       | initiative to connect the whole town with it's own hobby network.
       | I bought some directed antennae to connect, and it worked, but I
       | seem to recall it was line-of-sight WiFi stuff and thus very
       | limited range.
       | 
       | I also worked for an airline that had line-of-sight laser
       | communications, but again it was only short distances.
       | 
       | Are there any modern alternatives that would make such a people
       | sponsored/focused hobby network possible?
       | 
       | The first thought I had was long-wave radio, but I cant imagine
       | there is much bandwidth there, but might be okay for a text only
       | protocols?
        
       ___________________________________________________________________
       (page generated 2025-10-30 23:01 UTC)