4 ms·

ICANN has failed us, Nominet has failed us, Public Interest Registry has failed us. We definitely need some better system where a single entity going rogue can'

ICANN has failed us, Nominet has failed us, Public Interest Registry has failed us. We definitely need some better system where a single entity going rogue can't affect millions of domain owners. I can see two approaches. One is where we make it technically impossible for DNS operators higher in the tree to mess with us. That's what the GNU Name System proposes with the "hyper hyper local root", a petname system which supports delegation, to defeat censorship. Another one is a domain name cooperative like CCOR.org attempted to build. So the organization serving our TLD is in fact operating transparently in the interests of its members paying administrative fees (think <1$/domain/year)... not a business selling trying to make profits. In my view we need both! Replace the corrupt pro-business ICANN with a democratic non-profit foundation, and at the same time make it technically harder for bad actors to mess with everyone. Having a (domain) name on the network should be a free and accessible public service for all. They're half a dozen lines of text (counting glue records and DNSSEC anchors) in a database, it's OUTRAGEOUS we have to pay for that at all. In the early nineties, it used to cost money to host just a few kilobytes of information, and that's why we introduced administrative fees for domain names. Why are we still paying >10$/year for less than a kilobyte of information hosting just so that shady businesses can make a profit?

johnchristopher 6y ago
But how would that work in practice ? Like letsencrypt ?: what's needed to incite enough people to move on so that it becomes feasible and a reality (while keeping the old system alive for legacy reasons and for the time it takes for thing to settle down) ?
- southerntofu 6y ago
  If we're talking about GNS to replace the DNS protocol, it's in fact backwards-compatible and intends to recognize ICANN as root zone, as was the case previously. If we're talking about a domain name cooperative, that would act as a registrar and potentially as a TLD in the future, all we need is a non-profit organization, some bylaws, and enough good-will people to make it work. Maybe OpenNIC would be it?
CallingOutBS 6y ago
> One is where we make it technically impossible for DNS operators higher in the tree to mess with us. That's what the GNU Name System proposes with the "hyper hyper local root", a petname system which supports delegation, to defeat censorship. DNS and our existing registries are too embedded in too many critical systems, businesses and workflows for it to be replaced by a niche unproven system, over the relatively small cost of domains. Improving the governance of existing structures is feasible and useful.
- selfhoster11 6y ago
  Those new systems retrofit into the existing DNS system very easily. There is no excuse not to try new approaches.
robalfonso 6y ago
Quote>>> They're half a dozen lines of text (counting glue records and DNSSEC anchors) in a database, it's OUTRAGEOUS we have to pay for that at all. This is not what you are paying for. Every time this argument comes up it hides the real problem...what you are paying for is the legal wrangling, security overhead and general compliance that comes with the internet. Phishing, trademarks, spam, business disputes,and valid legal requests will all still exist no matter who runs things. Those costs are what your domain fees really have to cover. It’s been a very long time since domain costs were highly dependent on the technical requirements of the data they represent.
- southerntofu 6y ago
  > Phishing, trademarks, spam, business disputes,and valid legal requests will all still exist no matter who runs things. Of this list, only "valid legal requests" is something a TLD should be concerned with in my opinion. And even that is questionable, given that law is often unjust, favoring the powerful and abusing the powerless. We already have strong regulations holding internet hosts accountable for what they're hosting, and many countries (like France) have censorship at the DNS level. Isn't that more than enough to deal with abuse? > It’s been a very long time since domain costs were highly dependent on the technical requirements of the data they represent. Yes, and that's the problem. When you pay 1$/month for web hosting, you're not paying a thousand times the price because of hypothetical legal concerns. Why would it be different with DNS? Why is it normal that a line in a zone file on some TLDs costs more than an actual VM in a datacenter?!
  - robalfonso 6y ago
    TLD's have various compliance requirements around abuse, which does include the afore mentioned. There is significant volume related to this compliance as well as ensuring registrars maintain compliance. It's the nature of the existing arrangement so it's not hypothetical. Of course the individual cost is low for a single item, however, businesses have never operated on that principle. They aggregate the profit and the risk across a pool, no surprise here.
marcinzm 6y ago
>a democratic non-profit foundation Isn't that what Nominet basically was and here we are? >Why are we still paying >10$/year for less than a kilobyte of information hosting just so that shady businesses can make a profit? To provide a minimum barrier to people simply writing a script and registering literally every domain in existence. Either to resell or troll. Probably also to provide a financial paper trail in case of abuse, illegal activity or other reasons.
- doublerabbit 6y ago
  > To provide a minimum barrier to people simply writing a script and registering literally every domain in existence. Either to resell or troll. Probably also to provide a financial paper trail in case of abuse, illegal activity or other reasons. But that happens anyway, what has it discouraged?
  - marcinzm 6y ago
    >But that happens anyway, what has it discouraged? Just because something isn't great doesn't mean it can't become worse or that the current solution isn't making it better than it could be.
    - southerntofu 6y ago
      Sure, it might get worse. But it may as well get better. Implying that changing paradigm can only make things worse is a fallacy.
      - marcinzm 6y ago
        > Implying that changing paradigm can only make things worse is a fallacy. So is saying that changing the paradigm will make things better or stay the same. I'm not making a data driven argument for not changing things, I'm simply pointing out that you're not making one either.
        6y ago
dooglius 6y ago
Tor's Onion Services are pretty much ideal for this, and provides additional anti-censorship guarantees by providing anonymization for both parties.
- southerntofu 6y ago
  Onion names are a very good global crypto-secure naming scheme, but they don't solve the human-meaningful part of Zooko's triangle. GNS has some similar design properties, but "solves" the triangle by tying two different naming systems: petnames for locally-meaningful/secure names, root names for global/meaningful names. Also as good as tor's design is, it requires some level of centralization with hidden service directories, which GNS doesn't. Disclaimer: i'm a happy tor user for many years, criticism here is in good faith (posting this from the tor network).
tzs 6y ago
If I could I would give you 50 up votes just because you wrote 5 paragraphs about what you see as problems with the current system without suggesting that it be replaced with some kind of blockchain.
- southerntofu 6y ago
  We probably need a blockchain-based commenting system to make sure you can afford 50 upvotes as long as you have enough miners running ;-) Who cares about sybil attacks when you can run a proper ponzi scheme and get rich! /s

johnchristopher 6y ago

But how would that work in practice ? Like letsencrypt ?: what's needed to incite enough people to move on so that it becomes feasible and a reality (while keeping the old system alive for legacy reasons and for the time it takes for thing to settle down) ?

southerntofu 6y ago
If we're talking about GNS to replace the DNS protocol, it's in fact backwards-compatible and intends to recognize ICANN as root zone, as was the case previously. If we're talking about a domain name cooperative, that would act as a registrar and potentially as a TLD in the future, all we need is a non-profit organization, some bylaws, and enough good-will people to make it work. Maybe OpenNIC would be it?

CallingOutBS 6y ago

> One is where we make it technically impossible for DNS operators higher in the tree to mess with us. That's what the GNU Name System proposes with the "hyper hyper local root", a petname system which supports delegation, to defeat censorship. DNS and our existing registries are too embedded in too many critical systems, businesses and workflows for it to be replaced by a niche unproven system, over the relatively small cost of domains. Improving the governance of existing structures is feasible and useful.

selfhoster11 6y ago
Those new systems retrofit into the existing DNS system very easily. There is no excuse not to try new approaches.

robalfonso 6y ago

Quote>>> They're half a dozen lines of text (counting glue records and DNSSEC anchors) in a database, it's OUTRAGEOUS we have to pay for that at all. This is not what you are paying for. Every time this argument comes up it hides the real problem...what you are paying for is the legal wrangling, security overhead and general compliance that comes with the internet. Phishing, trademarks, spam, business disputes,and valid legal requests will all still exist no matter who runs things. Those costs are what your domain fees really have to cover. It’s been a very long time since domain costs were highly dependent on the technical requirements of the data they represent.

southerntofu 6y ago
> Phishing, trademarks, spam, business disputes,and valid legal requests will all still exist no matter who runs things. Of this list, only "valid legal requests" is something a TLD should be concerned with in my opinion. And even that is questionable, given that law is often unjust, favoring the powerful and abusing the powerless. We already have strong regulations holding internet hosts accountable for what they're hosting, and many countries (like France) have censorship at the DNS level. Isn't that more than enough to deal with abuse? > It’s been a very long time since domain costs were highly dependent on the technical requirements of the data they represent. Yes, and that's the problem. When you pay 1$/month for web hosting, you're not paying a thousand times the price because of hypothetical legal concerns. Why would it be different with DNS? Why is it normal that a line in a zone file on some TLDs costs more than an actual VM in a datacenter?!
- robalfonso 6y ago
  TLD's have various compliance requirements around abuse, which does include the afore mentioned. There is significant volume related to this compliance as well as ensuring registrars maintain compliance. It's the nature of the existing arrangement so it's not hypothetical. Of course the individual cost is low for a single item, however, businesses have never operated on that principle. They aggregate the profit and the risk across a pool, no surprise here.

marcinzm 6y ago

>a democratic non-profit foundation Isn't that what Nominet basically was and here we are? >Why are we still paying >10$/year for less than a kilobyte of information hosting just so that shady businesses can make a profit? To provide a minimum barrier to people simply writing a script and registering literally every domain in existence. Either to resell or troll. Probably also to provide a financial paper trail in case of abuse, illegal activity or other reasons.

doublerabbit 6y ago
> To provide a minimum barrier to people simply writing a script and registering literally every domain in existence. Either to resell or troll. Probably also to provide a financial paper trail in case of abuse, illegal activity or other reasons. But that happens anyway, what has it discouraged?
- marcinzm 6y ago
  >But that happens anyway, what has it discouraged? Just because something isn't great doesn't mean it can't become worse or that the current solution isn't making it better than it could be.
  - southerntofu 6y ago
    Sure, it might get worse. But it may as well get better. Implying that changing paradigm can only make things worse is a fallacy.
    - marcinzm 6y ago
      > Implying that changing paradigm can only make things worse is a fallacy. So is saying that changing the paradigm will make things better or stay the same. I'm not making a data driven argument for not changing things, I'm simply pointing out that you're not making one either.
      - 6y ago

dooglius 6y ago

Tor's Onion Services are pretty much ideal for this, and provides additional anti-censorship guarantees by providing anonymization for both parties.

southerntofu 6y ago
Onion names are a very good global crypto-secure naming scheme, but they don't solve the human-meaningful part of Zooko's triangle. GNS has some similar design properties, but "solves" the triangle by tying two different naming systems: petnames for locally-meaningful/secure names, root names for global/meaningful names. Also as good as tor's design is, it requires some level of centralization with hidden service directories, which GNS doesn't. Disclaimer: i'm a happy tor user for many years, criticism here is in good faith (posting this from the tor network).

tzs 6y ago

If I could I would give you 50 up votes just because you wrote 5 paragraphs about what you see as problems with the current system without suggesting that it be replaced with some kind of blockchain.

southerntofu 6y ago
We probably need a blockchain-based commenting system to make sure you can afford 50 upvotes as long as you have enough miners running ;-) Who cares about sybil attacks when you can run a proper ponzi scheme and get rich! /s