RFC - Decentralised Naming System

What happens if someone makes a program that generates and registers all the possible DNS names in existence. Does that make them the Supreme Overlord of All SAFE DNS?

I know it is impossible to generate all the DNS names in existence, but it is within the realm of probability for someone to generate the most common names and hold them hostage, putting the SAFE Network in the same crappy centralized DNS predicament the current internet is in.

How do we solve this issue without taking away users freedom?

Network confidence ratings? If the system is able to evict people from a domain based on popular action, you could have a situation where Facebook has been gazumped, but everyone wants SAFE book, and no one wants ā€œIs this your domain? Give me money!ā€

An option to flag a DNS entry as bad would allow you to choose a ratio of visits-to-flags. Obviously this moves the problem to a question of trusting reports, but it provides a decentralised mechanism to avoid domain squatting.

Hi, if @janitor excuses me for restating my views, I share this concern. The best answer I have found so far is that a network aimed to break a monopoly held by others shouldnā€™t build a monopoly of its own. Naming systems can be multiple.

Interoperation can be ensured via a two-tier addressing system where at the lower level unique binary addresses are used (aka ā€œcomplexā€, onion-style) - and resolution of these addresses to routing info is done by MaidSafe core network. At the higher level human readable addresses are mapped to binary ones. Each user may be using a different high level naming system of his choice. These high level naming systems are simply normal applications running on top of the core MadeSafe network.

There can still be one official high level naming system run by MadeSafe foundation itself. Names in this system can be sold for money. This naming system will come preconfigured on safe browser at installation time which hopefully will make it popular enough for organisations to pay money for registration.

Under this proposal each user would be able to construct his own human readable address space by combining different high level naming system. Each user would be expected to keep a list of private bookmarks mapping human readable names like ā€˜bankā€™ to the binary address his bank has given him on paper possibly via a q-code. The level of trust to all addresses resolved via naming systems would need to be carefully computed using ā€œweb of trustā€ techniques. Trust levels will be something a user would need to be acutely aware of.

Writers of all applications on MadeSafe network would need to be aware that different users may be using different naming systems. This means for example that all hyper links on all web pages would need to use the low level binary addresses. Reverse lookup can then be used to present a human-readable address of the new page to the user.

I agree that this all looks a bit awkward, but then we, future MadeSafe users, hope for MadeSafe to free us from the monopolies not build a new one. And if we really want it that badly we may need to put up with the associated increase in complexity of naming and learn to understand trust levels in distributed networks.

Guys youā€™re like those doomsday astronomers.

Firstly, you canā€™t just ā€œgenerateā€ a MaidSafe address and claim it forever. You need to actually have an actual node come up and stay online.

Secondly, thereā€™s no (as far as I know) a name resolution solution, so thereā€™s nothing to occupy yet.

Thirdly, letā€™s say thereā€™s an alias system one can bring up 100K nodes and register ā€œall the good namesā€, but it doesnā€™t prevent you from registering a crappy name. Itā€™s only a matter if you like (or donā€™t) First Come First Serve, but itā€™s not a matter of can you have a mycrappyalias54321.safe, Iā€™m sure youā€™ll be able to get something like that.

1 Like

A small note: isnā€™t

    mycrappyalias54321.safe

somewhat in the same league as

    duskgytldkxiuqc6.onion

both are hard to remember and require the use of local bookmarks? And local bookmarks can be organised in a tree of foldersā€¦ and if people start sharing them we get back exactly to where I started - to a multitude of naming systems.

Yes, you might end up with a name thatā€™s close to impossible to remember, but thatā€™s not ā€œmonopolyā€, you have an unlimited number of (crappy) names at your disposal.

The naming system would be one (*.safe), bookmarks/redirects/aliases would map many-to-1. But SAFE canā€™t control what people do on the client, this many-to-one aliasing can always be done.

Also please everyone note that my proposal to auction ā€œall the good namesā€ would ensure that the names hoarder would make the Foundation so rich that it could easily fund continuous development of the platform over a period of several years without raising money.
You want a crappy name? You can have it.
You want a nice name? Buy it (i.e. donate $50 to the Foundation)

Wouldnā€™t this namespace also end up a bit chaotic?
The existing DNS is hierarchical making it easier on humans.

I appreciate the need to raise funds. I might even secretely dream of working for the foundation :slight_smile: So I do want it to have money even if that dream (most likely) never comes true. So what Iā€™m trying to do here is to balance the need to raise funds against a design which in my view might deliver the best value to the humanity. Please see aboveā€¦

[quote=ā€œcatbert, post:23, topic:4235ā€]
There can still be one official high level naming system run by MadeSafe foundation itself. Names in this system can be sold for money. This naming system will come preconfigured on safe browser at installation time which hopefully will make it popular enough for organisations to pay money for registration.[/quote]

This was my best shot so far in trying to satisfy both needs. In this official namaspace the foundation can enforce whatever policy it deems sensible - for example remove names of fraudelent sites, perform dispute resolution between trademark holders etc Might running such a namespace actually end up being a better business mode?

Another angle: imagine a legal dispute about arises.

Under the current proposal: the Foundation sold name ABC to DEF but GHI wants for itself and sues the Foundation. What does the Foundation do? Does it even have the technical means to reverse the grant of name ABC to DEF? Or has it gone into Bitcoin style block chain and can not be undone? Does the Foundation need to pay the penalty to GHI?

Under catbert proposal: firstly there are multiple naming systems, most of them not run by the Foundation - so Foundation has no responsibility for what they do. If however Foundation has granted ABC to DEF in its own official namespace and GHI wins a court battle the foundation can easily give ABC to GHI because under this proposal the naming system is not something built into the core of the network but a simple app like a relational database running on top of it.

Also if the naming service is not something built into the core of the network but an app the foundation can spawn a separate legal entity managing the official namespace. This entity would pump proceeds back into the foundation but in case of a legal case it would shield the founcation bearing all the responsiblity itself.

ā€¦ and should this extra legal entity collapse of a judicial brutality that wonā€™t do that much harm to SAFE network. Okay the Network will loose its most popular naming service but then it will be very soon replaced. This organism should survive the death of any of its parts!

I donā€™t think I want the Foundation to take the role of ICANN, thatā€™d be a point of centralization and a thus a vulnerability. The Foundation members can be pressured by government agencies, their keys can be lost or stolen, etc. No SAFE functionality should be dependent on the Foundation, ideally it should only fulfill an advisory role for development and the community.

If I remember correctly the network itself has a reward scheme for developers who release SAFE software updates that actually get adopted. If that works, we donā€™t need the Foundation for dev funding.

Iā€™m not principally against payment for domain names, but then the SAFE network itself should be paid. Coins used in such payments would then be recycled (destroyed and later re-issued to farmers etc) by the network. This would make SAFE far more robust economically, because right now continued farming rewards rely solely on income from data uploads. Creating more sources of income for the network should be an important goal, and DNS is a good candidate.

4 Likes

This is an important point. The foundation should not control DNS for the same reason it should not control user names and passwords and Safecoin. The more autonomous the network is the better.

1 Like

It wouldnā€™t.

In my proposal it merely sells fancy aliases (in line with those proposals who want to add aliasing).
All native addresses (alphanumeric SAFE equivalent of biy0wp6edsf.onion) would be usable at all times and continue to work independently of the alias system.

The overlay system is what was proposed by Christopher in RFC - Decentralised Naming System), that is the part that Iā€™d put up for auction if people really need user friendly host names (personally I donā€™t see a need for it).

A decentralized organization is always a bit chaotic because there is no benevolent dictator who could impose order from above. (Example: the current maximum Bitcoin block chain size debate).

I explained this in another topic (also related to SAFE name resolution): there is no better value for the poor than when the rich subsidize SAFE Network for them (by buying/registering highly desirable name aliases).

I moved 194 posts to a new topic: RFC - Decentralised Naming System II - continuous auction (by Seneca)

Sorry about the tin foil hat rant, but letā€™s talk in the realm of reality. When the network guess live, I take the week off. I run my node and start registering names. I start with all three letter combinations ie. ABCThis is a classic example of the tragedy of the commons.

Add:

3 letter word = 17,576 possible combinations

After messing around with the DNS example in this weekā€™s update, I figure I can register one name every two seconds. If I take a 15 minute break every hour, that means I could register 1,350 names an hour. If I register names eight hours a day, it would only take me two days to register all names with a three letter combination.

What do you mean with the last sentence? This isnā€™t TOR, you donā€™t need to keep a node/server online to have your domain/website to persist.

[quote=ā€œjanitor, post:31, topic:4235, full:trueā€]
In my proposal it merely sells fancy aliases (in line with those proposals who want to add aliasing). All native addresses (alphanumeric SAFE equivalent of biy0wp6edsf.onion) would be usable at all times and continue to work independently of the alias system.

The overlay system is what was proposed by Christopher in RFC - Decentralised Naming System), that is the part that Iā€™d put up for auction if people really need user friendly host names (personally I donā€™t see a need for it).[/quote]

@janitor, I do understand.

So long as the biy0wp6edsf.onion is supported anything can be built on top. Good.

On the other hand we may need to be very careful. If you decree now that there is a single global namespace then this is what is going to get encoded in SafeBrowser and other software:

  • web pages will link to each other via aliases
  • web services will use aliases for endpiont information
  • aliases will be printed in newspaper ads

There seems to be a lot of debate on how to replace DNS.
First-come-first-served? Auctions? An LTD feeding money back to Foundation?

Wouldnā€™t it be wise to leave the door for future open? To test alternatives in real life?
Weā€™re software people, we know that nothing can replace testing.

True. I guess my point is you and me could use different resolution systems (chaos) but each of us could see a sensible hierarchical namespace (convenient for human brain).

And I totally agree. Do you think however that this could be achieved by running a nicro-ICANN as a separate LTD sponsoring the Foundation? And having this mincro-ICANN manage the most popular namespace - but not the only namespace in existance? The popularity in my view could come from the following sources:

  • pre-installed in SafeBrowser (like Google search in Firefox)
  • usefullness (the LTD would throw out fraud, resolve disputes)
  • we can ask users to make use of it to support the network

Also what about the legal argument? Wouldnā€™t it indeed be safter to manage the namespace via a separate organisation?

I mean for those alias names for actual safe addresses - the way Christopher proposed it once you create it, itā€™s mapped to your safe address, isnā€™t it?

I donā€™t know, I am not a distinguished thinker or DNS guru. I just know how much I donā€™t know and because of that I am a minimalist.

Very few people are concerned about (example) high quality install guides and how-toā€™s but those are the actual and important things at this moment.
But itā€™s boring and doesnā€™t lend itself to stimulating intellectual discussions, so we engage in these complicated discussions about an area none of us are competent about. :slight_smile:

The way I understood Chrisitian:

  • a symbolic address (hardtoremembername12345.safe) is mapped directly to routing info (in the form of a hashmap)

Catbertian proposal was:

  • a binary address derived from public key (biy0wp6edsf.onion) is resolved by SAFE to routing info natively
  • SafeBrowser uses an ordinary application running on top of SAFE to resolve symbolic addresses (wikipedia.official.safe) to binary ones (biy0wp6edsf.onion)
  • each user is free to choose which naming application to use (an application is identified with its .onion address) and to combine them

Itā€™s true that I havenā€™t even started looking at the codebase. If that disqualifies me from the discussion Iā€™m fine with that. However if not then my point is that deciding that a single global naming service will exists is an important decision which likely will not be possible to undo later. Iā€™d feel much SAFE-r if either

  • a decision was made that two-tier system will be used with multiple naming services
  • decision on naming service was postponed and only biy0wp6edsf.onion were used for the moment
1 Like

Sorry - I didnā€™t mean to imply youā€™re not competent in C, but I assumed none of us is really a DNS or name resolution expert.

crux of the matter - a major social and economic decision: if app writers start assuming presence of one global namespace now this may be impossible to reverse later