DIANNA: the IANA Decentralized design concept

[pent]

==UPDATED: 02 Mar, 2012

"You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete" Buckminster Fuller

"Never doubt that a small group of committed citizens can change the world." Margaret Mead

Quotes from http://p2pfoundation.net/

==

After inspecting a bunch of distributed DNS creation attempts (NameCoin, BitDNS, DNSchain) I wrote a new system design concept.

DIANNA - Decentralized Internet Assigned Names and Numbers Authority. Primary usage: Decentralized, authoritative, P2P DNS System.

This system is aimed to get rid of any central Internet Authorities (ICANN, IANA and others) which are vulnerable to pressure from some sorts of interested parties and groups. Yes, they are vulnerable to Internet Censorship attempts.

But the first close objective of this system is to provide authoritative DNS for such anonymous networks as I2P, Tor, Freenet.

DIANNA design is based on alternate Bitcoin block chain. DIANNA is not a fork of Bitcoin, but an extension.

==Design version 1.5 02 Mar, 2012

http://dianna-project.org/wiki/Design_Overview
v1.5 http://dianna-project.org/wiki/Design_Changelog

====

The authoritative domain reply lookup will be as follow:

For first, DNS client queries for particular domain and network returns a last domain transaction hash and block hash. Highest block wins - as always. Here client can verify that block hash is present in local headers chain and has a particular height.

For the second, client queries the network for Merkle Tree branch for needed domain transaction and transaction data itself. Here he can verify that transaction data are correct by reassembling Merkle Tree and comparing its root hash against local stored block header in chain.

Since client ensured that network returned *valid* *last* transaction for this domain, he can easily resolve domain into VALUE containing in transaction output.

Peace a cake

I need volunteers to code this tree of freedom. Primary, I need the project manager which will coordinate programmers. For the first steps I can be ideologist, project manager and programmer in one But I really need a help.

And please draw your attention: this is not Bitcoin fork, but an extension.

[1713559242]

1713559242

[1713559242]

1713559242

[1713559242]

1713559242

[1713559242]

1713559242

[ptshamrock]

WOWI! That looks cool ! Great Amount of Work u put in there !

[slush]

Dianna - DNS system on steroids :-)

[Raoul Duke]

Now this is something. It even has a marketable name and great features. I can see this taking off, honnestly.

Man, you may be the one to convince me to get into the mining business with this DNS system.

[Mushoz]

Really great idea! I love the fact you have something laid out, with arguments and ideas, rather than just a vague description of what you want to do. This could materialize really fast and become something awesome! This has potential Is it going to support merged mining by the way? May have missed it, but I don't think I could find anything about it on the wiki.

[pent]

Thank you.

Contributors wanted! C++ coders, donations also welcome! All donations will go to coders' work. I only coordinate this project.

[pent]

Dianna - DNS system on steroids :-)

 

[pent]

Is it going to support merged mining by the way? May have missed it, but I don't think I could find anything about it on the wiki.

I still din't get the purpose of merged mining. This looks like a cheat rather than something really useful. It is profitable only for one side (aux chain). I don't want to substitute DIANNA's chain to bitcoin or any other.

Also they say merged mining is full of bugs and possible double spent situations.

The main questions, why is this for? What for DIANNA should put its chain to risk? For possible popularity increase? The idea and design makes the popularity - not tricks.

If people like it, they will mine, they will use. All other - blasphemy

[DeathAndTaxes]

Is it going to support merged mining by the way? May have missed it, but I don't think I could find anything about it on the wiki.

I still din't get the purpose of merged mining. This looks like a cheat rather than something really useful. It is profitable only for one side (aux chain). I don't want to substitute DIANNA's chain to bitcoin or any other.

Also they say merged mining is full of bugs and possible double spent situations.

The main questions, why is this for? What for DIANNA should put its chain to risk? For possible popularity increase? The idea and design makes the popularity - not tricks.

If people like it, they will mine, they will use. All other - blasphemy

People liking it has nothing to do with mining at least not any significant hashing power.

Without merged mining if you chain is less profitable to miners than Bitcoin you are asking them out of the  goodness of their heart to lose money helping you out.  While some will most won't and that will keep the network small, weak, and vulnerable to attack.

[pent]

Does anybody know how much power aux chain gets from parent chain? At least average?

Maybe we just have to switch to another sort of cryptography functions, which use floating point for example? The MM smells bad for me, its like a parasite. Needs additional research.

[Rassah]

Just finished the paper. Some questions:

You mention a part of the system includes compensation/ rewards for miners. In what form are those rewards distributed? Some type of coin that lets them register more names (like in Namecoin)? Some alternate version of Bitcoin that only has use as a currency? Something else?

Will this allow for merged mining? (hit post, walked away, didn't see it didn't post due to new replies) Namecoin mining doesn't pay a high reward, but since it is bootstrapped to Bitcoin mining, mining Namecoin is essentially free, and any reward above 0, no matter how small, makes it still be worth it. The other benefit is thatwith MM, the mining infrastructure is already there, and it's HUGE, giving the alt chain instant protection against attacks. Starting from scratch would mean people would have to dedicate time and resources specifically to this, which, since the rewards wind won't have much value at the beginning, will mean that people mining at the start will pretty much have to all be volunteers.

Are there any considerations for database sizes? I've read somewhere (take with grain of salt) that the current ICANN DNS database is over 10 terrabytes in size. That's not something regular home users can support. Will this database/blockchain eventually have to be relegated to a few large centralized data storage places?

Are there any considerations for growth? The part that concerns me about Namecoin is that block rewards are constant for long periods of time, which means the maximum number of registrations per year is limited. Will your chain generate a decreasing reward like Bitcoin, level rewards (eg 50 per block) for ever,  growing exponentially (3% increase every year to keep up with population/technology/other growth), or something else?

[DeathAndTaxes]

Does anybody know how much power aux chain gets from parent chain? At least average?

NMC is roughly 40% of the BTC network ~4TH/s.

Maybe we just have to switch to another sort of cryptography functions, which use floating point for example? The MM smells bad for me, its like a parasite. Needs additional research.

Maybe you misunderstood me.  The efficiency (or lack thereof) isn't what matters.

Say I have a 5970.  I can earn 0.5 BTC per day with it.  If performing proof of work for an alternative xCoin I can get 0.1 BTC (equivelent) per day then the xCoin network is asking me to forgo 0.4 BTC revenue to "help" xCoin.

Now lets abstract that further.  A 5970 costs ~$400.  $400 (and some time/skill) "buys" me 0.5 BTC per day in revenue.  Now lets say xCoin works best on TI-85 graphing calculators.  If $400 in calculators only nets me 0.1 BTC (equivelent) you are still asking me to take a lower return on my capital.  How you perform the proof of work isn't all that important.

Not saying you should merge mine but you should look at the economics of it.

You choices really are:
New GPU optimized algorithm  (directly competing against BTC daily revenue).
A new CPU based proof of work (since GPU miners have idle CPU)
Merged Mining GPU (SHA-256) proof of work
Merged Mining CPU (Scrypt) proof of work.

[matthewh3]

How about doing merged mining with Litecoin CPU mining then your not competing against NMC or anyone.

[pent]

In what form are those rewards distributed? Some type of coin that lets them register more names (like in Namecoin)? Some alternate version of Bitcoin that only has use as a currency? Something else?

The DIANNA will have its own financial block chain like Bitcoin. If we cut the domain operations, it will function like Bitcoin cryptocurrency with its own block chain and own "coins".
Those coins will be used also like alternate currecy.

Actually, DIANNA operates like NameCoin, but:
- registration fee is returned to miners (not destroyed) and caused additional proportional difficulty overhead to be not free
- registration fee defined by free market agreements, not by system. System can only suggest its value at first stages like bitcoin tx fee
- coins are not used as a domain name carrier

The registration fee of DIANNA's chain will be defined by communty like fees in "post-apocalyptic" Bitcoin (when coinbase dropeed to low values at lately stages).

Will this allow for merged mining? (hit post, walked away, didn't see out didn't post due to new replies) Namecoin mining doesn't pay a high reward, but since it is bootstrapped to Bitcoin mining, mining Namecoin is essentially free, and any reward above 0, no matter how small, makes it still be worth it.

I am still not considered with this. Merged mining needs additional research. See posts above.

Are there any considerations for database sizes? I've read somewhere (take with grain of salt) that the current ICANN DNS database is over 10 terrabytes in size. That's not something regular home users can support. Will this database/blockchain eventually have to be relegated to a few large centralized data storage places?

The first thing is yes, this will be heavy.
The second is the words of Satoshi:

The networks need to have separate fates.  BitDNS users might be completely liberal about adding any large data features since relatively few domain registrars are needed, while Bitcoin users might get increasingly tyrannical about limiting the size of the chain so it's easy for lots of users and small devices.

So the DIANNA won't be so popular due to big data sizes. Users will have to have a good hardware to run the full client.
Although, the lightweight clients (i.e. DNS clients) are possible. They will just listen to network and do not write whole database, but only needed namespace.

However I already thought to put the full block chain to DHT database and on clients keep only reasonable its part, other part will represented only as headers. If client need a full block record which it hasn't, he queries DHT for it. Since client has a hash of this block, he can verify whether DHT returned valid block.

Are there any considerations for growth? The part that concerns me about Namecoin is that block rewards are constant for long periods of time, which means the maximum number of registrations per year is limited. Will your chain generate a decreasing reward like Bitcoin, level rewards (eg 50 per block) for ever,  growing exponentially (3% increase every year to keep up with population/technology/other growth), or something else?

I think the Bitcoin emission model is fine. 50 coins means 5E9 "coin quantums". One coin is int64 = 1E8. I think this is totally fine. Correct me if I mistake.

[pent]

How about doing merged mining with Litecoin CPU mining then your not competing against NMC or anyone.

I don't competing with anyone. I just want Internet to be free. Like speech.

I need to research merged mining for considerations about its use.

[pent]

I just thought the solution to put full chain on DHT is very good

Found a doc about designing low-latency DHT.

pdos.csail.mit.edu/papers/dhash:nsdi/paper.pdf

One more research Anyone help? I do not going to own this project, it is a property of sociaty. And sociaty help needed.

[Rassah]

How about doing merged mining with Litecoin CPU mining then your not competing against NMC or anyone.

I don't cometing with anyone. I just want Internet to be free. Like speech.

I need to research merged mining for considerations about its use.

I think a big question you will have to answer is how will you get it started. Initially the coins will have very little value, relying on people donating their time and hardware for very little reward (likely negative profit), and the initial size of the network will mean 51% attacks or general mayhem will be very easy to carry out. It's the same issue that plagues all alt coins.
I'm not entirely sure why the parallel separate currency is needed, either, since the part that is used to register already has its own value. Can't the registration fee itself be set by the domain buyers, and miners decide whether the fee is high enough? (Admittedly, I don't understand why Namecoin has fixed registration fees)

I do hope this idea gets developed. I'm just concerned it may not be able to start.

[Raoul Duke]

Man, I wish I knew how to program so I could help you somehow.

Anyway, if you need something translated into portuguese or help to promote it I may be able to help you. Beyond that, I dunno.

[pent]

I think a big question you will have to answer is how will you get it started. Initially the coins will have very little value, relying on people donating their time and hardware for very little reward (likely negative profit), and the initial size of the network will mean 51% attacks or general mayhem will be very easy to carry out. It's the same issue that plagues all alt coins.

This you said divides on three parts:
1. The problem of initial demand. The initial demand for this system is in anonymous networks. Many-many people put their effords to invent similar system. Since it will work and lightweight client integrated in I2P, I sure ThePiratebay will be the first domain client. And this will be enough for big massmedia advertisement, like SilkRoad advertising bitcoin ))
2. The problem to motivate miners to mine. Its not a problem, I can mine by myself.
3. The main problem: 51% attacks. Here is a big field for thoughts. Either to use MM, which is totally alpha, or consider using more advanced cryptography which can deal with exponental network power growth.

I'm not entirely sure why the parallel separate currency is needed, either, since the part that is used to register already has its own value. Can't the registration fee itself be set by the domain buyers, and miners decide whether the fee is high enough? (Admittedly, I don't understand why Namecoin has fixed registration fees)

Separate currency needed to motivate miners to secure authoritative records and manage network.

I don't see a clear way to connect Bitcoin currency chain with alternate key/value chain. Financial and key/value transactions need to be in one solid chain.

Refering to my document:

Domain transactions are similar to financial ones and exist because of them, because a
record verification is an action which requires payment to motivate the participants to process it.
Consequently, domain transactions as well as financial transactions will have commissions
associated with them.
It became clear from the BitDNS discussion and looking into NameCoin that we will need to
draw a line between domain and financial transactions. Processing each type of transactions is a
different activity with its own complexity and compensation.
Nevertheless, introducing an additional block sequence for domain transactions may bring
about problems with synchronizing the two block sequences together. Let me remind you here that
the financial block sequence is quasi-stable and a part of it can be rewritten by the system at any
time.

[pent]

Man, I wish I knew how to program so I could help you somehow.

Anyway, if you need something translated into portuguese or help to promote it I may be able to help you. Beyond that, I dunno.

Thanks. If you are able to promote, please do it It would be better if more people will examine this design.