elective-stereophonic
elective-stereophonic
Nxt 2.0 design singapore
Please login or register.

Login with username, password and session length
Advanced search  

News:

Latest Stable Nxt Client: Nxt 1.12.2

Pages: 1 ... 45 46 [47] 48 49 ... 51

Author Topic: Nxt 2.0 design  (Read 242483 times)

apenzl

  • Hero Member
  • *****
  • Karma: +248/-10
  • Offline Offline
  • Posts: 2498
    • View Profile
    • Nxter.org
Re: Nxt 2.0 design
« Reply #920 on: March 02, 2016, 12:17:34 am »

Assets being global does this mean that the assets we have today can be traded on childchains in these chains' "local" currency?
Yes, this is one important goal of the design. When ask/bid order transactions are submitted on a child chain, they will be denominated in the local currency of that child chain, even if the asset issuance transaction was done on another child chain. And a dividend payment can be submitted on a different child chain, resulting in distribution of that child chain local currency to shareholders.

Since some asset issuers may not want their assets to be tradeable on all child chains, or vice versa, this should be configurable to allow restricting an asset to only one child chain, or a child chain to not accept trades of assets originating from other chains. These are all details to be worked out.

I see. Thank you for confirming that.

The creation of a SuperBTC sidechain = a decentralised Nxt AE < > BTC market.

BUMP:

To make this discussion on topic please try to avoid "forward looking" statements such as:
"The value of token X will increase/decrease causing the value of token Y to increase/decrease - this will destroy NXT"

In a world where the value of any given token increases or decreases 10% an hour on a daily basis due to reasons that nobody understands, it's really not serious to make such predictions about the future. Let alone base decisions on such predictions.
Regarding the "this will destroy NXT" postfix that seems to accompany so many messages in this thread. Relax, come down, take a big breath.
Everybody here would like to see NXT succeed.
We know changing an existing system is always difficult but NXT must not stagnate like Bitcoin it must evolve.

What would be helpful is if you think about how your application, business, asset etc can be adapted to take advantage of this new design and the scalability increase it will offer.
Then see if you can understand how it will work. Does it make sense for your application to use it's own child chain ?
For example, over the last year or so, I've been approached by several application developers which wanted to issue an MS currency but had a problem paying fees in NXT.
With NXT 2.0 these applications can issue a child chain token and also pay transaction fees in this token.

Not only that, the users of this applications can now pay very low transaction fees or not pay any fees at all, as long as they will make sure their prunable transactions are included in the forging chain blocks. This is a real world problem solved by this design. There are many others.

Asset to Asset trading is another potential problem solved by this design. Think about it.

I encourage you think about your existing use case and its limitations. Can we solve them with a child chain ? If not, why not ?
« Last Edit: March 02, 2016, 12:24:59 am by apenzl »
Logged

Jean-Luc

  • Core Dev
  • Hero Member
  • *****
  • Karma: +816/-81
  • Offline Offline
  • Posts: 1610
    • View Profile
Re: Nxt 2.0 design
« Reply #921 on: March 02, 2016, 03:22:30 pm »

I've also been wondering about this area. Especially that sentence from the original post, "Anyone can create a ChildchainBlock transaction."

We need a name for the nodes that create ChildchainBlock transactions, so I'm going to call them "bundlers" because they bundle up NXT transactions. One approach for a bundler would be to accumulate NXT transactions until the total NXT fees justify a ChildchainBlock, and then broadcast it. Holding onto it becomes increasingly risky as we get closer to when the next fNXT block is due, because network delays might mean the forger doesn't get it in time. Is there any motive for a bundler to delay broadcasting it at all? When another NXT transaction comes in, they can add it to their ChildchainBlock and broadcast the updated version. Then for every NXT transaction over the fee size, we could get a new ChildchainBlock broadcast that is far bigger than the one NXT transaction it adds. It feels like it could be very inefficient in bandwidth. Add that anyone can create ChildchainBlocks, then we could get scores or hundreds of nodes all creating their own ChildchainBlocks from the same NXT transactions in the hope that the forger picks them and they make a profit.

There are some optimizations we can do here to reduce the increase in network traffic. Propagating a ChildchainBlock transaction does not require also sending the JSON of all transactions that it includes - it only needs to include the transaction id's. Peers most likely already have the actual unconfirmed transactions, and if a node is missing one or two, we could request them separately from the peer that sent us the ChildchainBlock transaction.

Also, even if multiple bundlers submit ChildchainBlocks for the same set of child chain transactions, those don't need to be propagated further unless they offer to pay higher fee. That is, if a node already has a ChilchainBlock for the same transactions, it would just ignore and drop any further such ChildchainBlocks received from others, unless they pay more. Thus different segments of the network will end up having different ChildchainBlocks waiting to be included in the next block, but what matters at the end is which one made it first to the node that is going to forge the next block.

Quote
Alternatively, and this may address your point, maybe we won't get any ChildchainBlock transactions broadcast at all. I don't really see what value the bundlers add, at least for an established child-chain like NXT. They are middle men. They aren't doing anything the forgers can't do for themselves more efficiently, and yet they are skimming off some of the profit. If I were a forger I think I'd ignore ChildchainBlocks broadcast by other nodes (except to copy their NXT transactions). I'd build my own instead, and include them in my fNXT block when it is my time to forge. That way I never pay fNXT fees (or rather, I pay them to myself). We'd end up with a similar situation to today, where the NXT bundlers are just the fNXT forgers. The effective NXT block time would be 1 per minute, the same as for the fNXT chain.

We will probably have this as the default, for the NXT child chain, if you are forging you will also be doing bundling of NXT transactions by default in the blocks that you forge, unless turned off. This way a forger will be accumulating NXT from the fees included in the child chain, and just signing a bundling transaction paying the fNXT to itself. Unless the node receives a bundling transaction from someone else, willing to pay more fNXT, in which case the rational choice would be to accept it.

If a forger wants to get fNXT and not keep the NXT from the fees, he would turn bundling off. And the opposite, if someone wants to convert fNXT to NXT, he would be doing bundling even for blocks forged by others (i.e., will be broadcasting ChildchainBlock transactions). This way one can do a slow exchange of small amounts of fNXT/NXT, without paying the high forging chain transaction fees.
Logged
GPG key fingerprint: 263A 9EB0 29CF C77A 3D06  FD13 811D 6940 E1E4 240C
NXT-X4LF-9A4G-WN9Z-2R322

Windjc

  • Sr. Member
  • ****
  • Karma: +59/-17
  • Offline Offline
  • Posts: 360
    • View Profile
Re: Nxt 2.0 design
« Reply #922 on: March 02, 2016, 10:37:04 pm »

Well the mastermind economists behind Nxt 2.0 have done a stellar job so far. Somehow, they have figured out a way to make a huge fundamental "upgrade" announcement have a catastrophic impact on a price that was already completely oversold. Great job!
Logged

farl4bit

  • Hero Member
  • *****
  • Karma: +210/-45
  • Offline Offline
  • Posts: 3466
    • View Profile
    • Crypto Advies
Re: Nxt 2.0 design
« Reply #923 on: March 02, 2016, 11:48:55 pm »

Just spoke Riker in reallife on bitcoin Wednesday and I have 100% trust in the 2.0 design. The Nxt big thing.  ;D
Logged

neofelis

  • Hero Member
  • *****
  • Karma: +74/-12
  • Offline Offline
  • Posts: 568
    • View Profile
Re: Nxt 2.0 design
« Reply #924 on: March 04, 2016, 05:56:23 pm »

All my support for 2.0.  It is visionary and pushes crypto to where it needs to go for global utilization.  Multiple blockchains, all interdependent.  We are all betting on the future and I think this is where it's headed.

Yeah for JLP and all the devs.

 :) :)
Logged

lopalcar

  • Hero Member
  • *****
  • Karma: +99/-15
  • Offline Offline
  • Posts: 561
    • View Profile
Re: Nxt 2.0 design
« Reply #925 on: March 04, 2016, 11:24:13 pm »

First: I have almost no clue about how difficult or even possible would be this suggestion, just an idea which I don't know if was mentioned and dismissed before

Suggestion: Why not one token building in parallel two chains "already know as fnxt chain and nxt childchain" "transactions on nxt childchainblockchain don't need to be bundled, they are automatically stored by the forgers as childchain blocks in the forgingBlockchain (the bundler and the forger are the same person)". The rest of childchains works as suggested till now, but the value of nxt as we know till today can remain untouched.

And one question about childchains: Private keys will open the same account in every childchain? Or will be a different algorithm selected by the issuer for generate accounts?
Logged

Jean-Luc

  • Core Dev
  • Hero Member
  • *****
  • Karma: +816/-81
  • Offline Offline
  • Posts: 1610
    • View Profile
Re: Nxt 2.0 design
« Reply #926 on: March 05, 2016, 02:46:18 pm »

Suggestion: Why not one token building in parallel two chains "already know as fnxt chain and nxt childchain" "transactions on nxt childchainblockchain don't need to be bundled, they are automatically stored by the forgers as childchain blocks in the forgingBlockchain (the bundler and the forger are the same person)". The rest of childchains works as suggested till now, but the value of nxt as we know till today can remain untouched.
Doesn't work, this is just another perpetuum mobile attempt. Besides a childchain block is a bundle of transactions, so of course those transactions do need to be bundled. Who does it doesn't change anything.

Quote
And one question about childchains: Private keys will open the same account in every childchain? Or will be a different algorithm selected by the issuer for generate accounts?
Accounts are global, the private key to account mapping will remain unchanged.
Logged
GPG key fingerprint: 263A 9EB0 29CF C77A 3D06  FD13 811D 6940 E1E4 240C
NXT-X4LF-9A4G-WN9Z-2R322

farl4bit

  • Hero Member
  • *****
  • Karma: +210/-45
  • Offline Offline
  • Posts: 3466
    • View Profile
    • Crypto Advies
Re: Nxt 2.0 design
« Reply #927 on: March 05, 2016, 04:20:36 pm »

In the interview between Lior (Riker) and Marc de Mesel is some good explanations and reasons about the Nxt 2.0 design. Be sure to check it out!  :o

http://youtu.be/tN4FjZ-31uk
Logged

lovely89

  • Jr. Member
  • **
  • Karma: +12/-0
  • Offline Offline
  • Posts: 66
    • View Profile
Re: Nxt 2.0 design
« Reply #928 on: March 08, 2016, 05:05:15 am »

Is there a comparison of the new proposed design and lisk anywhere? I'm wondering, if nxt 2.0 launhces and implements automated transactions, will it be a similar design to lisk?
Logged

Cassius

  • Hero Member
  • *****
  • Karma: +207/-18
  • Offline Offline
  • Posts: 2459
  • Rather be a pirate than join the navy
    • View Profile
Re: Nxt 2.0 design
« Reply #929 on: March 08, 2016, 07:43:18 am »

Is there a comparison of the new proposed design and lisk anywhere? I'm wondering, if nxt 2.0 launhces and implements automated transactions, will it be a similar design to lisk?

The two projects will be almost entirely dissimilar.
Logged
I head up content for BitScan, crypto business hub.

Kiomansk

  • Jr. Member
  • **
  • Karma: +12/-1
  • Offline Offline
  • Posts: 55
    • View Profile
Re: Nxt 2.0 design
« Reply #930 on: March 08, 2016, 01:46:56 pm »

Is there a comparison of the new proposed design and lisk anywhere? I'm wondering, if nxt 2.0 launhces and implements automated transactions, will it be a similar design to lisk?
AFAIK Lisk is mostly an offshoot of Crypti or a continuation of that project.

It's closer to Ethereum than it is to NXT.
Logged

farl4bit

  • Hero Member
  • *****
  • Karma: +210/-45
  • Offline Offline
  • Posts: 3466
    • View Profile
    • Crypto Advies
Re: Nxt 2.0 design
« Reply #931 on: March 08, 2016, 06:31:23 pm »

So are the devs going to start with the development of 2.0? I want this as soon as possible in the air. A lot of competition in the crypto scene.  ::)
Logged

Jean-Luc

  • Core Dev
  • Hero Member
  • *****
  • Karma: +816/-81
  • Offline Offline
  • Posts: 1610
    • View Profile
Re: Nxt 2.0 design
« Reply #932 on: March 08, 2016, 09:11:28 pm »

I am already working on 2.0.
Logged
GPG key fingerprint: 263A 9EB0 29CF C77A 3D06  FD13 811D 6940 E1E4 240C
NXT-X4LF-9A4G-WN9Z-2R322

TheWireMaster

  • Sr. Member
  • ****
  • Karma: +27/-0
  • Offline Offline
  • Posts: 356
    • View Profile
    • NXT Folks
Re: Nxt 2.0 design
« Reply #933 on: March 08, 2016, 09:21:37 pm »

Logged
NXT-5WW2-XQ63-CFGM-G7YAJ

Nxter

  • Hero Member
  • *****
  • Karma: +61/-7
  • Offline Offline
  • Posts: 597
    • View Profile
Re: Nxt 2.0 design
« Reply #934 on: March 08, 2016, 09:45:44 pm »

I am already working on 2.0.
Those are good news :)
Logged

Marc De Mesel

  • Hero Member
  • *****
  • Karma: +228/-83
  • Offline Offline
  • Posts: 643
    • View Profile
Re: Nxt 2.0 design
« Reply #935 on: March 08, 2016, 11:05:24 pm »

durerus

  • Hero Member
  • *****
  • Karma: +106/-8
  • Offline Offline
  • Posts: 527
  • user-owner
    • View Profile
Re: Nxt 2.0 design
« Reply #936 on: March 08, 2016, 11:52:54 pm »

Interesting! BCNext's concept of Parallel Chains looks very similar to the fNXT design proposal: https://bitcointalk.org/index.php?topic=345619.msg5585029#msg5585029:

Quote
Quote
Quote from: Come-from-Beyond on March 08, 2014, 11:09:35 AM
RFC: Parallel Chains concept

Only TL;DR version, coz noone would read a full one.

Master Chain

Contains only checkpoints of all slave chains. Checkpointing is done once a day and only when 1440 blocks r built on top of the corresponding slave chain. Master chain is never pruned. Growth rate is [32 bytes * numberOfSlaveChain] per day.

Slave Chains

Contains only 1 type of transactions. Different currencies can be implemented as different chains. Forgers can choose what chains to secure. The market balances TPS rate.

Migration plan

Create slave chains that implement all transactions types existing in Nxt. Add a new type for checkpointing. Reject attempts to include non-checkpointing transactions into the master chain.

Side-effects

The Chinese could use a separate currency inside their borders for very high TPS rates. Only checkpointing transactions have to bypass Great Firewall of China. Speculators provide currency exchange service - the business they love to do. NXTs become "tokens", users buy them for fiat to spend for fees, it's similar to prepaid coupons/tickets for provided services.

How do we prevent bloating the master chain?

New chains need to be registered at the master chain for a fee.

Logged

apenzl

  • Hero Member
  • *****
  • Karma: +248/-10
  • Offline Offline
  • Posts: 2498
    • View Profile
    • Nxter.org
Re: Nxt 2.0 design
« Reply #937 on: March 09, 2016, 12:10:43 am »

Interesting! BCNext's concept of Parallel Chains looks very similar to the fNXT design proposal: https://bitcointalk.org/index.php?topic=345619.msg5585029#msg5585029:

Quote
Quote
Quote from: Come-from-Beyond on March 08, 2014, 11:09:35 AM
RFC: Parallel Chains concept

Only TL;DR version, coz noone would read a full one.

Master Chain

Contains only checkpoints of all slave chains. Checkpointing is done once a day and only when 1440 blocks r built on top of the corresponding slave chain. Master chain is never pruned. Growth rate is [32 bytes * numberOfSlaveChain] per day.

Slave Chains

Contains only 1 type of transactions. Different currencies can be implemented as different chains. Forgers can choose what chains to secure. The market balances TPS rate.

Migration plan

Create slave chains that implement all transactions types existing in Nxt. Add a new type for checkpointing. Reject attempts to include non-checkpointing transactions into the master chain.

Side-effects

The Chinese could use a separate currency inside their borders for very high TPS rates. Only checkpointing transactions have to bypass Great Firewall of China. Speculators provide currency exchange service - the business they love to do. NXTs become "tokens", users buy them for fiat to spend for fees, it's similar to prepaid coupons/tickets for provided services.

How do we prevent bloating the master chain?

New chains need to be registered at the master chain for a fee.

 :D

I re-visited that one too, a few weeks ago.

We had a somehow more fruitful brainstorm back then, which is summarised here http://nxter.org/nxt-parallel-chains/

farl4bit

  • Hero Member
  • *****
  • Karma: +210/-45
  • Offline Offline
  • Posts: 3466
    • View Profile
    • Crypto Advies
Re: Nxt 2.0 design
« Reply #938 on: March 10, 2016, 11:03:27 am »

I am already working on 2.0.

Nice!!  ;D

Logged

mr001

  • Jr. Member
  • **
  • Karma: +7/-1
  • Offline Offline
  • Posts: 71
    • View Profile
Re: Nxt 2.0 design
« Reply #939 on: March 10, 2016, 06:30:42 pm »

Logged
Pages: 1 ... 45 46 [47] 48 49 ... 51
 

elective-stereophonic
elective-stereophonic
assembly
assembly