About merged mining

Quote from: Bitinvestor on February 26, 2014, 05:28:37 PM

It seems that merged mining for scrypt has arrived:

https://litecointalk.org/index.php?topic=15666.0

That makes me want to build a rig and start PoW mining again!
I think Novacoin should implement merged mining as well because it would make us truly energy efficient.

You are right about energy efficiency but there are a some problems:

  • AUX PoW is cost-free for a miner;
  • There are a few big mining pools which will be able to introduce a centralization into PoW mining.

These things are serious problem because one or two big LTC pools will take a major part of PoW emission for free. I presume that we shouldn’t introduce a possibility of this scenario, especially if the pool operator decides to keep mined coins for himself instead of distribution. Because it allows somebody to get a significant part of stake for free, which contradicts to the network security model.

Some fixes

Some fixes were merged to resolve address book page and listtransactions issues (special thanks to svost for his investigation).

https://github.com/novacoin-project/novacoin/pull/9
https://github.com/novacoin-project/novacoin/pull/10

  • Qt: fix known addressbook bugs (duplicate addresses etc.);
  • Core: Avoid trying to parse outputs that aren’t relevant to CWalletTx::GetAmounts;
  • Core: Supress “address” key in transaction details, when the destination isn’t recognized;
  • Qt: copyrights update.

Windows nosetup builds:

0.4.4.6 update3: [ AMD64 | i686 ]
0.4.4.7 bugfix4: [ AMD64 | i686 ]

Linux, OS X builds and installers will be published later.

Protocol update: preconditions and prospects

A new proposal for Proof-of-Work signing protocol is currently under development. It would provide stakeholders an ability to vote for Proof-of-Work blocks they’ve received. What is interesting, it’s possible to implement without breaking a compatibility with nodes that doesn’t support this protocol.

A main idea is quite simple:

1. Proof-of-Work miner submits a new block over the network;
2. New block is accepted as a “candidate” (score = 1);
3. Each node builds a deterministic list of stake generators from the past (64 items, for example);
4. Stake holder appends own signature to vchBlockSig field (currently it’s not used) and then submits signed block over the network;
5. Once enough signatures collected (16 of 64, for example) new block is accepted into block chain as a “full value” block (score is equal with provided by the current proof-of-work score policy).

This process would be able to prevent malicious miner from creating a secretly pre-mined chain in order to perform 51%+ attack even if he has 99% of hashing power and significant stake volume. Because he can’t collect enough signatures while keeping his chain secret.

Release 0.4.4.7

Current 0.4.4.7 testing branch runs pretty stable for now. This branch is available here for public testing:

https://github.com/novacoin-project/novacoin/tree/current

Basic list of changes:

1) Pruned transactions serialization support.

Pruned transaction records are now saved into coinsDB. These records are composed of 13-14 bytes of metadata (block height, coinbase/coinstake flags, block and transaction timestamp). This resolves all syncronization issues. Though currently there is no lifetime limit for such records, cleanup for old records will be implemented in the future versions of NVC client (this cleanup will be performed after reaching the 520-blocks maturity window, I guess).

2) Conflicted transactions handling.

Valid XHTML 1.0 Transitional Valid XHTML 1.0 Transitional Valid CSS Valid CSS      I find a bug/typo or I have a suggestion
Adapted for Lynx-like browsers
2013-2017 © Zloy & Co
For donations: 4ZLoY6T8wXKgrbZM9gL8fQt48NUzGo2yAg