Who witnesses a Segregated Witness transaction and what incentive do they have to witness the extended block containing signatures?

2

This video claims Segregated Witness is insecure because no one has an incentive to process (witness) the extended signature block.

https://www.youtube.com/watch?v=VoFb3mcxluY

It claims that as the hashing power increases over 50%, you would be a fool (not a Nash equilibrium) to waste processing power on extended blocks containing the signatures, when the real money is in the transactions containing fees. This means anyone can feed a block with any transaction they want, since no one is checking signatures. (Paraphrased.)

From what I understand, Segregated Witness transaction addresses start with a 3 and are multi-signature addresses. So you have to sign, and someone (who?) also has to sign for the transaction to go through. All of these signatures are placed into an extended block.

  • Is the extended block also placed on the main blockchain?
  • Who processes the extended block containing signatures?
  • What incentive do they have to process the extended signature block?
  • Does the extended block have fees?
  • Who is the 3rd party signer for segregated witnesses multi-sig addresses?

Chloe

Posted 2017-12-19T21:00:59.550

Reputation: 568

type 3 doesn't necessarily mean, you are using multisig. It merely defines sigscripts, that are hashed (hence P2SH - pay to script hash). What is inside the hash, you cannot know - hash is a one way function... So you can have a segwit tx with one input and one output. – pebwindkraft – 2017-12-19T22:35:17.407

So it uses the signature field to include a hashed input & output address? But I thought the whole point of Segregated Witness was to eliminate (blank out) the signature field which took up 60% of a transaction and block? Can you add more details to an answer? – Chloe – 2017-12-22T20:33:53.203

no, I should leave the sigscript link away (non native english). I just want to say, that addresses appear in TX_IN and TX_OUT section. These addresses can be "of type 3". The way the TX_IN or TX_OUT are setup, is defined in "https://bitcoin.org/en/developer-reference#raw-transaction-format" and a very good overview in Andreas' book "Mastering Bitcoin", which is also online available...

– pebwindkraft – 2017-12-22T22:29:05.313

Answers

2

Since exchanges and merchants won't accept the blocks, miners who mine invalid blocks won't get paid. A miner could take that risk if they wanted to, but it's no risk to the system as a whole. Only the longest chain that consists only of valid blocks is honored by non-miners.

I could imagine a miner trying to mine on top of a new block without first checking the extended block. But they'd almost certainly start checking the extended block and switch if it turned out to be invalid. This takes an acceptable risk of mining an invalid block while ensuring that nobody has any incentive to deliberately mine invalid blocks.

David Schwartz

Posted 2017-12-19T21:00:59.550

Reputation: 46 931

"Won't accept the blocks..." which blocks? Can you be more precise? The block that says everyone just sent me 100 BTC is perfectly valid. It contains no signatures. As long as the hash is correct, it's valid. 1) Is the extended block also placed on the main blockchain? 2) What incentive do miners have to process the extended block? 3) Are there fees in extended blocks? Can you give an example of an extended block? – Chloe – 2017-12-22T20:54:51.993

@Chloe No, it's not valid. Unless the extended information exists with the hash cited in the block, the block is not valid. Nodes that check this will reject the block as invalid. Miners have to process the extended block because if they don't, they risk mining an invalid block (one for which the extended information does not exist and have the correct signatures) which will be rejected by every full node that checks. – David Schwartz – 2017-12-27T19:32:57.570

Asked: 2017-12-19T21:00:59.550

Viewed: 162 times

Active: 2017-12-19T21:12:27.503