Solving block propagation orphan risks through mining empty blocks (on top of headers only)

[seweso]

The easiest way to solve orphan risks due to block propagation problems is to let miners mine empty blocks onto headers only.

Obviously this should not be confused with SPV mining where blocks are never downloaded and verified. If a block is verified and correct normal blocks should be generated. And if a block verified and deemed wrong the miner should return to mining on top of the previous block.

The risks is having confirmations which do not say anything about the validity of the previous block. But if we also decide that empty blocks do not count as a confirmation, then nothing of value should be lost.

If the blocksize limit is increased, mining a few empty blocks (right after the previous one) should not be such a big deal.

I think this is the easiest solution to solve orphan risks/cost stemming from slow block propagation, and it suits with BU.

 

[freetrader]

I'm trying to understand the reasoning behind your argument, but I could use some help.

We have a problem - orphan risk due to block propagation issues.

How exactly does allowing miners to mine empty blocks which don't count make the situation any better (except pay miners for what seems like useless spinning)?

Wouldn't this allow miners to collude to raise their max block size beyond what block propagation actually supports, and then even the majority of miners could just "pretend to keep up" by churning out empty blocks, collecting some "feel good money" while they make plans to upgrade (if they do at all)?

If that's the case it would only reduce the efficiency of mining as a whole, no?

It would seem to me that to make real headway we have to fix the actual problem by improving block propagation (*), and be careful of offering incentives that might actually delay that.
Good (**) miners should plan ahead for capacity, just like other service providers.

Please correct me if I'm completely mistaking the point.

(*) which is part an infrastructure problem, but there is probably a lot of room for optimization in the software (I don't know about gossip networks, but it seems @Gavin Andresen is studying this area)

(**) in the sense that they'll survive in the market

 

[sickpig]

@seweso,

Sergio Demian Lerner seems to agree:


block post about SPV mining as a solution to block propagation:

https://bitslog.wordpress.com/2016/01/08/spv-mining-is-the-solution-not-the-problem/

really interesting, among the other things it contains a few critics to Corallo relay network.

 

[seweso]

How exactly does allowing miners to mine empty blocks which don't count make the situation any better

Because then bigger blocks (with increased propagation time and verification time) will not have increased orphan risks.

 

[theZerg]

Did anybody read my paper? www.bitcoinunlimited.info/1txn

 

[seweso]

I'm going to read it. Hope I understand everything

 

[seweso]

So with the current state of the network (and without block propagation optimizations) we would arrive at a natural maximum of 30Mb blocks.

There is no tragedy of the commons because miners would already not be incentives to create infinite blocks just by looking at Bitcoin itself. So no external economic pressure needed.

So, that's wonderful right?

 

[theZerg]

Yes and 30mb is with a huge fee. With realistic fees prob just a few mb.