 |
March 15, 2017, 04:56:04 AM |
|
This "centralization" issue because of nodes is ridiculous. Bitcoin is a proof of work consensus system. The power resides in those that have proof of work: the miners. Nowhere else. What is that power ? It is the power to decide:
1) what new block to make, and what transactions to include
2) on what previous block to build
3) according to what protocol
in other words, the pools.
But this power is balanced by the power of users to determine market value. Pools are not going to do things that make users crash seriously the market value.
When you look at the pools, bitcoin is already an oligarchy. The pool bosses can easily sit together in a room, and represent a large majority of hash rate.
Yes, I hear you say, they are only the pools, they don't possess the mining hardware. But mining hardware doesn't decide about blocks. Mining hardware is only providing hashes for pools. The choice an owner of mining hardware has, is to join pool X or pool Y, and that will solely be determined by how much pool X or pool Y pay him for his hash rate. The mining hardware doesn't determine protocol, chain on which to build or whatever.
So how can block size centralize even a bit more bitcoin ? By the network quality between miners. The network latency BETWEEN MINERS makes those miners that have a poor network connection TO THEIR PEERS receive the latest blocks later than the others, and makes them put their blocks at disposal of the others later than their peers' blocks. The time lost to get the latest block, and the time lost when putting your block at disposal, is wasted hash rate, and higher probability to get your block orphaned.
THAT is the ONLY disadvantage larger blocks bring, or, the only "economy of scale" that larger miner pools have: they can invest in better network links between them.
Imagine that bitcoin has, say, 20 important mining pools. Then the full power over bitcoin is in the hands of these 20 entities.
Now, imagine that 18 of these pools are linked together with a 100 MB/s network backbone but that miners 19 and 20 have only 1MB/s links. Currently, they have to receive a block and it takes them 1 second, and when they send out their block, it also takes a second.
This makes that they have 2 seconds lost on 10 minutes. If now, the block size rises to, say, 50 MB, they would lose 50 seconds at the start when receiving a block, and 50 seconds at the end, when sending a block. They would lose almost 2 minutes on the 10 minutes: when other miners have 10 minutes to hash a block, they only have 8 real minutes. 20% hash rate lost.
So, 50 MB blocks would seriously hinder a 1MB/s linked miner: he would lose 20% of his investment.
In fact, the wise thing to do, would be to upgrade his network link. Compared to the cost of 20% of his hash rate, a higher network capacity might be a good investment.
But there's of course worse: miners with good links might send out some rubbish to their peers to keep their links saturated. But that goes in two directions: if miners start to DDOS one another, they might get hurt more than they are hurting the other miner. I think that the miner starting to DDOS his peers is going to get cut off. So they won't play that game.
So block size starts to accelerate bitcoins (unavoidable) centralisation from the moment that blocks are so big, that the network infrastructure to transmit them becomes more expensive than the cost of proof of work itself, and that only the biggest factories of proof of work can permit themselves such good network connections. I think that this must run in the size of many GB blocks before this becomes relevant.
Ah, yes, and the nodes of people ? Doesn't matter. They don't mean anything. Bitcoin is a proof of work system. If you don't deliver proof of work, you don't have anything to say.
|
