During a panel discussion on Bitcoin scaling at the recent State of Digital Money event in Los Angeles, the idea that a larger block size limit would lead to further centralization of bitcoin mining was debated by the four participants on the panel: Airbitz CEO Paul Puey, derivatives trader Tone Vays, Yours CEO Ryan X. Charles, and Bitcoin Core contributor Eric Lombrozo.
The topic was first brought up by Vays in response to a statement from Puey regarding the need for the block size limit to be set by the free market. From Vays’s point of view, removing the block size limit completely and allowing it to be set by nodes on the network would lead to a situation where more powerful miners try to increase the block size limit for their own advantage.
“That scares me in [terms of] having a decentralized Bitcoin a lot,” said Vays.
As a counter to Vays’s claim, Charles pointed out that the more expensive aspect of mining is the actual mining process rather than the costs associated with operating a full node.
“Increasing the block size to 2MB has absolutely nothing to do with mining centralization,” Charles added.
Puey agreed with Charles’s point that the cost of an increased block size limit is nothing compared to the costs of mining equipment and electricity. “It’s nothing; it’s a rounding error,” he stated.
In Puey’s view, a better tool for the further decentralization of bitcoin mining would be financial incentives that push the industry to more countries around the world. Right now, a high percentage of Bitcoin’s network hashrate is assumed to be located in China due to the share of the overall network hashrate held by China-based mining pools.
When Lombrozo chimed in on this topic, he agreed that the costs associated with the system resources required to operate a full node are not the key issue with an increased block size limit that could lead to further mining centralization. In Lombrozo’s view, the real issue here is latency.
“The bigger the block gets, the longer it takes to propagate across the network,” said Lombrozo. “This is just a matter of physics.”
If it takes a long time for a block of transactions to get around the network on average, then miners with the best connections to a majority of the network hashrate can benefit from that latency.
“If you get a block a few seconds after everyone else already got the block, then everybody else has an advantage over you because they’re already mining on that block,” said Lombrozo.
Longtime Bitcoin developer and researcher Peter Todd gave a presentation that covered the issues around block propagation and miner centralization at Scaling Bitcoin Montreal back in 2015. It should be noted that there have been many improvements related to block propagation implemented since Todd’s talk.
Charles countered this issue of network latency and stated that miners will not want to create blocks that are too big because they will want the blocks to be small enough for other miners to receive the blocks and build on top of them in a timely manner.
“It is the case that miners want other miners to build on top of their block or they don’t get the money,” said Charles.
“Unless they’re the biggest miners and they mine on top of their own blocks and they don’t care if they get orphaned by the other miners because they’re so small compared to them,” Lombrozo countered.
According to Lombrozo, a few seconds of advantage in terms of block propagation can add up to considerable amounts of money when you’re talking about large mining operations. In his view, this would encourage big miners to centralize in data centers and have their own internal networks where they relay blocks to themselves (or a cartel of miners) first in order to gain an advantage. The block would then be relayed to other miners after the larger miners had already gotten a head start.
Here, Lombrozo is referring to a strategy known as selfish mining.
At this point, Vays noted that the places where bitcoin mining facilities are located, such as in the mountains of China, are not areas of the world that already have the best internet connections by default.
While Charles acknowledged this point, he also stated, “It’s still true that actually getting a sufficiently fast Internet connection is a tiny fraction of the actual expense for [miners]. It’s really like a non-issue I think.”
Charles added that those who can afford to build bitcoin mining facilities in the mountains can probably also afford to drag an internet cable there with them.
Vays responded that this is his entire point: Those who are building the most efficient mining facilities in the mountains can afford the better Internet connection and the little guy can’t; therefore, the increased need for a low latency connection caused by bigger blocks leads to further centralization.
At this point, Puey stated that it’s not necessarily increases to the block size limit that causes that issue.
“We’re not disagreeing that mining centralization is not a good thing,” said Puey. “It’s just, what’s causing it?”
As Puey previously indicated, he believes there are multiple factors at play.