Topic: Innovation in the crypto scene - Saito Network block production mechanism (Read 51 times)

Originally posted by David Lancashire to https://org.saito.tech/blog/ - September 9, 2019.

We periodically receive correspondence from people working to wrap their heads around Saito. One common issue people have is the idea that not using mining to produce blocks creates a threat to the network as similar to POS:

"With work as a percentage of the transaction fee, I believe there is an issue with selecting who gets to produce a block. Paying to produce blocks is an issue because no real world resources are spent. This means that you can try multiple times to produce blocks that benefit you."

As the Saito whitepaper explains, block producers are paid when golden tickets are included in subsequent blocks. So there is no way to produce blocks that benefit them more than others. And even if they could, what would be the point? Block-flooding defenses prevent attackers from spamming the network, while multiple blocks split miner focus and reduce the chance of a payout at all.

The mining payment is ungameable because it is costly to win. The routing payment is ungameable because it is costly to win: you only get money by paying more to the network than you get back. Honest nodes have zero problems with this – they are sharing fees given to them by real network users. But the same approach is a huge problem for attackers as it forces them to haemorrhage cash.

As a result, when someone attacks a Saito network, the proper response for the honest nodes is to sit back and light-up a cigar as the attacker empties his wallet into theirs. In the unlikely but delightful case that a Saito network is targeted by a state actor or Silicon Valley behemoth able to sustain their cost-of-attack for some time, impatient users may consider pushing up the burn fee to speed up the pace at which the wealth transfer completes, while more forward-looking participants may consider selling the attackers more tokens lest they run out and be forced to stop their wealth transfer prematurely.

Either way, since no-one has an unlimited supply of money, all attacks must necessarily end. At that point, other network participants may consider purchasing their own Swiss chalet or South American ranch with their share of the windfall profits.

But what about grinding attacks? If attackers can create blocks but then their attack fails they can get their money back. So if the attackers is trying to reorganize the blockchain, but they aren’t willing to spend enough money to overwrite the chain then the attack fails then the attacker gets their money back. So attacks are costless, no?

This question is trapped in the POW mindset, which says, “producing a block is expensive, therefore producing a series of blocks is expensive.” But do we really care about the first property or the second?

    1. Saito guarantees the cost of writing blocks in time. There are no “grinding attacks” that lower this cost — to produce blocks for less money attackers must increase the time between blocks. If attackers want to create a slower fork for less money of course they can do that… there are no shortage of POW or POS that do exactly this.
    2. This question implies there is some “hack” as in many POS networks where minting a special block somehow cheats the payment mechanism and reduces the cost of attack. But this is not the case. There are no “savings” to be had by creating multiple starting blocks for a fork. Attackers who solve golden tickets burn money on mining. Attackers who do not have no hope of getting their money back.
Those concerned that Saito makes it possible to purchase blocks at all should be more concerned about these attacks in POW and POS networks. It is TRUE that you can always buy blocks in Saito, but you can also do this in POW networks (the BCH / BSV fork involved an attack of this sort) and the situation will only get worse as the block reward falls and mining must be paid for out of transaction fees alone. POS networks are worse since there are the sort of grinding attacks you are talking about, staking pools are rapidly becoming a thing, and the ROI from transaction fees will only ever incentivize a small amount of staking on the lowest layer (which competes with economic activity within smart contracts as well as the broader economy).

The reality is that all blockchains that depend on external markets for their security function (hashpower / capital) are experiencing the emergence of business models that monetize the ROI from mining / staking by opening control of the market to anyone with a bank account and a willingness to spend money. This is one reason it is so important to solve the 51 percent attack and shift to genuinely secure mechanisms like Saito. True security comes with security functions that ensure attacks are ALWAYS expensive, as this is the only way to make sure they eventually bankrupt the attacker and stop.