Introduction
Censorship resistance is enabled in large scale peer-to-peer networks like Ethereum or Bitcoin, since even one node could allow a message to be broadcasted to the rest of the network. However, when a large scale sanction on a particular service is made, as with Tornado Cash, then the entire community seems to rally with it, thereby, censoring the acts of the individuals associated with Tornado Cash. While this may be ethically sound, this has now set a precedent for cancel-culture, or SEC-induced sanctions to be viable for any decentralized protocol in the future. While others claim that this is also thwarted by forks of the original protocol, and that the SEC or any related authority cannot efficiently place sanctions on all look-alikes, due to the operational overhead, the quality of these look-alikes is often ignored, and therefore results in sub-optimal outcomes, which is worse for censorship resistance.
Societal Good vs Network Good
To truly prevent censorship, we must also acknowledge that with censorship resistance, we allow anti-social individuals to actively participate in the network. While they may not be a bad actor to the network, i.e, they may participate in the network with 99.9%+ uptime, have 0 equivocations, etc, but, they may be a bad actor in society. The fact that someone is a bad actor in society often changes with time, whereas a bad actor in the network, is a bad actor in the network for all time to come. This fundamental difference is what causes the pushback from those who are in control, in fear of the chaos that could be caused by societal bad actors. In order to have a truly decentralized network without censorship, it is imperative to include all good actors of the network, regardless of any other aspect which may label them as good or bad.
This does not mean that networks which emphasize societal good-ness shouldn't exist. Such networks could be useful in generating a reputation for arbitrarily good actors, which could power a whole suite of protocols which rely on an individual's reputation (Example - Augur, oracles, etc). The individual's reputation can be built over a long period of time, but slashed almost immediately based on their actions. This could prevent someone from acting adversely, at the cost of censoring their true action.
The game-theoretic aspect of censorship resistance is indeed unique, where a network does not want it's peers to Dox themselves to hold them accountable, nor does it want to enable societal bad actors from participating in the network. This is partially solved by having different networks with different use-cases, as mentioned above.
Barrier of Entry
A paper published in 2007, mentions censorship resistance of the data stored on a peer-to-peer network, but ignores the participation aspect of nodes on the same network. This also circles back to the different consensus mechanism's that blockchains use, like PoW, and PoS. An interesting fact is that Proof of Work was used in email spam prevention, years before it was actually used on the Bitcoin blockchain. Many have varied opinions on the use of PoW vs PoS, with reasons like energy efficiency, barrier of entry, etc. An analysis of barrier of entry to compare the two is required.
PoW
The barrier of entry with Proof of Work systems is dependant on the difficulty of the computation, which increases with the number of blocks produced by the network (trivially). The higher the difficulty to produce the next block is, the more expensive it is to acquire the hardware which can produce the next block quickly and efficiently. This allows only a subset of actual individuals to participate in PoW networks, and facilitated the creation of organizations like Mining Pools. Mining pools account for 1/5th of Bitcoin's hashrate, which reduces the decentralization of the network, and in turn, censorship resistance.
At the time of writing this article, it costs approximately $2000 to procure a Bitcoin mining rig, which turns a profit only after a year of mining.
The average salary in India, with a population of over 1.2 Billion individuals, is 4554 a year. It is impractical for an average person to spend a little under half their yearly net income on a Bitcoin mining rig, with the hope that it is able to keep up with the difficulty with time. There are a lot more costs associated with running a PoW rig, like maintainance, optimal cooling, hardware upgrades, etc.
Conclusion: Lower income individuals are censored by PoW
PoS
The barrier of entry with PoS systems is relatively lower in terms of monetary amount over a long period of time. However, the initial investment is large, since the stake is used to secure the network.
It costs approximately $59,000 to set up an Ethereum validator, and running costs to make sure that it has the least downtime as possible. This is much higher than the initial cost for PoW rigs, but the hardware costs and maintainance are cut down massively. Due to this higher initial cost, this leads to Staking Pools forming, which pool together the stake of multiple users, which, again, reduces decentralization, and in turn, censorship resistance.
Conclusion: Lower income individuals are censored by PoS
PoW, PoS, and Sybil Resistance
The reason that it is very expensive become a block producer on PoW or PoS based systems, is to prevent Sybil attacks. Sybil attacks are caused by the lack of identity on either networks, which again, if identities were a part of these networks, would enable censorship resistance.
This is a very difficult problem to solve, in a way that does not compromise confidentiality, integrity and authenticity of the peers.
Conclusion
While censorship-resistance may not be mature in the current state of peer-to-peer networks, there is much research being done in the space, from notable teams who run Monero, Secret Network, ZCash, and Status.im.
Thanks for reading!