Censorship and Censorship Resistance

People in the blockchain space often talk about censorship and censorship resistance. It is often claimed that censorship resistance is one of the key properties that decentralized, blockchain systems can provide at a level not possible with previous technologies. But exactly what is censorship resistance?

What is censorship exactly?

It helps to start with a fairly precise definition of censorship. By censorship, I mean discriminatory treatment of transactions for non-technical reasons. This doesn’t include conditions where the system as a whole fails for everyone nor does it consider situations where transactions become uneconomical due to escalating fees. It specifically covers the case where transactions face discriminatory treatment due to external characteristics such as the particular account or balance the transaction operates on.

Censorship also exists on a continuum that I like to describe as running from soft to hard. Hard censorship is the worst kind of censorship. By hard censorship I mean a situation where a particular operation that is permissible according to the system’s neutral rules is somehow made impossible for an extended period of time. By soft censorship I mean a situation where some party is advantaged or disadvantaged in transaction execution.

Hard Censorship in Blockchain Systems

All of the blockchain systems that claim censorship resistance have some properties in common. All system state and rules are public, which means that anyone can determine whether a transaction is valid or not by testing it against the system’s rules. Transactions can be submitted into the system by anyone because they are self-contained, that is, they contain internally all that is needed to determine their validity.

If someone is a victim of hard censorship, they have already tried to make their transaction public by submitting it to the network. Every participant in the network knows the network state and so can tell that the transaction is valid. Anyone can also submit the transaction themselves to the network if they wish to.

This means that any case of hard censorship is easily demonstrable. The victim can easily make the transaction public anonymously and anyone who sees it can tell that it is valid and submit it the network themselves. If the censorship remains, everyone will know it since they will see that the transaction was not accepted by the network.

Since hard censorship is easy to demonstrate when it happens and the victim has every incentive to do so, it is likely that if there are not reported instances of hard censorship, hard censorship is not happening. In general, any blockchain project that does not have complaints of hard censorship quite likely is not actually experiencing hard censorship. Of course, this does not mean that it has good resistance to hard censorship, just that nobody has done it yet.

Hard censorship is a deal killer for most use cases for public blockchains used to move digital assets. A system cannot remain neutral and decentralized if disfavored players can be easily excluded for non-technical reasons.

Soft Censorship in Blockchain Systems

Soft censorship is less annoying but harder to demonstrate. Say a large bitcoin mining pool decides to delay a particular account’s transactions by a few seconds prior to including it in blocks it mines. This is technically soft censorship. But even if they happen to mine a block that doesn’t include a particular transaction due to this rule, it’s entirely plausible that they simply didn’t see the transaction on time.

Unlike hard censorship, soft censorship only significantly affects some use cases. For a store of value, soft censorship really doesn’t much matter. As a means of exchange, soft censorship is a bit of an issue, but not very much. For a decentralized exchange, soft censorship resistance is important because soft censorship can be used to game an exchange with techniques like front running.

Censorship in Bitcoin and other PoW Systems

Bitcoin has excellent resistance to hard censorship unless a majority of miners conspire over an extended period of time. Such conspiring seems unlikely for a variety of reasons including that miners would decrease the value of their own investment in mining hardware.

However, bitcoin has virtually no resistance to soft censorship. A miner who engages in soft censorship likely reduces their transaction fee revenue slightly, and that does provide some incentive to handle transactions fairly. More importantly, there is virtually no reason to engage in soft censorship in bitcoin because there is nothing to be gained by delaying a transaction for a few blocks and, absent a broad conspiracy, no single miner can significantly affect any particular transaction originator. For the uses cases bitcoin targets, soft censorship is not really a problem.

Censorship in XRP Ledger

The XRP Ledger has excellent resistance to hard censorship. A majority of validators would need to conspire to censor in the short term. As explained above, such an instance would be easily provable and the victim would have every incentive to report it. Unless the network’s stakeholders want hard censorship (which seems unlikely since decentralized blockchains are much less useful and valuable with hard censorship) they will replace the validators that participated in the censorship (easily determined, since they sign their messages) with ones that did not participate in the censorship.

The XRP Ledger also has excellent resistance to most soft censorship. If a transaction is flawless and is broadcast prior to a ledger’s close time, every honest validator will vote to include it and only 50% is required. In other words, a 51% attack (a network-level failure) is required to soft censor.

More importantly, soft censorship on the XRP Ledger, if significant, is detectable and provable — the validators who voted to exclude the flawless transaction issued signed statements. While soft censorship is fundamentally part of bitcoin’s design, soft censorship is only tolerated in the XRP Ledger if the network’s users opt to tolerate validators conspiring to censor.

This permits the XRP Ledger to target use cases sensitive to soft censorship, use cases that PoW-based systems like bitcoin and Ethereum are far less suited for. This includes the XRP Ledger’s decentralized exchange.

A decentralized exchange built on a PoW chain will find miners increasing their revenues by gaming the exchange through things like front running on every block they mine. They won’t have a choice because mining is fundamentally a race to the bottom. Whoever can mine at the lowest cost with the greatest return will push the difficulty up by increasing the amount of mining they do until those less efficient are cut out. If gaming the system through soft censorship increases rewards significantly, miners will find they have little choice but to do so.

One thought on “Censorship and Censorship Resistance

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s