If Ripple acts maliciously token holders could, theoretically, fork the XRP Ledger. In fact, this is one of the features that makes the network decentralized, argues CTO David Schwartz. Here’s how to actually do it and why it ultimately raises concerns around Ripple’s control over XRP.
How to fork XRP Ledger
In a technical discussion last week, Ripple CTO David Schwartz revealed it’s relatively easy to fork XRP Ledger.
If a group of nodes adopts the same UNL, which doesn’t include any nodes outside of their group, then that would allow them to do the equivalent of a fork, agreed Schwartz. This would create an entirely new tradeable asset using the history of the XRP Ledger from that point forward.
Schwartz added that “the forked off group could then make any code and amendment changes that they wanted.”
A successful fork could allow a group to make enormous changes to a blockchain. Like in Bitcoin, those who control code changes and amendments could, for example, create new coins or increase the supply cap.
In December of last year, Schwartz clarified that forking XRP Ledger is even easier than forking a proof-of-work (PoW) coin like Bitcoin.
“If a group wants to fork the XRP ledger, all they need to do is agree on system rules and roughly agree on validators that will follow those rules. It’s much easier than forking a PoW chain because you don’t need miners and you don’t become insecure if you can’t incentivize them.”
Holding Ripple accountable
The ability to fork the XRP Ledger is critical because it’s one of the primary arguments David Schwartz and others backing XRP use to defend its decentralization. Ripple is emphatic that XRP Ledger is, in fact, not controlled by the corporation.
Any two groups who don’t agree can fork the XRP Ledger, says Schwartz:
“If nobody agreed with Ripple, Ripple would have to choose between being alone on a worthless network or giving in.”
“I do think that fork freedom which ensures value creators can’t be coerced into running with rules they don’t like is one of the most important aspects of decentralization.”
However, it seems practically implausible to fork the XRP Ledger, raising questions over the nature of Ripple’s control over XRP. To better understand this dynamic, it’s important to understand the nuances of XRP Ledger’s consensus algorithm.
Primer on Ripple Protocol Consensus Algorithm
XRP Ledger has a unique method of consensus, known as Ripple Protocol Consensus Algorithm (RPCA). Like Bitcoin, anybody can spin up a node and begin participating as a “validator” on XRP Ledger. Unlike Bitcoin, these validators are not financially incentivized.
“If you run an XRP Ledger server to participate in the network, the additional cost and effort to operate a validator is minimal. This means that additional incentives, such as the mining rewards in Bitcoin, are not necessary,” states the technical FAQ. “The primary incentive to run a validator is to preserve and protect the stable operation and sensible evolution of the network.”
Reaching consensus on XRP Ledger
For XRP Ledger, consensus is determined through agreement between parties running the Ripple Server software. Each node running this software is called a “server” in Ripple’s technical documents. Each server maintains a list of other nodes they trust “not to collude in an attempt to defraud the network,” called the unique node list (UNL).
Consensus requires 80 percent of a server’s UNL to agree on a transaction. All transactions that meet the 80 percent threshold are included in the ledger and finalized. Overall, if 80 percent or more of “trusted” nodes cannot reach consensus then finalization of the ledger halts until the threshold is achieved.
For comparison, Bitcoin nodes come to consensus on the state of the ledger based on “block height,” which abstractly represents the chain which has had the most mining power support it. Assuming a miner successfully mines a block, they can include whatever transactions they wish (usually based on the attached fee amount) as long as those transactions are valid based on the rules of the network.
There are so many other advantages to our approach. Cheaper. Faster. No unconfirmations or double spends. Better censorship resistance. Security not dependent on token value. What do you do in a PoW system if you don't like the rules 51% of hashing power wants?
— David Schwartz (@JoelKatz) June 8, 2019
On a network level, the valid ledger that represents the XRP token is abstractly based on the collective ‘reputation’ of a smaller set of all servers. The more times a server is “trusted” via inclusion on another server’s UNL the more importance it gains on the network. Said again by one analyst, the nodes with the largest number of UNL connections have the greatest influence inside the network.
The default UNL
Ripple determines the default UNL list of trusted nodes. From the onset the corporation chooses 33 of the most ‘reputable’ servers (out of over 1,000), which ultimately determines which transactions are included in the ledger and which version of XRP is canonical. 7 of these nodes are operated by the company directly.
“Ripple provides a default and recommended list which we expand based on watching the history of validators operated by Ripple and third parties,” states the technical FAQ.
This means the security and fidelity of the XRP Ledger hinges on this trusted set of servers. It is also a heavy point of contention among critics.
The vast majority of stakeholders choose to trust this default list since there is little reason to deviate. More broadly, research indicates people tend to choose the default the vast majority of the time. Research from influential cryptographer Peter Todd supports this assertion.
Since the nodes with the largest number of UNL connections have the greatest influence inside the network, that means Ripple has substantial control over XRP Ledger by controlling the default list. A number of influential crypto analysts have brought up centralization concerns around XRP.
Comparisons to other cryptocurrencies
In proof-of-work cryptocurrencies, miners are one stakeholder that lobbies for changes on a network. And, miners are financially incented to do so.
Bitcoin alone generates over $20 million in revenue for miners every 24-hours from block rewards and fees, giving these miners the resources needed to lobby other network participants. It also makes it more likely that economic value will actually follow a fork, as seen in Bitcoin Cash’s fork from Bitcoin; and Bitcoin SV’s split from Bitcoin Cash.
In proof-of-stake systems, token holders themselves are empowered to participate. Cryptocurrencies such as EOS, Tezos, and Augur have dedicated enormous resources to building mechanisms to poll token holders and hold founders accountable. Collectively, these processes are called “on-chain governance.” (Though, voter participation among these projects has thus far been abysmal.)
Other cryptocurrencies such as TRON, Cardano, NEO, and arguably XRP are organized in such a way where a single entity has enough influence to fundamentally alter the underlying rules of their networks. Subsequently, this means they have a large influence on the value of underlying coins.
If the Justin Sun wants to change something on TRON it will likely get implemented. For now, IOHK calls the shots for Cardano. The NEO Foundation controls a huge treasury and determines which nodes get to operate on its blockchain.
Different kinds of decentralization
Technically, these networks are ‘decentralized’ in two of the ways Vitalik Buterin described in his famous 2017 essay, “The Meaning of Decentralization.” The highlighted crypto projects are architecturally decentralized, in that the system can handle several physical computers on the network breaking down—and logically decentralized: if you cut the network in two halves they each could continue existing independently.
But, in terms of political decentralization, the organizations behind the highlighted networks still have undue influence. Vitalik Buterin calls this “political centralization.”
As an example, Bitcoin is politically decentralized because if any one entity, such as Bitmain or the Bitcoin Foundation, were suddenly eliminated then the network would continue existing relatively unscathed. For XRP, if Ripple suddenly disappeared it’s unlikely the token would continue to see appreciable use.
Forking XRP Ledger is impractical
Ripple CTO David Schwartz uses the technical ability for people to fork XRP Ledger as a reason why it’s decentralized. Yet, doing so in a way that’s economically meaningful is implausible.
As Schwartz said himself, forking a large cryptocurrency and having any economic value follow the fork is incredibly difficult. Nonetheless, there are hundreds of examples of this happening with proof-of-work coins.
However, doing so with XRP Ledger doesn’t make any sense because of the control Ripple has over the network.
Economically, Ripple and insiders at the company control more than 50 percent of the total XRP supply. In terms of trading, Ripple plays a large role in the liquidity of the token, paying market makers to stimulate trading volume. The company is a large contributor to the supply of the coin, selling XRP in larger and larger allotments as a percentage of overall trading volume.
Finally, in terms of the XRP Ledger network, Ripple controls the default node list. The overwhelming majority of the network trusts this list of nodes. If the nodes with the largest number of UNL connections has the greatest influence on the network then by extension Ripple has the greatest influence on XRP Ledger.
Nevertheless, Ripple argues that it doesn’t create a “centralized system”:
“The XRP Ledger network is opt-in. Each participant directly or indirectly chooses its UNL. Should Ripple stop operating or should Ripple act maliciously, participants could change their UNLs to continue using the XRP Ledger.”
But, to assume fragmented and non-incentivized stakeholders will make an effort to fork XRP Ledger—who are also in the minority in terms of political power—in cases of negligence on the part of Ripple seems optimistic.
Empower XRP holders
If Ripple cares about decentralization then there are a few things it can do to further distribute political power on XRP Ledger.
As a reference, EOS and Tezos are attempting to empower other stakeholders on their respective networks. By giving block producers and token holders a way to organize and express their opinions a fork can actually materialize and hold weight in terms of decision-making.
If Ripple wanted to construct ways to poll token holders and incorporate on-chain governance, it could. But token holders aren’t the primary customers Ripple is targeting, banks are. So, for the time being, the only meaningful thing an XRP holder can do to compel Ripple to make changes is by hitting them in the pocket-book by selling their XRP.
That said, one small thing Ripple could do is eliminate the default UNL and allow participants to select nodes on their own. The company even says it will do this, “eventually”:
“Eventually, Ripple intends to remove itself from this process entirely by having network participants select their own lists based on publicly available data about validator quality.”
There’s little stopping Ripple from making this change. But, the company has no reason to give up control, either.