Blockchain governance: how Decred iterates upon Bitcoin
It is in the nature of open source software to evolve with collaboration and fork under contention. The same is true for blockchains and cryptocurrencies. As blockchains face the hurdle of mass adoption, coordination and collective decision making can contribute to the success of a project. While Bitcoin’s protocol is the standard for most cryptocurrencies as it relates to network consensus, it is arguably deficient as a mechanism for achieving social consensus among stakeholders.
This article depicts how achieving social consensus on intra-protocol changes can lead to disruptive events on the Bitcoin (or Bitcoin Cash!) network, and examines how Decred’s protocol allows these changes to occur in a more efficient and stable manner. This article will use diagrams to illustrate which entities can influence protocol evolution and shed some light on the high-level power structures that govern most blockchain based cryptocurrencies.
- Social consensus = Peers in agreement
- Intra-protocol changes = Changes to the consensus rules
- Consensus failure = A change to the rule-set that results in chain splits
- Sustained chain split = When there are multiple competing versions of the same blockchain for an extended period of time
Entities in the visual diagrams:
Diagrams are used in this article to illustrate the power dynamics within a cryptocurrency network. These diagrams are made up of the following entities:
- Mining Nodes refer to nodes that choose to secure the network because they see value in running a particular software implementation, for example Bitcoin Core. As networks become competitive, miners could congregate in mining pools to get consistent payouts.
- Staking Nodes (specific to Decred) refer to nodes that choose to time-lock their coins by running a particular software implementation. These nodes are able to weigh in on validating blocks, vote on proposed intra-protocol changes and earn a portion of the block reward.
- Economic Nodes are exchanges, merchants, and businesses where the value of the network and its token is derived from. Economic nodes rely on software implementations from a corporation, foundation and a community of open source developers for updates and patches.
- Users refer to the people using the network token to transact value. They can also provide feedback to developers and incentivize businesses to adopt the cryptocurrency as a payment method.
- Developers are usually grouped in teams that work on the most dominant software implementation(s) of the network. They provide other entities with software updates and formulate a roadmap.
Examining Bitcoin’s PoW based governance
It is possible to create a diagram of the power dynamics that govern a pure PoW cryptocurrency (such as Bitcoin) using the described entities. The diagram below shows the model of a PoW network in equilibrium. None of the entities have complete control and if anyone is removed, the value proposition of the network changes. While in equilibrium, this model works well.
When disputes arise over changes to the protocol, things can get complicated very quickly. There are a good deal of examples that such disruptive events occur, for example the recent Bitcoin Cash war and the historical Bitcoin forks. The pure PoW model is arguably deficient as a mechanism for achieving social consensus among stakeholders.
In pure PoW governance there is no tamper-proof pathway for stakeholders to resolve disputes. The resolution pathway is either providing feedback to the developers via forums/websites/polls or through arbitration by the market itself. One of the biggest issues with the reliance on market forces is that it allows for the possibility of consensus failure.
Let’s examine this in more detail (summarized in the diagram below):
- Disputes occur over intra-protocol change X. Social consensus is gauged (via polls, meetings, social media). A competing software implementation emerges, backed by developers with alternative sources of funding.
- Developers provide updates, stating their intent to resolve problem X. Some users show support to the new developers and upgrade to the competing software implementation (containing the new rule-set).
- Signaling can occur to help convince miners, however this is not a requirement (see UASF/UAHF).
- A few mining pools can decide to support the opposing implementation; in this case a hash power split occurs.
- If this is a contentious change, miners can start a hash power war causing chain splits, reorganizations, or even sustained chain splits.
- Since there is no way to formalize intra-protocol decisions, the only option is resolution through arbitration by the open market.
- Economic nodes have to update their software and adapt to the new status quo. In the case of a sustained chain split, they have to decide whether they add support for the new fork(s) of the cryptocurrency.
By looking at the above model, it can be observed that:
- Users do not have a tamper-free pathway to participate in intra-protocol decision making. Conflicts are resolved via (market) arbitration.
- Mining nodes can cause consensus failure by supporting alternative rule-sets, since they are not bound by any agreements.
Consensus failure in PoW networks
Since PoW networks have no formal mechanisms to resolve intra-protocol disputes, consensus failure is likely to occur. As defined in the introduction, consensus failure will result in a chain split, often causing the blockchain to reorganize. This could mean a loss of transaction finality, one of the most important features of cryptocurrencies.
Usually, the higher the number of confirmations, the more confidence one can have in the permanence of a transaction. During consensus failure, however, no amount of confirmations are safe. The branch on which your transaction takes place has to win and cause a reorganization. During consensus failure it is dangerous for transactions to occur as there are increased risks of losses.
The reality of this danger is demonstrated by Jimmy Song in his article on UASF BIP148 Scenarios and Game Theory:
You should be very careful about transacting after August 1. There are scenarios where your transaction will get wiped out no matter which fork you choose. Both chains have some wipeout risk and I imagine most people will be hesitant to do anything until there’s some resolution.
Apart from the loss of transaction finality, consensus failure could also result in sustained chain splits. In this scenario the network forks into two different cryptocurrencies that share the transaction history up to a certain point. In his article on the dangers of sustained chain splits, Noah Pierau explains the risks of sustained chain splits in more detail.
It should be abundantly clear that resolving disputes in pure PoW networks is a messy, sub-optimal and often damaging process.
Examining Decred’s Hybrid PoW/PoS governance
Decred iterates upon pure PoW networks in several ways. Most importantly, it introduces staking nodes, providing a pathway for users to weigh in on intra-protocol decisions. Apart from improving blockchain governance, the hybrid model augments the security of network.
Just like in a pure PoW network, the PoW miners provide computational power to find blocks. However, each block they produce needs to be checked by the PoS nodes. This means that PoW miners are prevented from extending the blockchain without explicit approval of stakeholders. Users can become stakeholders and formally participate in the protection and evolution of the protocol through time-locking their DCR coins in exchange for tickets.
An internal ticket market rewards users with a portion of the block reward as an incentive to participate. For intra-protocol changes, PoS nodes are able to vote on their preferred version of the blockchain. Changes with >75% of the votes will automatically activate. It is in the best interest of stakeholders to make informed decisions, as they are directly exposed to the downsides of bad decision making (their coins are locked). More details on this unique voting system can be found in the Decred documentation.
Benefits of the Hybrid approach
Decred’s hybrid protocol incentivizes users to become stakeholders and keeps miners compliant. In a scenario where hashing power is split, miners cannot purposely or accidentally cause chain splits since they are unable to build on top of unapproved blocks (visualized in the image below).
It is important to note that the issues affecting pure PoW blockchains can theoretically occur in Decred. However, its hybrid protocol is designed to be highly resistant to them. A good explanation of Decred’s fork resistance can be found in this Reddit post by Dave Collins.
In addition to fork resistance, it is substantially more expensive to attack Decred compared to a pure PoW network such as Bitcoin. Here is an article on how expensive it is to conduct a majority attack: Decred’s hybrid protocol, a superior deterrent to majority attacks.
Cryptocurrencies operate on networks run by sovereign entities. These actors are able to agree on the current consensus rules, although their opinions may clash if there are intra-protocol changes required. Gauging social consensus in Bitcoin is not tamper-proof and entails severe coordination issues. In order to resolve disputes around intra-protocol changes, most PoW networks face the risks of reorganizations and sustained chain splits.
Decred iterates upon Bitcoin by solving the coordination problems that plague pure PoW blockchains. By allowing users to obtain a seat at the table through PoS voting, anyone with skin-in-the-game can participate in intra-protocol decision making. The hybrid approach also keeps miners compliant since blocks are checked before they are added to the chain. With this unique blockchain governance model, Decred is ready for the future.
EDIT: in this article we have examined Decred’s on-chain governance model. Decred also has off-chain governance, in the form of Politeia. Read more about this in the introduction to Decred governance.
Credits (slack/matrix names): haon, bee, davecgh, s_ben, and others who have provided feedback and answered questions.