In blockchain, decentralization refers to the transfer of control and decision-making from a centralized entity (individual, organization, or group thereof) to a distributed network. Decentralized networks strive to reduce the level of trust that participants must place in one another, and deter their ability to exert authority or control over one another in ways that degrade the functionality of the network.
Decentralization is not a new concept. When building a technology solution, three primary network architectures are typically considered: centralized, distributed, and decentralized. While blockchain technologies often make use of decentralized networks, a blockchain application itself cannot be categorized simply as being decentralized or not. Rather, decentralization is a sliding scale and should be applied to all aspects of a blockchain application. By decentralizing the management of and access to resources in an application, greater and fairer service can be achieved. Decentralization typically has some tradeoffs such as lower transaction throughput, but ideally, the tradeoffs are worth the improved stability and service levels they produce.
Decentralization should be applied where it makes sense. Just because it’s a blockchain application doesn’t mean it needs to be 100% decentralized. The goal of any blockchain solution is to deliver what the users of that solution need, and this may or may not include certain levels of decentralization. To better understand decentralized networks, the table below breaks out how decentralized networks compare to the more common centralized and distributed networks.
Centralized | Distributed | Decentralized | |
---|---|---|---|
Network/hardware resources | Maintained & controlled by single entity in a centralized location | Spread across multiple data centers & geographies; owned by network provider | Resources are owned & shared by network members; difficult to maintain since no one owns it |
Solution components | Maintained & controlled by central entity | Maintained & controlled by solution provider | Each member has exact same copy of distributed ledger |
Data | Maintained & controlled by central entity | Typically owned & managed by customer | Only added through group consensus |
Control | Controlled by central entity | Typically, a shared responsibility between network provider, solution provider & customer | No one owns the data & everyone owns the data |
Single Point of Failure | Yes | No | No |
Fault tolerance | Low | High | Extremely high |
Security | Maintained & controlled by central entity | Typically, a shared responsibility between network provider, solution provider & customer | Increases as # of network members increase |
Performance | Maintained & controlled by central entity | Increases as network/hardware resources scale up and out | Decreases as # of network members increase |
Example | ERP system | Cloud computing | Blockchain |
Each network architecture has its benefits and tradeoffs. For example, decentralized blockchain systems, unlike distributed systems, typically prioritize security over performance. So, when a blockchain network scales up or out, the network becomes more secure, but performance slows down because each member node must validate all data being added to the ledger. Adding members to a decentralized network can make it safer, but not necessarily faster.
Every blockchain protocol, decentralized Application (dApp), Decentralized Autonomous Organization (DAO), or other blockchain-related solution adopts varying levels of decentralization. The adoption level is typically based on the maturity of the solution, the time-proven reliability of its incentive models and consensus mechanisms, and the ability of the founding team to strike the right balance. For example, many DAOs have various components at different stages of decentralization: oracles (i.e., third-party services that provide smart contracts with external information) may be partly decentralized, smart contracts might be fully centralized, while the governance process for adjusting parameters is community-driven and decentralized.
On a broader scale, decentralized blockchain solutions are being explored and adopted by organizations of every type, size, and industry. Some notable examples include applications that provide immediate foreign or emergency aid to those who need it most, without the mediation of a bank, government or third-party entity. Or applications that give people the ability to manage their own digital identities and data. Today, social media platforms, companies, and other organizations sell this information without the individual seeing any benefit. A decentralized approach would help make it equitable for all.
Contura Energy, a leading U.S.-based coal supplier, has depended on an outdated letters of credit system to manage its international trade payments. These letters of credit, issued by an intermediary bank on behalf of its client, serve as a guarantee of payment for buyers. While this system is trustworthy, it’s also manually driven, slow, and highly inefficient.
Contura Energy understands the importance and value of digitizing and automating their letters of credit process. The challenge they face, however, is enabling mutual trust and verification between sellers and buyers. They are working with AWS on a decentralized, blockchain-based innovation that provides a more efficient, cost saving, and less risky system to manage international trade payments. This decentralized solution also increases transparency, giving all parties real-time visibility into the data and documentation.