Overview
Hyperledger Fabric is an enterprise-level blockchain framework implementation. It is one of the open-source projects under the Linux Foundation. It has a highly modular and configurable design. It has complete member management and governance measures and supports pluggable consensus protocols. For example, industries such as banking, insurance, health care, and supply chain can use various mainstream programming languages to develop smart contracts according to their own requirements and dock with their own businesses.
Hyperledger Fabric does not need to use traditional means such as issuing tokens, mining, and Proof of Work to encourage participants. Targeting the Consortium Blockchain Scenario, its differentiated design makes Fabric an underlying engine with high performance and wide recognition.
Features and Advantages
Hyperledger Fabric has the following features:
Modular Design
Pluggable sorting service, providing a unique sequence for the occurrence of transactions for numerous blockchain participants.
Pluggable membership management service, using a unified method to confirm the legitimate identities of participants on the blockchain.
Pluggable endorsement and verification policies that can be flexibly configured for upper-level applications to adapt to various business scenarios.
Isolated container environment that can support running smart contracts developed in various mainstream programming languages.
The ledger supports flexible configuration and supports using multiple databases as status data caches.
Blockchain with Access License
Traditional blockchains have no access control mechanism, and anyone can participate in them anonymously. To handle the untrustworthiness and malicious behavior caused by this scenario, proof-of-work (PoW) is often used for transaction consensus, and mining is used to encourage participants. This method has lower performance, high transaction costs, and huge energy consumption.
In fact, the blockchain across organizational accounts needs to complete access control, that is, each participant performs transactions and consensus with their identity known. In such cases, the behavior of enterprises on the blockchain can be audited and supervised, and participants can have a certain degree of trust rather than absolute trust. Against this background, we can flexibly select a consensus mechanism according to the actual situation, use the traditional crash fault tolerance mechanism (CFT) to gain higher performance, or use the Byzantine fault tolerance mechanism (BFT) to enhance the anti-attack of the blockchain network.
In permissioned blockchain mode, the risk of participants maliciously destroying the network through smart contracts is significantly reduced. Whether modifying network configuration, introducing new blockchain participants, or installing and deploying new contracts, verification must be carried out according to predefined recognition policies. This policy can easily identify various malicious parties and perform corresponding actions according to the governance model.
Smart Contract
Smart contracts in Hyperledger Fabric are called "chaincode". Smart contracts can be developed using mainstream development languages such as Go, NodeJS, and Java, and executed in the corresponding container environment.
In traditional blockchains, the execution of smart contracts uses a "sequential execution" architecture. Its execution result must be deterministic; otherwise, consensus may not be attained. To solve the non-deterministic problem, many platforms require smart contracts to be written using specific scripts or specific languages (for example, Solidity). This requires developers to be familiar with the development of specific languages; otherwise, it may bring various unpredictable errors. In addition, since multiple transactions are sequentially executed on all nodes without distinction, performance and scale are limited. All nodes must have complete protection measures to resist the impact caused by malicious contracts.
Fabric has introduced a new "pre-execution-sorting-verification" architecture, which can split any transaction into three steps to address the flexibility, scalability, performance, and privacy issues that the above "sequential execution" may bring. The operation steps are as follows:
1. Perform a transaction in pre-execution and check the correctness of its result. If correct, endorse/sign the result.
2. Sort batch transactions through the sorting service.
3. According to the specific verification policy, check the execution outcome of the transaction. If the check passes, apply it to the ledger.
Fabric supports developing smart contracts in multiple mainstream languages. Before a transaction participates in sorting, in conjunction with the global sorting service and the verification mechanism before final storage, the transaction is pre-executed and endorsed. Nodes are set flexibly to eliminate uncertainty.
Privacy Protection
In a public blockchain network without access control, PoW is used as the consensus algorithm by default, so transactions are executed on each node. This makes the smart contracts and the data being processed completely non-confidential. The code of each smart contract and the content of transactions are visible to all nodes in the network.
For many commercial applications, this lack of confidentiality design can bring many problems. For example, in the supply chain network, partner A can see the smart contract and transaction content in this way, obtain more discounts or promotions from it, and use it as a means to consolidate the relationship between both parties or promote additional sales. In the securities industry, investors do not want their target prices to be known by other competitors. To address such privacy requirements, blockchain needs to support multiple methods to meet the needs of various industries and audiences:
Data encryption, a simple and commonly used method. Over time and with the innovation of technology, historical encrypted data on the chain may still be cracked by powerful new type calculation tools.
Zero-Knowledge Proof (ZKP), an additional approach to solving such problems. However, designing a zero-knowledge proof model and computation for information requires a lot of time and computing power. Currently, it is only an optional means in small range scenarios. In the future, Fabric will design and support more privacy protection solutions based on zero-knowledge proof.
Leverage methods such as controlling the transmission and access authorization of key information. Only necessary nodes are aware of the core data, and other nodes have no visibility into this part of the data. Hyperledger Fabric has implemented channels and private data sets in this way.
Pluggable Consensus Algorithm
In Fabric, the sorting service is designed in a modularized way. Users with development capabilities can redesign and replace it.
In the currently available versions of Fabric, the CFT sorting service based on Kafka and Zookeeper is used to sort transactions. In subsequent versions, Fabric will implement the Raft sorting service in etcd/Raft and the fully distributed BFT sorting service for users to choose.
Open-Source Community