Abstract
The blockchain architecture with distributed and decentralized as the main features often face problems such as long transaction confirmation time, low system throughput, and waste of computing resources. These problems are exactly what the traditional centralized architecture can solve. In order to effectively take advantage of the high security, high reliability of the blockchain, and high efficiency of traditional centralized architectures, this paper presents an innovative hybrid blockchain approach, Hybrid-chain. This architecture consists of three node types: service center nodes for efficient transaction confirmation, dynamic transaction verification nodes participating in transactions, and random verification nodes responsible for fairness verification. Experiments show that compared with traditional blockchain architecture, Hybrid-chain reduces transaction sending time by 96%, shortens confirmation time by 95%, and reduces block information storage space by 79%.
Introduction
Traditional blockchains like Bitcoin face a "scalability trilemma." Because every node in the network must verify and store every single transaction, the system becomes sluggish. In peak periods, Bitcoin confirmation times can spike significantly, while centralized processors like Visa handle thousands of transactions per second. Conversely, while centralized systems are fast, they lack transparency and are susceptible to data manipulation. The Hybrid-chain architecture seeks to bridge this gap, providing the traceability of a blockchain without the bottleneck of global consensus.
Architecture and Methods
The Hybrid-chain model categorizes nodes into three distinct roles to streamline processing and reduce redundant verification:
- Service Center Nodes: A small cluster (which can be centralized or a mini-blockchain) that acts as the primary billing and coordination hub.
- Dynamic Transaction Verification Nodes: These are the actual parties involved in a transaction (sender and receiver). Unlike standard blockchains, these nodes only record and verify transactions relevant to them.
- Random Verification Nodes: Nodes selected at random to act as "auditors" to ensure fairness and prevent collusion between the service center and the transacting parties.
Mixed Consensus and Selective Storage
The system utilizes a mixed consensus approach, selecting mechanisms based on transaction characteristics (e.g., PoW for the central cluster to maintain security and PoS for verification nodes to ensure speed).
To address storage bloat, Hybrid-chain uses a "need-to-know" model. Instead of every node storing the entire multi-gigabyte history of the chain, individual nodes store their own transaction history and the cryptographic hashes of the blocks immediately before and after theirs to maintain the integrity of the chain.
Evaluation and Results
The researchers evaluated the architecture using p2pSim, simulating 6,000 nodes and 10,000 transactions. The results showed massive performance leaps over traditional blockchain architectures:
- Latency: Transaction sending time was reduced by 96%, and confirmation time dropped by 95% (from ~500 seconds down to ~18 seconds in the simulation).
- Storage: The total space occupied by the blockchain across all nodes was reduced by 79% (from 52,800 MB down to 11,003 MB).
- Synchronization: Block synchronization between nodes took only 2 seconds, compared to 30 seconds in the traditional model.
Key Insights and Limitations
The core innovation of Hybrid-chain is relevance-based verification. By limiting verification to the involved parties plus a few random auditors, the "network noise" is drastically reduced. The architecture acknowledges that total decentralization is often overkill for business applications, treating decentralization as a spectrum rather than a binary choice.
However, the "Service Center" remains a more centralized point of failure than a truly global network. Future work is required to explore the optimal number of random nodes needed to prevent localized attacks and to test the system under real-world network conditions involving varying node uptimes and malicious actors.
Summary updated