A high-performance, scalable distributed ledger utilizing a Directed Acyclic Graph (DAG) structure for transaction validation and consensus. It is implemented in OCaml and leverages the Avalanche consensus protocol to achieve high throughput, low latency, and security against Sybil attacks. The design focuses on decentralization, scalability, and resilience, making it a strong candidate for real-world applications.
- DAG-based Architecture: Transactions form a DAG rather than a linear chain, improving concurrency and scalability.
- Avalanche Consensus: A robust, metastable consensus mechanism optimized for high throughput and transaction finality.
- Sybil Attack Resistance: Utilizes staking mechanisms and network sampling to mitigate Sybil attacks.
- Asynchronous Transaction Processing: Transactions are validated in parallel without a single leader bottleneck.
- Efficient State Propagation: Optimized gossip protocol for rapid transaction dissemination across the network.
- Decentralization: Every member of the blockchain structure is able to access the entire distributed database.
- Distributed Consensus: Implements a consensus algorithm to ensure that events are consistently applied across all nodes in a distributed environment.
- Scalable Architecture: The system is designed to scale horizontally to handle high throughput and large datasets in a distributed environment.
- Transactions are stored in a DAG, where each transaction validates multiple past transactions.
- Avoids traditional bottlenecks seen in linear blockchain architectures.
- Ensures high throughput by allowing multiple transactions to be confirmed simultaneously.
Consensus is done using Avalanche consensus protocol, which is a high throughput, parallelizable protocol. The transitive voting optimizs transaction throughput and the subsampling ensures that the number of consensus messages a node sends during a query is always consistent regardless of sample size. This protocol also allows for configurable parameters improving overall system flexibility.
-
Nodes repeatedly query a small, randomly selected subset of other nodes to determine transaction validity.
-
Achieves consensus efficiently with low communication overhead.
-
Provides strong probabilistic guarantees of finality after a small number of rounds. The Avalanche consensus mechanism provides probabilistic finality with the following plugable parameters:
-
k: Number of peers to query (default: 20)
-
α: Quorum size (default: 15)
-
β: Decision threshold (default: 15)
- Nodes communicate using gRPC, ensuring efficient, high-speed message passing.
- Gossip protocol for rapid propagation of transactions and votes.
- Each transaction contains:
- Unique identifier
- Pointers to parent transactions
- Digital signatures for authentication
- Consensus metadata
- Transactions are considered finalized once they receive enough confidence from the network.
- Byzantine fault-tolerant (BFT) properties ensure resilience to malicious nodes.
- Nodes are resistant to Sybil attacks via staking and probabilistic sampling.
- Transaction Submission: A user submits a transaction.
- Parent Selection: The transaction references a small number of previous transactions in the DAG.
- Network Propagation: The transaction is broadcasted to peers using a gossip protocol.
- Sampling & Querying: Validators repeatedly query random node subsets to determine the transaction’s validity.
- Consensus Formation: Once a transaction accumulates enough confidence, it is considered finalized.
- Ledger Update: The finalized transaction is permanently added to the DAG.
- Global State Synchronization: Nodes maintain a consistent view of the DAG through periodic synchronization.
- Horizontal Scaling: More nodes improve network security and validation speed.
- Low Latency Consensus: Avalanche achieves sub-second finality for most transactions.
- Reduced Overhead: Efficient querying reduces the cost of network communication.