Pharos Network is a high-performance modular Layer 1 public blockchain, with core advantages in scalability, interoperability, and security. Pharos is dedicated to providing efficient execution, consensus, and customized support for the next generation of blockchains, boasting a processing capacity of 50,000 transactions per second (50K TPS), a throughput of 2 Gigagas, and DP level 5 performance (the highest standard among current EVM-compatible chains).

This article will deeply analyze the architectural design of Pharos Network from the following dimensions: consensus mechanism, execution model, storage innovation, network layer optimization, pipeline efficiency, and modular SPN framework.

Pharos Network Architecture Analysis

Pharos Network adopts a modular and customizable layered architecture design, achieving the dual goals of high performance and flexible upgrades through the decoupling of the execution layer, consensus layer, settlement layer, and data availability layer. Below, we will analyze the core technical design layer by layer.

• Network Layer

- Using an optimized P2P Gossip protocol for low-latency message transmission.
- Built-in cross-SPN (Special Processing Network) communication channel

• Consensus Layer

- Based on the adaptive asynchronous BFT consensus protocol, it combines high fault tolerance with low latency characteristics.
- Transaction sorting optimization algorithm, reducing the load on validation nodes and improving efficiency.

• Execution Layer

- Dual Virtual Machine Support: EVM Compatible Layer + WASM High-Performance Execution Environment
- Equipped with and supporting LLVM-based execution engines, parallel transaction processing, optimistic execution, and SALI (Smart Access List Inference)

• Settlement Layer

- Verify the cross-chain re-staking mechanism of nodes to achieve shared security within the ecosystem.
- Sub-second transaction final confirmation in multi-execution environments

• Data Availability and Storage Layer

- Pharos Store: An efficient blockchain storage model supporting multi-version control
- Using ADS push compression technology and Delta encoding, dynamic I/O optimization, to reduce disk read and write overhead.

• SPN (Special Processing Network) Framework

- SPN allows developers to customize the blockchain execution layer and freely configure the consensus model.
- Semi-independent operation mode: while enjoying the security and data of the mainnet, maintain independent scalability.

Next, we will delve into the key technological implementations and innovative designs at various levels.

Network Layer

The network layer is the foundation of all communications in the system, utilizing an optimized P2P Gossip protocol to efficiently disseminate blocks, transactions, and consensus messages. The protocol features adaptive bandwidth allocation capabilities, allowing it to dynamically adjust resources based on real-time network load, ensuring high-speed transactions and data transmission even under high-pressure conditions. Additionally, this layer supports cross-SPN communication channels, providing low-latency interoperability support for different execution environments.

Pharos Consensus Mechanism

Pharos Network adopts an innovative consensus design that combines BFT fault tolerance mechanisms with asynchronous execution and fast path commitment protocols, enhancing system throughput, transaction fairness, and final confirmation speed. Unlike traditional BFT implementations, Pharos does not require a fixed leader role and round-based communication, but allows multiple validating nodes to initiate, verify, and commit transactions in parallel, effectively avoiding the performance bottleneck caused by a single leader.

This consensus mechanism also implements dynamic transaction sorting, which not only ensures the fairness of transactions and reduces deviations in the block creation process but also maintains the consistency of global transaction order in cases of significant network fluctuations. Notably, the fast path execution mechanism allows transactions that meet specific conditions to bypass the complete consensus process and directly enter the confirmation stage, thereby reducing the average confirmation time and improving the overall throughput of the system. This well-designed consensus mechanism enables the Pharos Network to support high throughput and low latency stable operation on a global scale.

Main Features:

• Asynchronous execution architecture
• Dynamic Trading Sorting
• Fast Path Execution
• Optimized leader rotation and voting mechanism

Pharos Execution Layer

The core of the Pharos execution layer lies in its innovative execution engine, which provides developers with great flexibility through a dual virtual machine environment (EVM + WASM). The execution engine combines LLVM-based intermediate representation (IR) conversion and speculative parallel processing technology, employing multiple performance optimization measures, including opcode fusion, register promotion, and speculative execution. These optimization techniques enable smart contracts to execute in parallel while ensuring the determinism of execution results.

To further enhance execution efficiency, Pharos has introduced Smart Access List Inference (SALI) technology. SALI can predict the state entries that a contract will access through static or dynamic analysis, enabling the execution engine to group transactions with non-overlapping state access patterns for parallel execution, thereby avoiding state access conflicts. Additionally, this technology can preload contract state objects to accelerate execution speed.

Pharos also provides Ph-WASM, a WASM runtime environment specifically designed for blockchain. Ph-WASM not only supports deterministic high-speed execution but also natively supports programming languages such as Rust and Go, enabling developers to build high-performance, portable dApps and AI models directly on the chain.

Main Features:

• Support for optimistic execution and transaction batching
• Smart Access List Inference (SALI) Technology
• Ph-WASM environment supporting multi-language smart contracts

Pharos Store Storage System

Pharos Store is a scalable, high-performance blockchain storage system specifically designed to address the issues of state bloat and low I/O efficiency. This innovative data storage engine is designed to support verified versioned state storage. The system employs two core technologies: Delta encoding multi-version Merkle tree (DMM-Tree) and log-structured versioned page storage (LSVPS), effectively eliminating the efficiency problems present in traditional systems based on Merkle Patricia Trie.

Compared to the trie storage model of Ethereum, Pharos Store can achieve faster state retrieval, supports parallel Merkle processing, and reduces overall storage overhead by 80%. The system significantly reduces write amplification by only storing modified state incremental data and using a version-based indexing mechanism, resulting in an I/O throughput increase of up to 15.8 times.

Pharos Store adopts an append-only storage model, which ensures data security during concurrent writes and allows for quick recovery after node crashes. In addition, the system is capable of generating efficient verification proofs for light clients and supports intelligent hierarchical storage management for hot and cold data.

Main Features:

• Delta encoding multi-version Merkle tree (DMM-Tree) technology
• Log Structure Versioning Page Storage (LSVPS) Solution
• Lightweight SPN Asset Transfer Protocol

Pharos Pipeline

The Pharos pipeline adopts a multi-stage parallel execution architecture, effectively addressing the performance bottlenecks caused by synchronous state execution and sequential block validation in traditional blockchain systems. By parallelizing key processes such as execution, Merkleization, and state confirmation, it enhances the utilization efficiency of CPU, disk I/O, and network resources.

Main Processing Stage:

• Parallel Block Processing:

- Concurrent processing execution, Merkleization, and state updates

• Dynamic Resource Scheduling:

- Allocate CPU, I/O, and network resources based on real-time demand

• Flexible Confirmation Mechanism:

- Implement sorting, trading, and block finalization at different depths.

Each processing stage of the Pharos pipeline supports the dynamic allocation of resources based on real-time throughput demand. The system innovatively designs a flexible final confirmation layer, enabling dApps and clients to obtain early confirmation of transaction ordering and execution results before the block is fully confirmed. This feature is particularly important for high-frequency trading, DeFi, and other applications with high real-time requirements.

Through its intelligent adaptive scheduling algorithm and efficient resource coordination mechanism, the Pharos pipeline can achieve a throughput of over 200,000 TPS under a 64-core execution framework, while reducing latency by 30%-50% compared to standard blockchain pipeline models.

Pharos SPN architecture

Pharos SPN (Special Processing Network) is an execution environment tailored for specific applications, which are deeply integrated into the Pharos mainnet architecture while maintaining an independent operational logic and configuration system. Each SPN has a complete independent execution engine, a validation node cluster, a re-staking incentive mechanism, and a governance rules framework.

SPN architecture

• SPN Manager:

- Maintain lifecycle management, node registry, and governance rules execution.

• SPN Adapter:

- Promote cross-SPN communication and message verification

• Re-staking and shared validator security:

- Allow validators to re-stake assets into multiple SPNs.

SPN is very suitable for compute-intensive tasks, such as zero-knowledge machine learning proofs (zkML proofs), secure multi-party computation (SMPC), AI model training, and DeFi derivatives. They support mandatory inclusion and emergency mechanisms, ensuring that transactions are unreviewable while safeguarding users’ asset autonomy.

In terms of interoperability, SPN also interacts through the cross-SPN interoperability protocol, enabling atomic execution and data sharing across SPN and the main chain. Validators can re-stake $stPHRS into any SPN and receive corresponding $rstPHRS tokens, which can be used for governance, delegation, and liquidity mining.

Cross-SPN Interoperability

• Cross-SPN Messaging Protocol:

- Achieve seamless collaborative execution across multiple blockchain layers

• Resistance to censorship and emergency mechanisms:

- Protect users from mandatory trading reviews

Interoperability and cross-chain communication

Pharos Network natively supports cross-chain applications from the very beginning, providing two major communication protocols—internal (interactions within SPN) and cross-chain communication (connecting with external blockchains). The cross-SPN messaging framework ensures trustless message relaying between SPNs, using proofs and consensus signatures to verify authenticity.

Main features:

• Cross-chain message passing
• Cross-SPN Data Synchronization
• Multiple VM Smart Contract Interaction

These features provide key support for applications such as cross-chain DeFi, settlement of real-world assets (RWA), oracle networks, and collaboration of AI models across execution environments.

Overall, Pharos Network redefines the Layer-1 blockchain infrastructure standard by integrating modular design, high-performance processing, parallel computing, and composability at every level of our architecture. The combination of a unique adaptive consensus mechanism, a parallel execution engine, a multi-virtual machine architecture, a high-speed storage system, and a modular SPN framework enables developers to build applications with high scalability, censorship resistance, and security.

Pharos Network is dedicated to seamlessly connecting traditional finance (TradFi) with decentralized finance (DeFi), providing institutional-grade blockchain infrastructure support for AI-driven decentralized applications, private financial transactions, and scalable DeFi ecosystems.

Statement:

1. This article is reproduced from [TechFlow] The copyright belongs to the original author [TechFlow] If you have any objections to the reprint, please contact Gate Learn TeamThe team will process it as quickly as possible according to the relevant procedures.
2. Disclaimer: The views and opinions expressed in this article are those of the author alone and do not constitute any investment advice.
3. Other language versions of the article are translated by the Gate Learn team, unless otherwise stated.GateUnder such circumstances, it is prohibited to copy, disseminate, or plagiarize translated articles.