Optimizing Blockchain Storage: Smart Chain's Path-Based Storage System

Introduction

Operating a full node represents a significant responsibility that demands continuous attention and technical expertise. The traditional approach to node management has presented considerable challenges, particularly regarding offline pruning operations that have long plagued node operators across various blockchain networks.

Path-Based Storage Scheme, as introduced in Geth V1.13.0, offers a transformative solution to these persistent challenges. This innovative approach substantially enhances both performance and user experience by fundamentally changing how blockchain state data is stored and managed.

This comprehensive guide explores three critical aspects: the strategic reasons behind major smart chain's adoption of PBSS, the key performance improvements delivered by PBSS compared to traditional Hash-Based State Schemes, and the specific optimizations implemented in its PBSS implementation along with empirical testing results.

Why Did Smart Chains Adopt PBSS?

Before the introduction of PBSS, major EVM-compatible networks struggled with significant inefficiencies inherent in the hash-based storage scheme. The limitations of this approach became particularly acute due to unique operational characteristics: a remarkably fast 3-second block creation time and an exceptionally high gas limit of 140 million.

These specifications, while providing superior transaction throughput capabilities, created mounting storage pressures. The rapid accumulation of state data forced users to periodically halt their nodes for time-consuming offline pruning operations—a process exclusive to the hash-based state scheme that was introduced in Geth version 1.10.

PBSS addresses these critical concerns by introducing an online state prune feature that effectively reduces state bloat without requiring offline operations. Rather than forcing node operators to interrupt service, PBSS manages storage size dynamically through continuous background processes, delivering a seamless user experience. The performance improvements have been remarkable: the average time required to import a block has decreased from 698 milliseconds to 531 milliseconds, representing a substantial 24% reduction in block processing time.

Path-Based vs Hash-Based

The fundamental architectural difference between PBSS and the legacy hash-based model lies in their storage mechanisms. PBSS employs a path-based storage model that organizes trie nodes using encoded paths as keys, whereas the hash-based model relies on the content hash of individual trie nodes. This distinction creates significant cascading effects on system performance.

The path-based model's approach enables inline state data pruning and facilitates more efficient storage utilization of trie nodes. These improvements directly translate to enhanced system performance and reduced resource consumption.

The contrast in storage growth patterns is particularly illuminating. In hash mode, the state database expands rapidly at approximately 50GB weekly, reflecting the accumulation of historical state entries. Conversely, in path mode, storage growth exhibits a dramatically slower trajectory, increasing by less than 5GB per week. This represents a tenfold reduction in storage expansion rate, making path mode substantially more sustainable for long-term node operations.

Innovations in Smart Chain's PBSS

While PBSS provides a robust foundation, high-performance smart chains faced the unique challenge of managing an extraordinarily high transaction volume on their EVM infrastructure. To address this specific scenario, researchers and engineers implemented specialized adjustments and optimizations tailored to the distinct operational differences. These innovations directly target two significant challenges: the generation of empty blocks when validators operate under extreme transaction volumes exceeding 1,000 TPS, and notable performance fluctuations when processing large transaction batches.

Async Node Buffer

Innovation: High-performance chains have implemented an asynchronous node buffer strategy, fundamentally diverging from conventional synchronous approaches. Traditional networks with leisurely block times permit a synchronous node buffer process that pauses the main workflow when necessary to complete disk flushing and cache synchronization operations during block creation cycles.

Networks with rapid 3-second block intervals present a critical constraint: the network cannot tolerate the latency inherent in synchronous disk operations. To maintain optimal efficiency and prevent disruptions in continuous block generation, engineers have designed an asynchronous buffer at the disk layer that processes flush operations concurrently in the background, ensuring uninterrupted block production.

The asynchronous node buffer operates through the following process:

• The oldest differential layer commits its changes to the disk layer synchronously. Critically, the system does not wait for the node buffer to complete its flush to disk, even when buffer size exceeds predetermined thresholds.
• The node buffer associated with the disk layer performs database flushing operations continuously in the background, allowing other processes to proceed without delays or interruptions. These background operations maintain strict safety guarantees, protecting transaction integrity and state data consistency.
• The database continuously maintains complete snapshots of state data corresponding to specific historical block heights, ensuring comprehensive data consistency and availability for validation purposes.

Trie Node Cache Optimization

Innovation: High-performance chains have strategically adjusted trie node cache parameters to substantially improve system stability and eliminate unexpected performance variations. This optimization directly responds to challenges created by the asynchronous node buffer enhancement.

The increased frequency of disk flushing operations in the disk layer inadvertently elevated the likelihood of database compaction events. Database compaction, the resource-intensive process of merging and rewriting data to optimize storage and accelerate read performance, can significantly reduce system throughput and increase latency.

This challenge is addressed by doubling the buffer size in the disk layer, thereby reducing the frequency of compaction operations. A larger buffer capacity holds significantly more data before necessitating write or flush operations to the database, effectively reducing the total number of write cycles and data overwrites—the primary triggers for compaction events.

Complementing this strategy, an advanced memory management approach features a clean cache twice the size of the node buffer cache. This larger allocation substantially reduces the risk of data overwrites during synchronization between the node buffer cache and the clean cache. When the node buffer commits changes to disk through the clean cache, the expanded memory allocation ensures critical data remains readily accessible, proving essential for maintaining high throughput requirements and low-latency performance characteristics.

Disk Bandwidth Optimization

Innovation: High-performance chains have implemented optimized PebbleDB (an open-source database engine) to intelligently regulate flush and compaction operations, effectively preventing latency spikes that would otherwise degrade performance. This represents a significant upgrade from legacy system architectures.

Modern PBSS implementations replace traditional database systems with the optimized Pebble DB engine. Legacy systems operate without throttle mechanisms for flushes and compactions, consistently executing at maximum capacity and generating notable latency spikes for both write and read operations. Under high transaction volume conditions, these spikes create user-visible performance degradation.

In contrast, PebbleDB incorporates separate rate limiters for flushes and compactions, intelligently calibrating operation speed to actual requirements. This mechanism ensures operations proceed only as rapidly as necessary, preventing unnecessary strain on disk bandwidth and maintaining consistent performance under load.

PebbleDB has been increasingly adopted across EVM networks, with substantial effort invested in conducting comprehensive testing to fine-tune configurations specifically according to distinctive traffic patterns and transaction profiles, ensuring optimal performance for various operational contexts.

Testing Benchmarks

Comprehensive testing coordinated multiple validators with identical binary versions and standardized hardware configurations. One validator operated in hash mode using traditional database systems, while another operated in path mode using PebbleDB, with all other variables controlled.

Hardware configuration for pilot validators:

• AWS im4gn.4xlarge EC2 instance
• 16 CPU cores and 64 GB of memory
• 7TB Nitro SSD storage

Testing results conducted throughout 2024 and into 2025 demonstrated substantial performance improvements across critical metrics:

The empirical data reveals that path-based storage scheme achieves consistent and significant performance improvements across core operations. While this optimization approach utilizes greater disk bandwidth, PebbleDB's rate-limiting mechanisms effectively mitigate latency spikes that might otherwise impact system performance.

Conclusion

The implementation of Path-Based Storage Scheme represents a major milestone in blockchain infrastructure optimization. By replacing offline pruning with online state management and introducing specialized optimizations including asynchronous node buffers, enhanced cache strategies, and intelligent disk bandwidth management, high-performance smart chains have achieved substantial improvements in both storage efficiency and operational performance.

The 24% reduction in block import time, combined with the dramatically reduced storage growth rate (from 50GB to less than 5GB weekly), demonstrates that PBSS provides a scalable foundation for high-performance blockchain networks. The innovations developed—particularly the asynchronous node buffer and PebbleDB optimization—address real challenges inherent in managing high-throughput EVM-compatible chains and provide proven solutions for other blockchain networks facing similar constraints.

As blockchain networks continue to face mounting pressure from state bloat and performance demands, PBSS and specialized optimizations offer a practical pathway toward more efficient and sustainable blockchain infrastructure. These advancements position modern smart chains as technical leaders in EVM optimization and provide a compelling model for other EVM-compatible networks seeking to scale their operations while maintaining infrastructure stability and performance.

FAQ

What is BSC Storage? What are its main functions and purposes?

BSC Storage is a decentralized storage solution built on Binance Smart Chain, enabling efficient data storage and management for smart contracts. Its main functions include providing secure, scalable storage capacity, reducing on-chain costs, and accelerating transaction confirmation speeds for blockchain applications.

How to store and manage data on BSC Storage?

Download the BSC Storage SDK client and create a bucket. Upload files using the SDK, then access them via the designated URL format. Use organized folder structures and consistent naming conventions for efficient data management.

Is it safe to store data using BSC Storage? What risks should I be aware of?

BSC Storage offers secure data storage with encryption and access controls. Main risks include unauthorized access and data breaches. Enable strong encryption and strict permission management to ensure optimal security.

What are the advantages and disadvantages of BSC Storage compared to traditional cloud storage?

Advantages: Decentralized security, lower costs, user data ownership, censorship-resistant. Disadvantages: Lower accessibility, smaller ecosystem, requires technical knowledge, emerging technology with less stability.

How are BSC Storage costs calculated? What are the usage fees?

BSC Storage is permanently free with no usage fees. Only users with permissions to create scorecards, input data, or manage accounts require payment, charged monthly per user. Read-only access remains free and unlimited.

What types of files and data storage does BSC Storage support?

BSC Storage supports various file types including images, videos, documents, and backups. It efficiently handles large datasets and object storage for diverse data storage needs.