LETUS: A Log-Structured Efficient Trusted Universal BlockChain Storage

被引:1
|
作者
Tian, Shikun [1 ]
Lu, Zhonghao [1 ]
Zhuo, Haizhen [1 ]
Tang, Xiaojing [1 ]
Hong, Peiyi [1 ]
Chen, Shenglong [1 ]
Yang, Dayi [1 ]
Yan, Ying [1 ]
Jiang, Zhiyong [1 ]
Zhang, Hui [1 ]
Jiang, Guofei [1 ]
机构
[1] Ant Grp, Blockchain Platform Div, Hangzhou, Peoples R China
关键词
Blockchain; data storage; Merkle trie; multi-versioning;
D O I
10.1145/3626246.3653390
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
With the evolution of Web3.0 and decentralized applications (Dapps), increasing business logic is running on the blockchain. Blockchain storage, as the core infrastructure supporting the increasing volume of data, plays a crucial role. However, the performance and cost of blockchain storages are suffering seriously, which inspires us to re-examine the design of blockchain storage based on the blockchain data characteristics. We propose LETUS, a Log-structured Efficient Trusted Universal Storage for blockchain, providing cryptographic tamper evidence with excellent performance and resource efficiency. (1) LETUS breaks the traditional two-layered architecture and pushes down the Authenticated Data Structure (ADS) into the storage engine to enable fine-grained I/O optimizations. (2) LETUS proposes DMM-Tree which is a novel ADS combining the functionalities of Merkle tree and delta-encoding, significantly reducing storage consumption. (3) LETUS adopts a version-based indexing schema and manages the large volume of pages generated by ADS in a page store indexed by a B-tree variant. (4) LETUS provides a universal solution for different blockchains, such as public blockchains like Ethereum, BNB Smart Chain and AntChain as a representation of consortium blockchains. LETUS has been deployed in AntChain commercial applications, such as NFT and digital torch ignition for 2023 Asian Games. Experimental results also show that with LETUS, AntChain can achieve up to 15.8x improvement in throughput and 80.3% storage cost saving, Ethereum can achieve up to 10.1x improvement in throughput and 75.0% storage cost saving.
引用
收藏
页码:161 / 174
页数:14
相关论文
共 50 条
  • [31] An Efficient Bulk Loading Approach of Secondary Index in Distributed Log-Structured Data Stores
    Zhu, Yanchao
    Zhang, Zhao
    Cai, Peng
    Qian, Weining
    Zhou, Aoying
    DATABASE SYSTEMS FOR ADVANCED APPLICATIONS (DASFAA 2017), PT I, 2017, 10177 : 87 - 102
  • [32] On-line rollback in log-structured file systems
    Matthews, R
    Kearns, P
    COMPUTER APPLICATIONS IN INDUSTRY AND ENGINEERING, 2003, : 11 - 16
  • [33] The LHAM log-structured history data access method
    Muth, P
    O'Neil, P
    Pick, A
    Weikum, G
    VLDB JOURNAL, 2000, 8 (3-4): : 199 - 221
  • [34] On Log-Structured Merge for Solid-State Drives
    Thonangi, Risi
    Yang, Jun
    2017 IEEE 33RD INTERNATIONAL CONFERENCE ON DATA ENGINEERING (ICDE 2017), 2017, : 683 - 694
  • [35] The log-structured merge-bush & the wacky continuum
    Harvard University, United States
    Proc. ACM SIGMOD Int. Conf. Manage. Data, (449-466):
  • [36] On Integration of Appends and Merges in Log-Structured Merge Trees
    Gong, Caixin
    He, Shuibing
    Gong, Yili
    Lei, Yingchun
    PROCEEDINGS OF THE 48TH INTERNATIONAL CONFERENCE ON PARALLEL PROCESSING (ICPP 2019), 2019,
  • [37] The Log-Structured Merge-Bush & the Wacky Continuum
    Dayan, Niv
    Idreos, Stratos
    SIGMOD '19: PROCEEDINGS OF THE 2019 INTERNATIONAL CONFERENCE ON MANAGEMENT OF DATA, 2019, : 449 - 466
  • [38] LosPem: A Novel Log-Structured Framework for Persistent Memory
    Li, Sumin
    Huang, Linpeng
    ACM JOURNAL ON EMERGING TECHNOLOGIES IN COMPUTING SYSTEMS, 2020, 16 (03)