Lightweight Logging over the Blockchain for Data-Intensive Applications

Yuzhe Tang, Zihao Xing, Cheng Xu, Ju Chen, and Jianliang Xu

Abstract

With the recent success of cryptocurrency, Blockchain’s design opens the door of building trustworthy distributed systems. A common paradigm is to repurpose the Blockchain as an append-only log that logs the application events in time order for subsequent auditing and query verification. While this paradigm reaps the security benefit, it faces technical challenges especially when being used for data-intensive applications. Instead of treating Blockchain as a time-ordered log, we propose to lay the log-structured merge tree (LSM tree) over the Blockchain for efficient and lightweight logging. Comparing other data structures, the LSM tree is advantageous in supporting efficient writes while enabling random-access reads. In our system design, only a small digest of an LSM tree is persisted in the Blockchain and minimal store operations are carried out by smart contracts. With the implementation in Ethereum/Solidity, we evaluate the proposed logging scheme and demonstrate its performance efficiency and effectiveness in cost saving.
Type
Conference paper
Publication
In 2nd Workshop on Trusted Smart Contracts at International Conference on Financial Cryptography and Data Security
Date
March, 2018
Note
Workshop Paper, accepted to appear
Links