AgentEmulator_
A Simulation & Experimentation Platform for the Trustworthy Infrastructure of AI Agents
AgentEmulator is an open-source platform initiated by HuangLab (Sun Yat-sen University) for researching and building trustworthy infrastructure for AI agents. Grounded in blockchain as the trusted layer for behavior records and settlement, it enables researchers and students to experiment with agent identity, behavior auditing, payments, incentives, and governance. It is built on BlockEmulator-X, the successor of our widely adopted BlockEmulator testbed used by researchers in 100+ countries and regions.
AgentEmulator v1.0 (released Sept. 21, 2026) delivers an end-to-end pipeline: trace-driven behavior replay, on-chain DID registration & revocation, per-transaction payments settled on the blockchain, and automatic result visualization.
TestbedBlockchain Experimentation Platform
Testbeds, Datasets, and Codes used in our papers. Welcome to use them freely and cite our publications.
* BlockEmulator *
We made our experimental tool for blockchain research open-source. This tool is named BlockEmulator.
A. Visit our Websites of BlockEmulator
B. Introduction to BlockEmulator
- Initiated by HuangLab (a research group in the School of Software Engineering, Sun Yat-sen University, China), BlockEmulator is a blockchain testbed that enables researchers to verify their proposed new protocols and mechanisms. It supports multiple consensus protocols, particularly the cross-shard mechanism.
- The main purpose of this testbed is to help users (researchers, students, etc.) quickly verify their own blockchain consensus protocols and blockchain-sharding mechanisms. BlockEmulator is designed as an experimental platform that adopts a lightweight system architecture.
- It simplifies the implementation of industrial-class blockchains since BlockEmulator only implements the core functions of a blockchain, including the transaction pool, block packaging, consensus protocols, and on-chain transaction storage. It also supports common consensus protocols, such as Practical Byzantine Fault Tolerance (PBFT) and Proof of Work (PoW).
- In particular, BlockEmulator offers the system-level design and implementation for blockchain sharding mechanisms. For example, the cross-shard transaction mechanisms implemented by BlockEmulator include the following two representative solutions, i.e., i) Relay transaction mechanism proposed by Monoxide (NSDI’2019), and the BrokerChain protocol proposed by BrokerChain (INFOCOM’2022).
- BlockEmulator is oriented toward blockchain researchers because it provides a blockchain experimental platform for quickly implementing their own algorithms, protocols, and mechanisms. It also offers very helpful functions for researchers to help them collect experimental data, facilitating their experimental charts.
C. Papers related to BlockEmulator
The following papers from HuangLab’s publications have adopted BlockEmulator as an experimental tool.
- “Scaling Blockchain State Management via Heterogeneous Low-Resource Followers,” IEEE Transactions on Parallel and Distributed Systems (TPDS), Oct. 29, 2026. TPDS CCF-A 计算机系统结构顶刊 公众号介绍文章
- “Feather: A Two-Phase Propose-and-Endorse Consensus Protocol for Asynchronous Blockchains,” IEEE Transactions on Computers (TC), early-access, pp. 1-13, 2026. TC CCF-A top-tier Journal 公众号介绍文章
- “ShardCutter: A Protocol Achieving Scalability and Load Balance across Blockchain Shards,” IEEE Transactions on Networking (ToN/TNet), vol. 34, pp. 3393-3407, 2026. ToN CCF-A top-tier Journal 公众号介绍文章
- “LiquidityPool: Game-Theoretic Analysis of Stakeholder Revenue in Ranking-Dependent DeFi,” The Web Conference 2026 (WWW’26), June 2026, Dubai, UAE. WWW CCF-A top-tier conference 公众号介绍文章
- “BrokerChain: A Blockchain Sharding Protocol by Exploiting Broker Accounts,” IEEE/ACM Transactions on Networking (ToN/TNet), vol. 33, no. 4, pp. 1930-1945, 2025. ToN CCF-A
- “Justitia: An Incentive Mechanism towards the Fairness of Cross-shard Transactions,” IEEE International Conference on Computer Communications (INFOCOM), May 19-22, 2025, London, United Kingdom. INFOCOM CCF-A 公众号介绍文章
- “Account Migration across Blockchain Shards using Fine-tuned Lock Mechanism,” IEEE International Conference on Computer Communications (INFOCOM), 20-23 May 2024, Vancouver, Canada. INFOCOM CCF-A 公众号介绍文章
- “Broker2Earn: Towards Maximizing Broker Revenue and System Liquidity for Sharded Blockchains,” IEEE International Conference on Computer Communications (INFOCOM), 20-23 May 2024, Vancouver, Canada. INFOCOM CCF-A 公众号介绍文章
- BrokerChain: A Cross-Shard Blockchain Protocol for Account/Balance-based State Sharding INFOCOM 2022 CCF-A PDF ⤓
- Achieving Scalability and Load Balance across Blockchain Shards for State Sharding SRDS 2022 CCF-B PDF ⤓
- tMPT: Reconfiguration across Blockchain Shards via Trimmed Merkle Patricia Trie IWQoS 2023 CCF-B PDF ⤓
- MVCom: Scheduling Most Valuable Committees for the Large-Scale Sharded Blockchain ICDCS 2021 CCF-B PDF ⤓
- Scheduling Most Valuable Committees for the Sharded Blockchain IEEE/ACM ToN 2023 CCF-A PDF ⤓
Datasets & CodesOpen Research Artifacts
Dataset & Codes for Predicting Machine Failures
Background: This dataset is to implement the failure prediction using machine learning methods and AI approaches such as SVM, random forest, or deep learning algorithms. Besides the original dataset, I also provide two reports written by two visiting students when they performed a visiting-study in my lab in July 2019.
Huawei Huang, and Song Guo, “Proactive Failure Recovery for NFV in Distributed Edge Computing“, IEEE Communications Magazine, vol. 57, no. 5, pp. 131-137, March 2019.
Dataset & Codes for Predicting Server Failures
Background: This dataset is used to predict the failures of server machines that occurred in a datacenter. The related published papers are as follows.
Huakun Huang, Lingjun Zhao, Huawei Huang, Song Guo, “Machine Fault Detection for Intelligent Self-Driving Networks“, IEEE Communications Magazine, Vol. 58, Issue No. 1, pp. 40-46, January 2020. [RG-Page]
Huakun Huang, Shuxue Ding, Lingjun Zhao, Huawei Huang, et al., “Real-Time Fault-Detection for IIoT Facilities using GBRBM-based DNN“, IEEE Internet of Things Journal, Oct. 21, 2019. DOI: 10.1109/JIOT.2019.2948396 [RG-Page]
Dataset & Codes of MVCom ICDCS 2021
Background: This code-and-dataset shows how we implement the algorithms used in our paper, including the proposed SE algorithm and other 3 baselines (SA, DP, WOA).
Huawei Huang, Zhenyi Huang, Xiaowen Peng, Zibin Zheng, Song Guo, “MVCom: Scheduling Most Valuable Committees for the Large-Scale Sharded Blockchain“, ICDCS, July 2021. [RG-Page & PDF]
According to the requests of some readers, I provide all codes of the algorithms used in this paper such as the SE (stochastic exploration) algorithm (which is a Markov-based Algorithm, MA), SA algorithm, DP algorithm, and WOA algorithm. Some figure-plotting codes and partial data are also included. (Updated on Nov. 7, 2022, by Huawei HUANG)
Datasets of ContextFL ICDCS 2022
Background: This dataset is used to predict the CPU / Network Connection / App usages of mobile users when they are using their smartphones. The related published paper is as follows.
Huawei Huang, Ruixin Li, Jialiang Liu, Sicong Zhou, Kangying Lin, and Zibin Zheng, “ContextFL: Context-aware Federated Learning by Estimating the Training and Reporting Phases of Mobile Clients“, in proc. of IEEE International Conference on Distributed Computing Systems (ICDCS), 2022. [RG-Page] [公众号介绍文章]