Spatial-Temporal Heterogenous Graph Contrastive Learning for Microservice Workload Prediction
Authors: Mohan Gao, Kexin Xu, Xiaofeng Gao, Tengwei Cai, Haoyuan Ge
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on two datasets, including MS dataset obtained from Ant Group, which is one of the world s largest cloud service providers, demonstrate the superiority of STEAM. |
| Researcher Affiliation | Collaboration | 1Mo EKey Lab of Artificial Intelligence, Department of Computer Science and Engineering, Shanghai Jiao Tong University Shanghai, China 2Ant Group, Hangzhou, China |
| Pseudocode | No | The paper describes methods through textual descriptions and mathematical equations (e.g., equations 1-8) but does not contain a clearly labeled 'Pseudocode' or 'Algorithm' block. |
| Open Source Code | No | The paper does not provide any explicit statement about releasing source code, nor does it include a link to a code repository. |
| Open Datasets | Yes | MS contains workload information for 46 microservices over a 10-day period. Ali 1 is sourced from the Alibaba Cluster Trace, it contains workload records for 148 machines over 8 days. In the data preprocessing, any outlier values below 0 are set to 0. The datasets are divided into training and testing sets in a 7:1 ratio along the time dimension, with the last 24 hours of the training set used as the validation set. 1https://github.com/alibaba/clusterdata/tree/v2018 |
| Dataset Splits | Yes | The datasets are divided into training and testing sets in a 7:1 ratio along the time dimension, with the last 24 hours of the training set used as the validation set. |
| Hardware Specification | Yes | All models are implemented using Py Torch and trained using the Adam optimizer on a NVIDIA A10 GPU. |
| Software Dependencies | No | All models are implemented using Py Torch and trained using the Adam optimizer on a NVIDIA A10 GPU. The paper mentions PyTorch but does not provide a specific version number, nor other software dependencies with versions. |
| Experiment Setup | Yes | The hyperparameters for STEAM are set as follows: the learning rate is 0.001, the batch size is 128. The dimension of representation d is 16. All similarity thresholds are set to the value of the top 10% Pearson coeffi-cients. The kernel size for the temporal convolutions is set to 3, with two convolutional layers. The temperature parameter ζ is set to 0.5. The weight coefficient λ is set to 0.9. |