MagicScaler: Uncertainty-aware, Predictive Autoscaling
Summary: MagicScaler fuses multi-scale attention with Gaussian process regression to produce demand forecasts with quantified uncertainty, capturing scale-sensitive temporal patterns. An uncertainty-aware scaler uses a stochastic-constraint loss to trade off running cost vs QoS risk, validated on Alibaba clusters with superior results. (summarized by gpt-5-mini on Feb 09 2026)
Incoming Non-self Citations Over Time
Authors
- 1. Zhicheng Pan (Alibaba; East China Normal University)
- 2. Yihang Wang (Alibaba; East China Normal University)
- 3. Yingying Zhang (Alibaba)
- 4. Sean Bin Yang (Aalborg University)
- 5. Yunyao Cheng (Aalborg University)
- 6. Peng Chen (East China Normal University)
- 7. Chenjuan Guo (East China Normal University)
- 8. Qingsong Wen (Alibaba)
- 9. Xiduo Tian (Alibaba)
- 10. Yunliang Dou (Alibaba)
- 11. Zhiqiang Zhou (Alibaba)
- 12. Chengcheng Yang (East China Normal University)
- 13. Aoying Zhou (East China Normal University)
- 14. Bin Yang (East China Normal University)
BibTeX Citation
@article{pan_vldb23,
title = {{MagicScaler: Uncertainty-aware, Predictive Autoscaling}},
author = {Pan, Zhicheng and Wang, Yihang and Zhang, Yingying and Yang, Sean Bin and Cheng, Yunyao and Chen, Peng and Guo, Chenjuan and Wen, Qingsong and Tian, Xiduo and Dou, Yunliang and Zhou, Zhiqiang and Yang, Chengcheng and Zhou, Aoying and Yang, Bin},
journal = {PVLDB},
series = {{VLDB} '23},
volume = {16},
number = {12},
pages = {3808--3821},
doi = {10.14778/3611540.3611566},
url = {https://doi.org/10.14778/3611540.3611566},
year = {2023}
}
Incoming Citations (Sorted by Pagerank)
Showing 10 of 10 citing papers.
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 10 of 10 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 5,498 | SimpleTS: An Efficient and Universal Model Selection Framework for Time Series Forecasting | 2023 | VLDB |
| 2 | 4,398 | Real-time Workload Pattern Analysis for Large-scale Cloud Databases | 2023 | VLDB |
| 3 | 11,151 | Flux: Decoupled Auto-Scaling for Heterogeneous Query Workload in Alibaba AnalyticDB | 2024 | SIGMOD |
| 4 | 4,049 | Resource Elasticity for Large-Scale Machine Learning | 2015 | SIGMOD |
| 5 | 4,987 | Releasing Cloud Databases from the Chains of Performance Prediction Models | 2017 | CIDR |
| 6 | 1,092 | Automated Demand-driven Resource Scaling in Relational Database-as-a-Service | 2016 | SIGMOD |
| 7 | 9,943 | Eigen+: Memory Over-Subscription for Alibaba Cloud Databases | 2025 | SIGMOD |
| 8 | 10,695 | Oceanus: Enable SLO-Aware Vertical Autoscaling for Cloud-Native Streaming Services in Tencent | 2025 | SIGMOD |
| 9 | 10,694 | Managed Resource Scaling in Amazon EMR | 2025 | SIGMOD |
| 10 | 11,291 | OptScaler: A Collaborative Framework for Robust Autoscaling in the Cloud | 2024 | VLDB |