In recent years, encryption algorithms have become widely adopted as a trusted means of securing data transmission, effectively safeguarding against unauthorized access. However, providing raw encrypted traffic directly to researchers can still pose privacy risks, making it essential to implement desensitization measures before sharing such data. The main challenge is to desensitize the encrypted traffic in a privacyprotection way while preserving as much of the original traffic characteristics as possible to ensure the downstream tasks run properly. In this paper, we propose MADE, an encrypted traffic desensitization framework based on Masked AutoEncoder (MAE). This framework transforms the traffic desensitization task into a classifier-guided image reconstruction problem, using supervised signals to achieve state-aware traffic reconstruction, which ensures the utility as well as the safety of the desensitized traffic for downstream tasks. Moreover, MADE is highly scalable, permitting users to customize which data attributes to preserve in the traffic being processed, simply by modifying the classifier. Experimental results reveal that our proposed desensitization framework not only outperforms baseline methods but also exhibits high applicability across various scenarios.

Driven by the flexibility, reliability, and scalability of microservice architecture, an increasing number of enterprises are decomposing monolithic applications into microservices. However, existing deep learning-based decomposition methods rely heavily on partition number selection, which, if unreasonable, can lead to frequent microservice communication and reduced performance. Moreover, manual partition suggestions not only decrease automation but also fail to adapt to rapid business iteration, and existing methods inadequately capture the relationship characteristics between application classes. To address these issues, this paper proposes a novel graph-based partitioning technique, GGRME. It constructs a system dependency graph through static, dynamic, and semantic analysis, and then employs a self-supervised gated graph neural network combined with cross-supervised optimization for community detection to automatically generate microservice decomposition results. Experiments demonstrate that GGRME outperforms benchmark methods in 65 % of tests, yielding superior microservice decomposition performance.