Abstract: With the rapid development of the digital economy, data has evolved from an auxiliary resource that supports business operations into a critical production factor with independent economic value, and has been deeply integrated into government systems, industrial Internet platforms, and a wide range of intelligent applications. As data factorization continues to advance, data processing chains and system architectures are becoming increasingly complex, resulting in substantial changes in the exposure surfaces and propagation paths of security risks. In particular, in typical environments characterized by multi-source heterogeneity, high-frequency mobility, and cross-domain computing, data continuously flows across heterogeneous systems and processing stages, rendering traditional security control mechanisms based on static boundaries increasingly inadequate. Consequently, conventional security boundaries are becoming blurred and even ineffective in dynamic data processing scenarios. Under such conditions, security paradigms that rely solely on boundary protection and static hardening are no longer sufficient to characterize or mitigate the evolving risks arising throughout the data processing workflow. Therefore, it is imperative to systematically identify and analyze security risks at different stages from the perspective of the full data lifecycle, and further establish a set of key technologies to support lifecycle-oriented security risk prevention and control.