ICUBE5: A Conceptual Multidimensional Benchmarking Framework for AI-Native Secure 6G Edge Architectures

The trend towards sixth-generation (6G) networks requires a new network architecture capable of incorporating artificial intelligence (AI), distributed edge intelligence, security, governance, adaptability, and sustainability into increasingly heterogeneous computing and communication environments. Present methods to achieve such network architectures address only the optimisation, security, governance, or energy-efficiency issues separately, making it difficult to assess the maturity of a particular architecture across all of these dimensions. This paper proposes an AI-native secure edge architecture and introduces ICUBE5, a conceptual multidimensional benchmarking framework for assessing AI-enabled 6G edge architectures across five dimensions: AI Maturity, Security Resilience, Governance Alignment, Adaptive Stability, and Sustainability Efficiency. The framework provides a structured approach for normalising dimension-specific indicators and combining them through configurable weighting schemes that can be adapted to deployment priorities. Illustrative weighting scenarios are presented for general-purpose, industrial automation, smart-city, and energy-constrained edge environments to demonstrate how contextual priorities can influence the composite ICUBE5 score. A comparative analytical illustration involving cloud-centric, edge-centric, and AI-native secure edge architectural archetypes further demonstrates the application of the framework. Under the illustrative assumptions used in the analysis, the proposed AI-native secure edge architecture produces the highest composite ICUBE5 score, while the results remain conceptual rather than empirical measurements of deployed systems. The framework also incorporates security and governance considerations associated with adversarial machine learning, federated learning, trustworthy AI, and emerging regulatory requirements. ICUBE5 is therefore positioned as a benchmarking construct rather than a validated performance index. Future research will validate the framework through simulation and experimental testbeds, using measurable network, security, governance, adaptive-control, and energy-efficiency indicators. The proposed framework provides a basis for structured comparison and context-sensitive evaluation of emerging AI-native 6G edge architectures.