A Multi-Layered Framework for Secure Distributed Computing in Heterogeneous Cloud–Edge Environments Using Adaptive AI Orchestration
DOI:
https://doi.org/10.63282/3117-5481/AIJCST-V1I3P101Keywords:
Cloud–Edge Orchestration, Zero-Trust Architecture, Confidential Computing, Federated Learning, Secure Aggregation, Remote Attestation, Policy-As-Code, Differential Privacy, Reinforcement Learning, Bayesian Optimization, Slo-Aware Scheduling, Sbom, Supply-Chain Security, Ebpf Observability, Intent-Based NetworkingAbstract
We present a multi-layered framework for secure distributed computing across heterogeneous cloud–edge environments that couples zero-trust security with adaptive AI-driven orchestration. The architecture stratifies resources into device/edge, fog/metro, and cloud control planes, connected by a policy fabric that enforces least-privilege access, continuous attestation, and data-in-use protection via confidential-computing enclaves. An intent-based controller translates application SLOs (latency, throughput, cost, energy) and regulatory constraints into verifiable policies. An adaptive orchestrator combining deep reinforcement learning for online placement with Bayesian optimization for safe exploration schedules microservices, serverless functions, and dataflows while observing privacy-preserving telemetry (sketches and differentially private counters). Security services are embedded end-to-end: identity anchored in hardware roots of trust, SBOM-aware image admission, signed provenance on CI/CD supply chains, and runtime anomaly detection using graph embeddings of syscall and network traces. Data governance is maintained through federated learning and secure aggregation to keep raw data local, complemented by fine-grained lifecycle controls and automated compliance checks. A reference implementation on a Kubernetes-native substrate with eBPF observability demonstrates portable enforcement and low overhead. In emulated smart-city and industry-4.0 workloads, the framework reduces tail latency under bursty demand while preserving compliance boundaries and limiting blast radius during fault injections. This work unifies verifiable security and adaptive orchestration, offering a practical path to trustworthy, efficient cloud–edge computing at scale
References
[1] RFC 8446: TLS 1.3 (IETF). https://www.rfc-editor.org/info/rfc8446
[2] Dwork, C. & Roth, A. (2014). The Algorithmic Foundations of Differential Privacy (book PDF). https://www.cis.upenn.edu/~aaroth/Papers/privacybook.pdf
[3] McMahan, H. B., et al. (2017). Communication-Efficient Learning of Deep Networks from Decentralized Data (PMLR PDF). https://proceedings.mlr.press/v54/mcmahan17a/mcmahan17a.pdf
[4] Bonawitz, K., et al. (2017). Practical Secure Aggregation for Privacy-Preserving Machine Learning (ACM PDF). https://dl.acm.org/doi/pdf/10.1145/3133956.3133982
[5] Collaboration Based Cloud Computing Security Management Framework — Mohamed Almorsy, John Grundy & Amani S. Ibrahim, 2011. A security management framework for cloud computing with layered stakeholder/tenant aspect.
[6] A multi level security model for partitioning workflows over federated clouds, P. Watson, Journal of Cloud Computing: Advances, Systems and Applications, vol.1, Article 15, Jul 2012.
[7] Venkatarao Matte & L. Ravi Kumar, “A New Framework for Cloud Computing Security using Secret Sharing Algorithm over Single to Multi Clouds”, International Journal of Computer Trends and Technology (IJCTT), Vol. 4, No. 8, 2013.
[8] Syed Minhaj Ali & Zuber Farooqui, “Multi Layer Security System for Cloud Computing”, Computersciencejournal.org, published September 2014.
[9] Multi Layer Security System for Cloud Computing — Syed Minhaj Ali & Zuber Farooqui, 2014. Presents a multi layer security system in a distributed cloud context.
[10] A New Framework for Cloud Computing security using Secret Sharing Algorithm over Single to Multi Clouds — Venkatarao Matte & L. Ravi Kumar, 2013. Focuses on security in multi cloud environments, using secret sharing.
