A Bi-Objective Policy for Resilient and Sustainable SFC Management in Telco-Cloud Environments
Muhammad Hamza Shahab, Yogesh Sharma, Anish Jindal, Auday Al-Dulaimy · IEEE Access · 2025
Achieving high availability in Service Function Chains (SFCs) typically requires deploying redundant Virtual Network Function (VNF) instances. However, this redundancy increases resource usage and energy consumption, raising the carbon footprint of the service. Reducing redundancy lowers the energy consumption but impacts the service availability during failures, creating a critical trade-off between service availability and sustainability. To address this, a novel bi-objective policy is introduced that leverages a Tradeoff Aware Embedding (TAE) algorithm using greedy approach for resource provisioning, and a PSO-based Redundancy Optimizer (PRO) that dynamically adjusts VNF redundancy to meet service availability requirements without increasing the carbon footprint. To further enhance the fault tolerance, the policy integrates a Gradient-Boosting (GB) enabled failure prediction and remediation mechanism called Redundancy with Checkpointing (RPC), enabling proactive failure detection and rapid recovery. Simulation results using real-world failure traces demonstrated that the proposed policy with RPC achieved up to 99.987% availability while reducing the carbon footprint of the service by over 87% and service failures by more than 98% compared to the worst-case without redundancy and checkpointing mechanisms.