Service function chains deploying in hybrid scenarios
Chang Liu, Jing Ran, Hefei Hu · 2024
In the current network landscape, operators deploy network functions on physical devices (such as switches and middleboxes) as well as on Network Function Virtualization (NFV) devices (e.g., virtual machines on general-purpose servers), forming service function chains that meet specific requirements. This approach efficiently leverages the processing capabilities and deployment flexibility of both types of devices. However, the heterogeneity between devices may lead to the independent generation of multiple conflicting solutions within the network, making it challenging to identify an optimal deployment strategy. To address this challenge, this paper introduces a Taboo Search Algorithm Based on N-jump Search Solution Space (TSABNS). This algorithm aims to minimize the average latency cost of Service Function Chain (SFC) deployments while ensuring that most SFCs are successfully deployed within the network. Comparative simulations demonstrate that TSABNS significantly outperforms other algorithms, such as HOPE, in terms of deployment success rate and latency performance.