Context-Aware Service Function Chaining and Its Cost-Effective Orchestration in Multi-Domain Networks

The emergence of software-defined networking (SDN) and network function virtualization (NFV) facilitates policy-driven and dynamic service function chaining (SFC). In recent years, multi-domain SFC orchestration has attracted a significant attention, as it is a complicated problem involving many iss...

Full description

Bibliographic Details
Main Authors: Guanglei Li, Huachun Zhou, Bohao Feng, Guanwen Li
Format: Article
Language:English
Published: IEEE 2018-01-01
Series:IEEE Access
Subjects:
SFC
Online Access:https://ieeexplore.ieee.org/document/8387764/
Description
Summary:The emergence of software-defined networking (SDN) and network function virtualization (NFV) facilitates policy-driven and dynamic service function chaining (SFC). In recent years, multi-domain SFC orchestration has attracted a significant attention, as it is a complicated problem involving many issues, such as corporation of various administrations and integration of heterogeneous underlying technologies. However, as another critical problem for multi-domain SFC deployment, how to achieve context-aware service provision is still not widely discussed. Thus, in this paper, based on the hierarchical SFC architecture, we allocate metadata to share context information of packets and realize context-based and dynamic SFC in multi-domain networks. We design typical use cases and conduct related experiments about location-aware flow steering and denial of service attack elimination. Experimental results confirm the feasibility and elasticity of our proposed approach. Furthermore, to alleviate the resource over-provisioning problem of context-aware SFC orchestration, we propose a consolidation-based SF placement and chaining algorithm with an online feedback mechanism. We compare the performance of our algorithm with a related greedy algorithm and a tabu search algorithm. Simulation results demonstrate that the proposed algorithm is efficient to reduce the reject rate and improve the revenue-cost ratio, with a compromise on node load balance.
ISSN:2169-3536