Rerouting Strategies for Highly Available Virtual Network Functions

• Published in: IEEE transactions on cloud computing Vol. 9; no. 4; pp. 1592 - 1606
• Main Authors: Shang, Xiaojun, Li, Zhenhua, Yang, Yuanyuan
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE Computer Society 01.10.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Approximation; Approximation algorithms; Availability; Cloud computing; Heuristic algorithms; Heuristic methods; Links; locally rerouted path; Mathematical analysis; Middleboxes; Optimization; Prediction algorithms; randomized rounding; Ratios; Rerouteing; Routing; Servers; Strategy; virtual network function; Virtual networks
• ISSN: 2168-7161; 2168-7161; 2372-0018
• DOI: 10.1109/TCC.2019.2925110

The development of Virtual Network Functions (VNFs) migrates network functions from dedicated hardware to groups of commodity servers called network points of presence (N-PoPs). In this way, network services are redefined as interconnected VNFs called Service Function Chains (SFCs). The emerging of SFCs significantly reduces the cost of network services and improves scalability. However, the availability of SFCs brings new challenges since a failure of any N-PoP along an SFC affects its availability. In this paper, we propose two rerouting strategies to improve the availability of SFCs. First, we propose a local rerouting strategy to bypass the failed N-PoPs on SFCs using locally rerouted paths (LRPs). In the strategy, we formulate an optimization model to minimize the maximum load on links while deploying SFCs and LRPs to reduce the risk of congested links caused by local rerouting. We then propose an approximation algorithm to solve the optimization problem, preserving an approximation ratio of <inline-formula><tex-math notation="LaTeX">\mathcal {O}(\log (|V|))</tex-math> <mml:math><mml:mrow><mml:mi mathvariant="script">O</mml:mi><mml:mo>(</mml:mo><mml:mo form="prefix">log</mml:mo><mml:mo>(</mml:mo><mml:mo>|</mml:mo><mml:mi>V</mml:mi><mml:mo>|</mml:mo><mml:mo>)</mml:mo><mml:mo>)</mml:mo></mml:mrow></mml:math><inline-graphic xlink:href="shang-ieq1-2925110.gif"/> </inline-formula>, where <inline-formula><tex-math notation="LaTeX">|V|</tex-math> <mml:math><mml:mrow><mml:mo>|</mml:mo><mml:mi>V</mml:mi><mml:mo>|</mml:mo></mml:mrow></mml:math><inline-graphic xlink:href="shang-ieq2-2925110.gif"/> </inline-formula> is the number of N-PoPs in the network. We also propose an alternative heuristic algorithm to improve efficiency. Second, we propose a supplementary rerouting strategy with an online algorithm to provide supplementary rerouted paths when original SFCs and corresponding LRPs fail at the same time in the local rerouting strategy. The online algorithm is proved to have an <inline-formula><tex-math notation="LaTeX">\mathcal {O}(\log (|V|))</tex-math> <mml:math><mml:mrow><mml:mi mathvariant="script">O</mml:mi><mml:mo>(</mml:mo><mml:mo form="prefix">log</mml:mo><mml:mo>(</mml:mo><mml:mo>|</mml:mo><mml:mi>V</mml:mi><mml:mo>|</mml:mo><mml:mo>)</mml:mo><mml:mo>)</mml:mo></mml:mrow></mml:math><inline-graphic xlink:href="shang-ieq3-2925110.gif"/> </inline-formula> competitive ratio to the offline optimum. Finally, our extensive simulation results show that the proposed algorithms can provide highly available SFCs with less congested links.