Probe Station Selection And Fault Diagnosis Algorithm In The Network Virtualization Environment

Network virtualization, as a feasible approach to solve the problem caused by the ossification of the Internet, has caused research organization’s attention. Under this situation, this paper focuses on the problem of fault management in the network virtualization environment. Fault management can obtain the performance data of network, find the faults, locate the root and assure the proper and stable operation of network service, improve the availability and stability of network. However, due to the particular characteristics of network virtualization, traditional fault management technology cannot be applied to the new environment, especially for probe station selection and fault diagnosis. In order to solve these problems, we carry out the widespread research on the fault management in the network virtualization environment.(1) Probe station selection algorithmProbe station selection is an important research subject in the fault monitoring, which affect the evaluation of monitoring performance directly and has certain side effect on network workload. The problems, such as multi-layer network architecture, non-deterministic mapping relation, the dynamic network topology, bring to probe station selection new challenges. This paper proposes a probe station selection algorithm based on probabilistic dependency model and independent path, which is used in non-deterministic and dynamic virtual network environment. The algorithm in this paper can deal with the challenges of the non-determinacy and dynamicity to achieve good monitoring performance with the minimal cost, and we prove it by simulation.(2) Fault diagnosis algorithm The ultimate goal of fault diagnosis is to provide certain auxiliary information for fault recovery by resources reallocation and topology reconstruction of virtual network. But the flexible characteristics of virtual networks bring to virtualization fault diagnosis new challenges, such as network scalability, inaccessible substrate network fault information, incomplete and inaccurate network observations, dynamic symptom-fault causality relationships, and multi-layer complexity. We first set up a network monitoring model. Then the paper proposed a collaborative end-to-end fault diagnosis algorithm used in network virtualization environment. The algorithm can analyze the relationship of symptoms from multiple virtual networks and tackle with the challenges based on Bayes network. Simulations and experimental studies show that our solution can efficiently and accurately get a hypothesis set explaining all the observed symptoms, even in the large-scale network and with the effect of noise.