A method for fast switching of redundant networks between data centers

Last year´s earthquake in Eastern Japan sparked a growing demand for redundant structures in Web 3-tier model/servers, NICs and physical lines - which are required by conventional on-premises environments with high reliability and availability - to be directly migrated to cloud environments. In such environments, when applying ICMP connectivity checks and conventional fault detection methods such as IP-SLA, it takes time to recover from faults to allow for coordination between the core network and the user logical networks in the cloud. This paper proposes a method where a correspondence list of user logical networks and core networks is registered beforehand in a cloud management system. When a fault has occurred, the NIC of the virtual machine corresponding to the fault location is closed off, thereby resulting in a rapid switchover to a redundant configuration.