Time-varying resilient virtual network mapping for multi-location cloud data centers

摘要: Optical networks constitute a fundamental building block that has enabled the success of cloud computing. Virtualization, cornerstone computing, today is applied in networking field: physical network infrastructure logically partitioned into separate virtual networks, thus providing isolation between distinct operators (VNOs). Hence, problem mapping arisen: how to decide which resources allocate for particular network? In context, not just connectivity required, but also data center (DC) located at multiple locations, computation and/or storage. Given underlying anycast routing principle, operator some freedom specific DC these resources. this paper, we solve resilient optimally decides on both and multi-location resiliently using routing, considering time-varying traffic conditions. terms resilience, consider so-called VNO-resilience scheme, where resilience provided layer. To minimize resource capacity requirements, allow reuse The failures protect against include failures: hence backup resources, account synchronization primary DC. As optimization criteria, only usage minimization, aim limit reconfigurations from one time period next. We propose scalable column generation approach dynamic problem, demonstrate it case study nationwide US backbone network.