Unavailability-aware Shared Virtual Backup Allocation Model for Middleboxes
Fujun He, Eiji Oki · 2020
Network function virtualization paradigm enables us to implement network functions provided in middleboxes as softwares which run on commodity servers. This paper proposes an unavailability-aware backup allocation model with shared protection for middleboxes with comprehensively considering the failure, repair, and recovery behaviors of functions and backup servers. Multiple functions can share the backup resources on the backup server. The proposed model aims to find the assignment of backup servers to functions to minimize the maximum unavailability among functions. The multiple situations of failure, repair, and recovery of functions and backup servers lead to several different states for each function. The unavailability of function is estimated through analyzing all states that a function can be in. To compute the unavailability of middlebox for a given backup allocation, an analytical approach is developed based on the queueing theory. With the analytical approach, we introduce a simulated annealing heuristic to solve the backup allocation problem. The results reveal that, compared to a baseline model, the proposed unavailability-aware model reduces the maximum unavailability 11% in average in our examined scenarios.