Progressive Forwarding Disaster Backup among Cloud Datacenters

Abstract

The periodic disaster backup activity among geographically distributed multiple datacenters consumes huge network resources and therefore imposes a heavy burden on datacenters and transmission links. Previous work aims at least completion time, maximum utility or minimal cost, without consideration of load balance for limited network resources, likely to result in unfair distribution of backup load or significant impact on daily network services. In this paper, we propose a new progressive forwarding disaster backup strategy in the Software Defined Network scenarios to mitigate forwarding burdens on source datacenters and balance backup loads on backup datacenters and transmission links. We construct a new redundancy-aware time-expanded network model to divide time slots according to redundancy requirement, and propose role-switching method over time to utilize forwarding capability of backup datacenters. In every time slot, we leverage two-step optimization algorithm to realize capacityconstrained backup datacenter selection and fair backup load distribution. Simulations results prove that our strategy achieves good performance in load balance under the condition of guaranteeing transmission completion and backup redundancy.

Progressive disaster evacuation in cloud datacenter network

“…PDET is a two-step optimization algorithm consisting of safe datacenter selection (SDS) and bandwidth proportion allocation for concurrent evacuation transfers. As in our earlier work, 9,13 we choose SDN as running network environment to collect network status and split flows conveniently.…”

“…In our earlier work, 13 we propose progressive forwarding disaster backup (PFDB) algorithm consisting of capacity-constrained backup datacenter selection (CC-BDS) and fair link load distribution (FLLD). However, there are some key differences between PFDB and PDET: 2.…”

“…We define a time slot t ∈ { t 1 , t 2 , … , t r } as the time period from prediction to occurrence of a damaging event (including a certain disaster event or its spreading events). Different from disaster backup, 13 the time slot value is not fixed. We can leverage t to denote the deadline for all evacuation transfers in the current time slot.…”

“…Some researchers leverage TEN to convert the dynamic flow over time problem to static flow problem 8,12 . However, their strategies have not considered store‐and‐forward datacenter selection and routing searching jointly 13 . Therefore, their strategies cannot be applied to disaster evacuation activity in practice.…”

“…In our earlier work, we build redundancy‐aware TEN model, divide time slots and utilize forwarding capability by role‐switching method, and leverage two‐stage optimization algorithm to realize capacity‐constrained backup datacenter selection and fair backup load distribution in every time slot 13 . However, we have not considered how to improve evacuation efficiency through store‐and‐forward node selection and personalized bandwidth allocation.…”