Balancing the Migration of Virtual Network Functions with Replications in Data Centers

摘要

The Network Function Virtualization (NFV) paradigm is enabling flexibility,programmability and implementation of traditional network functions intogeneric hardware, in form of the so-called Virtual Network Functions (VNFs).Today, cloud service providers use Virtual Machines (VMs) for the instantiationof VNFs in the data center (DC) networks. To instantiate multiple VNFs in atypical scenario of Service Function Chains (SFCs), many important objectivesneed to be met simultaneously, such as server load balancing, energy efficiencyand service execution time. The well-known \emph{VNF placement} problemrequires solutions that often consider \emph{migration} of virtual machines(VMs) to meet this objectives. Ongoing efforts, for instance, are making astrong case for migrations to minimize energy consumption, while showing thatattention needs to be paid to the Quality of Service (QoS) due to serviceinterruptions caused by migrations. To balance the server allocation strategiesand QoS, we propose using \emph{replications} of VNFs to reduce migrations inDC networks. We propose a Linear Programming (LP) model to study a trade-offbetween replications, which while beneficial to QoS require additional serverresources, and migrations, which while beneficial to server load management canadversely impact the QoS. The results show that, for a given objective, thereplications can reduce the number of migrations and can also enable a betterserver and data center network load balancing.