Towards Adaptive Traffic Flow Optimizations with Unikernel-based VNFs

Jung, Alexander (2018) Towards Adaptive Traffic Flow Optimizations with Unikernel-based VNFs. In: UNSPECIFIED.

Abstract

VNFs mimic classical physical network operations such as routers and switches but enable easier deployment, malleability, and scalability. Unikernels as a Process (UaaP) or process-like VMs (pVMs) are one method of strongly encapsulating, securing, transporting and deploying VNFs on commodity hardware. The use of resources in this context has changed and has become out of touch with current networking practices particularly because this model is leading itself to highly dense and animated software-based network usage which in turn leads to suboptimal use of virtualized resources. In order to better utilize existing services, maintain clear separation of concerns and attempt to bring order to this network convolution, we argue for a new strategy for manipulating the dataplane for encapsulated services.