Elastic Slicing in Programmable Networks

Abstract

The concept of network slicing enables operators to provision multiple virtual networks on top of a single (shared) physical infrastructure. Adding elasticity to slicing, i.e., the ability to on-demand provision/release dedicated network resources, improves resource utilization. However, efficiently allocating and scaling slice resources, while maintaining state consistency, is challenging. Especially with P4-programmable network devices that process packets at Tbps speeds, controller-driven scaling of network functions would be too time-consuming, and data-plane scaling is needed.In this paper, we address this need, by developing a custom scaling protocol and framework that can consistently, with negligible delay, scale network slices and functions transparently to the slice end-users. We compare, via emulation and experiments on programmable hardware, our approach to state-of-the-art scaling techniques and demonstrate significant slice resource utilization improvements and scaling duration reductions.

Programmable Data Plane Applications in 5G and Beyond Architectures: A Systematic Review

“…Turkovic et al [ 86 ] propose a P4 switch-based network slicing framework that can handle time-sensitive tasks such as overload and underload detection, rerouting, and state transfer. A custom slice management protocol is implemented using a SM header to enable efficient slice management.…”

“…The study in [ 85 ] demonstrates that the proposed prototype can support up to 16,000 slices, while [ 80 ] reports that creating slices takes 400 ms, which is faster than slicing implementations that do not use programmable data planes. The network slicing framework in [ 86 ] shows an average delay seen at the end-host of 69.8 ms. The solution presented in [ 84 ] demonstrates that slices experiencing high traffic loads, defined as four or more flows within a slice, share the remaining bandwidth with a maximum difference of 4%.Furthermore, the bandwidth allocated to each flow within a slice varies only slightly, ranging from 0.54% to 8.23% of the total slice bandwidth.…”