Back to search

University of Glasgow

Leveraging NFV heterogeneity at the network edge

Abstract

dc:description.abstract

With network function virtualisation (NFV) and network programmability, network functions (NFs) such as firewalls, traffic load balancers, content filters, and intrusion detection systems (IDS) are virtualized and either instantiated on user space hosts using virtual machines (VMs), lightweight containers, or in the network data plane using programmable switching technology such as P4 or offloaded onto Smart network interface cards (NICs) – often chained together to create a service function chain (SFC), based on defined service level agreement (SLA). The need to leverage heterogeneous programmable platforms to support the in-network acceleration of functions keeps growing as emerging use cases come with peculiar requirements. This thesis identifies various heterogeneous frameworks for deploying virtual network functions that network operators can leverage in service provider networks. A novel taxonomy that provides network operators and the wider research community valuable insights is proposed. The thesis presents the performance gains obtained from using heterogeneous frameworks for deploying virtual network functions using real testbeds. In addition, this thesis investigates the optimal placement of vNFs over the distributed edge network while considering the heterogeneity of packet processing elements. In particular, the work questions the status quo of how vNFs are currently being deployed, i.e., the lack of frameworks to support the seamless deployment of vNFs that are implemented on diverse packet processing platforms – leveraging the capability of the programmable network data plane. In response, the thesis presents a novel integer linear programming (ILP) model for the hybrid placement of diverse network functions that leverages the heterogeneity of the network data plane and the abundant processing capability of user space hosts, with the objective function of minimizing end-to-end latency for vNF placement. A novel hybrid placement heuristic algorithm, HYPHA, is also proposed to find a quick, efficient solution to the hybrid vNF placement problem. Using optimal stopping theory (OST) principles, an optimal placement scheduling model is presented to handle dynamic edge placement scenarios. The results in this work demonstrate that employing a hybrid deployment scheme that leverages the processing capability of the network data plane yields minimal user-tovNF latency and overall end-to-end latency while fulfilling the placement of a diverse set of user requests from emerging use cases to speed up service delivery by network operators. The results also show that network operators can leverage the high-speed, low-latency feature of data plane packet processing elements for hosting delay-sensitive applications and improving service delivery for subscribed users. It is shown that the proposed hybrid heuristic algorithm can obtain near-optimal vNF mapping while incurring fewer latency threshold violations set by network operators. Furthermore, in addition to emerging edge use cases, the placement solution presented in this thesis can be adapted to place network functions efficiently in core network infrastructure while leveraging the heterogeneity of servers. The dynamic placement scheduler also minimises the number of latency violations and vNF migrations between heterogeneous hosts based on SLAs set by network operators.

Degree

thesis:*
Level dc:type.qualificationlevel
PhD
Grantor dc:publisher.institution
University of Glasgow
Year dc:date.issued
2024

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Adoga, Haruna Umar

Subjects

dc:subject × 1

Rights

Language dc:language
en

Identifiers

dc:identifier.*

Chain of custody

source
Harvested from
University of Glasgow
Base URL
theses.gla.ac.uk/cgi/oai2
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Adoga, Haruna Umar. Leveraging NFV heterogeneity at the network edge. PhD thesis, University of Glasgow, 2024. https://doi.org/10.5525/gla.thesis.84344