VNF Placement Optimization at the Edge and Cloud †

Abstract

Network Function Virtualization (NFV) has revolutionized the way network services are offered to end users. Individual network functions are decoupled from expensive and dedicated middleboxes and are now provided as software-based virtualized entities called Virtualized Network Functions (VNFs). NFV is often complemented with the Cloud Computing paradigm to provide networking functions to enterprise customers and end-users remote from their premises. NFV along with Cloud Computing has also started to be seen in Internet of Things (IoT) platforms as a means to provide networking functions to the IoT traffic. The intermix of IoT, NFV, and Cloud technologies, however, is still in its infancy creating a rich and open future research area. To this end, in this paper, we propose a novel approach to facilitate the placement and deployment of service chained VNFs in a network cloud infrastructure that can be extended using the Mobile Edge Computing (MEC) infrastructure for accommodating mission critical and delay sensitive traffic. Our aim is to minimize the end-to-end communication delay while keeping the overall deployment cost to minimum. Results reveal that the proposed approach can significantly reduce the delay experienced, while satisfying the Service Providers' goal of low deployment costs.

Multi-Access Edge Computing: A Survey

“…bandwidth and CPU). In [43] a VNF placement approach is proposed which considers the existence of edge servers and cloud servers in order to provide network services Optimization approach SDN NFV SFC 5G Slicing Yala et al [38] • • Cziva et al [40] • • Yang et al [41] • • Leivadeas et al [43] • Kaur et al [61] • [114] • • Ren et al [115] • Yang et al [116] • • Son et al [44] • • • Alameddine et al [117] • • Chen et al [118] • Li et al [119] • • • Chen et al [120] • Ahn et al [121] • • Li et al [122] • • Liang et al [123] • Blanco et al [126] • • Yang et al [124] • Liu et al [125] • Zhang et al [127] • Chen et al [128] • • Guo et al [131] • Tran et al [132] • Zhang et al [133] • Zhao et al [135] • Wang et al [143] • • • Sanchez et al [111] • Tun et al [145] • Martiradonna et al [146] • Taleb et al [113] • • • Mena et al [103] • Wang t al. [112] • Xiang et al [147] • • Tun et al [144] •…”

“…Leivadeas et al [43] • End-user devices randomly enter and exit from a service chain according to an identical and independent Poisson process. • Only three topologies that a service chain can follow.…”

Edge Computing Resource Allocation for Dynamic Networks: The DRUID-NET Vision and Perspective

“…Nevertheless, none of these approaches attempted to estimate the IoT workload, which in turn could significantly enhance the corresponding deployments. The authors of [13] and subsequently of [14] analyzed the resource allocation of a three-layer infrastructure (IoT, Edge, Cloud) under dynamic network conditions. However, they took into consideration the dynamic opt-in and out of IoT devices into the network, while ignoring their instantaneous workload generation.…”

Software Defined Networking (SDN) and Network Function Virtualization (NFV)

“…The fifth paper [5] focuses on the effective and efficient virtualized network functions (VNF) placement within edge and cloud infrastructures in order to provide enhanced networking services to Internet of Things (IoT) applications. In particular, the authors propose a novel approach to facilitate the placement and deployment of service-chained VNFs in a network cloud infrastructure that can be extended by using the mobile edge computing (MEC) infrastructure for accommodating mission-critical and delay-sensitive traffic.…”