IEEE Systems, Man and Cybernetics Magazine - January 2023 - 13

supply flexible services. Gao et al. [9] study the dynamic
resource allocation problem for VNF placement in satellite
edge clouds, which aims to minimize the network cost and
end-to-end delay. While a satellite network should be able
to accommodate as many user requests as feasible, the
VNF placement problem seeks to minimize the deployment
cost for each user request [10]. Zhang et al. [11], [12]
employ a demand-supply model to measure the VNF interference
since VNF consolidation might seriously affect
performance. They propose an Online Lazy-migration
Adaptive Interference-aware Algorithm
(OLAIA) for real-time VNF
deployment and cost-effective VNF
migration in a 5G network slice to
optimize the overall reward of
approved requests.
An enormous rise in data volrequests
with service chain requirements, and the service
chain requires the data traffic of the request to
pass through the VNFs in the chain one by one until it
reaches its destination. Xu et al. [19] consider admissions
of NFV-enabled requests of IoT applications in a
multitier cloud network. EdgeUsher is a solution that is
suggested to address the issue of where to best put
VNF chains to cloud-edge infrastructures to meet all of
their hardware, IoT, security, bandwidth, and latency
requirements [20].
However, none of the previous
umes has been caused by the
increasing number of physical
objects linked to the Internet and
using resource-intensive IoT applications.
In an end-to-end communication
paradigm, network operators
must find a way to manage restricted resource availability,
dynamic IoT traffic requirements, and dynamic IoT device
behavior [13]. To implement edge computing in 6G mobile
networks, effective resource allocation techniques are
needed to position the VNFs. Mendoza et al. [14] solve the
issue theoretically as a finite Markov decision process
(MDP) to achieve the theoretical performance constraint
for VNF placement in 6G networks (MDP). Latency is
always a key performance metric for applications on
the edge. To decrease the end-to-end latency, time to
response, and idling of the core network, Cziva et al. [15]
formulate the VNF placement problem and deploy VNFs at
the network's edge, near to the end users aiming to minimize
end-to-end latency. The SFC, which may require more
or less tight delay limitations, is required to solve the VNF
placement problem. Manzanares-Lopez et al. [16] propose
the Modified Priority-based Greedy (ModPG) heuristic, a
VNF placement method that takes into account not just
resource constraints like latency and bandwidth but also
the cost of instantiating VNFs.
Jin et al. [17] concentrate on jointly optimizing the
The objective of this
article is to minimize
the cost and profit as
the supplementary
objective in our model.
research works construct the
VNF placement problem in a view
of cost efficiency. This work considers
cost in an explicit method
as the resource al location of
VNFs in separate paths. In a practical
way, it is simple to select
resource-satisfied nodes first and
then allocate VNFs in these nodes
based on the requirements of different
applications. It is apparent
that to improve the profitability of
this system, other jobs in the same task should consider
the cost, respectively. A cloud-related work can be found
in [21], which aims to minimize the overall cost from the
metrics of latency and energy consumption. However, it
does not take the application's priority and significance
into consideration.
The System Model
In this section, we introduce the model of this article. The
industrial edge can be modeled as a graph
GE L= (, ),
where E represents the edge nodes in the system, and L
denotes the link set in the network. The VNFs can be
deployed in the system based on their requirements.
We define
Ee een12
= {, , ..., } as a set of edge systems distributed
to the manufacturers. The jobs in the system
are required to access the resource in the edge nodes.
The devices in the industry environment need to be
able to provide redundant resources for required tasks
for robustness.
For those virtualized functions, we denote the set of
resource use of edge servers and physical networks within
the latency constraints. To reduce overall resource
usage, they describe the VNF chain deployment issue as
a mixed ILP problem. The latency-aware EdgeCore SFCs
migration problem is proposed by Liang et al. [18] using
open Jackson networks and SFCs in edge-core networks.
Profit-driven heuristic search (PHS) and average utilization
based (AUB) are two SFC migration methods that
are suggested to effectively reduce the average latency of
all SFCs in edge-core networks. Quality of service (QoS)
is also important for the performance of the application.
Users request network services by sending service
VNFs as vi, which represents the designated VNFi, while
VNFi is reusable for a time. When considering reducing the
placement cost ci, the reusable VNFs can be applied as
c .ia,
For each vi, when engaged, some physical resources are
occupied (i.e., CPU cycles and dynamic RAM memory).
The capacity of this system can be described as P for the
CPU, O for memory, and W for the network.
The tasks in the industry edge contain several jobs, and
each job requires multiple VNFs in the system. To serve
the request rR! (where R represents the set of requests)
for an industrial task, the job needs to access a series of
dependent VNFs. In this article, we denote the service
request r as the set
r vvv vi,, which represent
= " 12 3
,, ,,
f
January 2023 IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE 13
IEEE Systems, Man and Cybernetics Magazine - January 2023

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - January 2023
