IEEE Systems, Man and Cybernetics Magazine - January 2023 - 11

of communication as the nodes are
more distant. In this article, we
investigate the resource allocation
problem for VNF placement in
industrial edge systems. The aim is
to minimize the overall cost of the
whole placement procedure,
including operating cost, communication
cost, and placement cost.
We formulate this VNF placement
problem as an integer programming
problem, and due to the complexity
of the problem, we propose
a heuristic method to solve it.
Numerical simulation shows that our proposed method can
significantly cut down the cost while meeting the multiple
resource constraints.
Introduction
With the development of information-physical convergence
systems and the Internet of Things (IoT), the connections
among various regions of the world are
becoming closer [1]. Global manufacturing companies lay
out digital businesses so that their products and equipment
will be more intelligent. To realize interactions
among people, products, and machines, there is an intention
to build highly flexible and personalized solutions
for products and services in an edge system. With the
assistance of the edge, the entire industry system will be
more intelligent and more closely networked, and many
components will be able to communicate and operate
more swiftly.
It is widely known that current mobile edge architectures
lack the scalability and flexibility to quickly adapt to
industries and support innovative services. Applying software-defined
networking (SDN) ideas in current architecture
design has lately gained a lot of interest as a solution
to the aforementioned problem. With the widely
researched network function virtualization (NFV) technologies,
an adaptive network can be established with
full virtualization and effective network slicing, which
can be tailored to the needs of various tenants (bandwidth
tolerance, latency sensitivity, etc.). Network functions
can be virtualized by being decoupled from the
underlying physical devices in an SDN/NFV-enabled network,
giving rise to VNFs that are employed as virtual
machines or containers. For a user request, a VNF placement
and scheduling process in the architecture can be
described as an end-to-end network service, where a number
of VNFs must be accommodated in physical devices
and applied when invoked. To be more precise, the VNF
placement of a service request requires specific functions
at the point of virtual network devices [2].
There is no doubt that the VNF carries different costs
in different locations. At the same time, in the industrial
environment, video streams dominate the total traffic of
In this article, we
investigate the
resource allocation
problem for VNF
placement in
industrial edge
systems.
the edge network. It is reasonable
to provide different allocation
strategies for different applications.
To reasonably share content
and allocate resources, the reduction
of sharing costs depends on
the location where the VNF is
deployed. Therefore, the placement
allocation of VNFs will greatly
affect the cost of the entire service
and the overall traffic in the network
because nodes with multiple
VNFs will bring large traffic. However,
in the industrial edge network,
the reasonable utilization of video applications is
regarded as a challenging task because the location of the
VNF determines the service performance of applications
and the resource overhead. It is worth noting that, without
considering the impact of mobility, the VNF is deployed
near the final data utilization to significantly reduce the
performance degradation caused by the transmission of
big data. In this case, the VNF is closer to the data generation
and utilization end to avoid a large amount of data
transmission. It is not advisable to blindly improve the
data transmission performance and ignore the overhead
brought by deployment. Therefore, it is necessary to consider
the impact of the placement position of the VNF on
the cost and the traffic constraint while meeting the performance
requirements of different applications.
The main contributions are summarized as follows:
◆ We study the resource allocation problem for VNF
placement in the industrial edge cloud.
◆ We formulate the VNF placement problem in the industrial
edge to minimize the overall cost.
◆ We propose a distributed VNF placement algorithm to
solve this problem and conduct extensive simulations
to evaluate the performance of our proposed method.
The remainder of this article is organized as follows.
The " Motivation and Related Works " section shows a motivation
example and related works. The " System Model "
section introduces the system model in the industry edge
system. We propose placement method in the " Placement
Method " section. In the " Numerical Results " section, we
evaluate the performance of our proposed method. Finally,
we conclude this article in the " Conclusion " section.
Motivation and Related Works
The Motivation Example
We consider a remote operation application as our example
of a critical application in industry. This application has
two types of jobs: video transmission and operation instructions.
We divide one remote operating task into video
and instruction transmission. First, we focus on the video
flow, which includes the current situation of devices, anomaly
detection, and a firewall to ensure the security of
January 2023 IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE 11
IEEE Systems, Man and Cybernetics Magazine - January 2023

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