Page 25 - ITU Journal Future and evolving technologies – Volume 2 (2021), Issue 2

ITU Journal on Future and Evolving Technologies, Volume 2 (2021), Issue 2

of spatial dimensions; then, the best partitioning that between the UE and the handover time reduces energy
minimizes the device’s energy consumption is chosen. consumption. Also, when the access point is changed,
The problem with this architecture, however, is the lack the re‑registration and reattachment process necessitate
of accuracy in loss, it does not accurately provide the additional energy and the previous energy consumed
amount of data loss. has to be taken into account in the calculation of the total
energy consumed. The proposed architecture is more
To minimize the consumption of energy during task suitable for mobile phones when initiating the handover
of loading and computation under both the main and process in a cloud computing environment and has not
edge processing delay limitations, Xianyan Hu et al. [15] been assessed for potential vulnerabilities yet.
proposed a computing architecture that comprises both
hybrid edge and central cloud, one macrocell with a Ren et al. [44] proposed an ef icient technique to ind the
Macro‑Base‑Station (MBS) and several small cells each optimal resource allocation solution that minimizes la‑
with a small base station, and a continual algorithm to tency in a multi‑user mobile edge computation of loading
ind a solution to the combinatorial mixed‑integer and system by developing a sub‑gradient algorithm. In this
non‑convex optimization problems. In this solution, the solution, the authors irst determined data segmentation
authors considered the delay of synchronizing end user’s methods by considering mobile devices {1, 2, 3, ⋯ , }
tasks, the end users’ tasks are assumed to already be syn‑ and a base station that links the devices to the
chronized in the problem formulation. To guarantee the cloud, the CPU and edge cloud compression capacity of
quality of services provided by the edge clouds, the edge the CPU and respectively, and the compression

processing latency constraints require the corresponding capacity of each device with the following constraint

latency to not exceed a targeted threshold. Furthermore, ∑ <= . Two compression models were irst

=1
to further reduce the complexity of solving the optimiza‑ considered, the Multi‑Access Model where one time slot
tion problem for reducing the total energy consumption is divided into several portions, reducing the data rate of
of the network during task of loading and computation, each portion; and Partial Compression Of loading Model
massive multiple‑input multiple‑output technology is where each ile can be partitioned in two parts with
applied at the multi‑antenna macro‑base‑station. one part compressed locally and the other in the edge
cloud. The proposed algorithm, which is based on the
An energy‑ef icient architecture based on service sub‑gradient method for similar non‑differential convex
provision was proposed by Hani et al.[43] to improve the problems, inds the optimal resource to be allocated
quality of service of the handover process in MCC. The with the aim to reduce the weighted sum latency of the
proposed architecture implies four layers, the media con‑ compression in all devices.
nectivity layer, the application layer, the Internet protocol
multimedia subsystem (IPMS) layer, and the communi‑ With the goal of reducing the delay of handling tasks
cation layer, and was implemented in C++. The Media execution and tasks failure of data partitioned based
Connectivity Layer (MCL) is responsible for connectivity applications, Nguyen et al. [45] proposed a fuzzy based
and media related operations and services, it includes the logic mobile edge orchestrator to segment tasks from
Media Resource Agent (MRA) and Media Resource Func‑ UEs and associate them to the appropriate edge servers.
tion Controller (MRFC). The application layer connects The proposed framework gets as input the network
to the IPMS layer to assure data communication and to and resources information such as bandwidth, size of
the cloud computing servers as an enterprise server. the task being processed, the characteristic of the edge
The IPMS layer is responsible for offering services such server’s virtual machine being used, and the latency
as web browsing, video streaming, videoconferencing, sensitivity associated with each task. It also involves a
email, the Internet, and handles the registration process fuzzi ication step where membership functions are set
used to obtain users’ location. This layer also integrates accordingly to transform the inputs into fuzzy values, and
a Call Session Control Function (CSCF) to associate the a defuzzi ication step where fuzzy values are transformed
users identity to the IP address; the function has three to normal values. The strategy to divide the execution of
parts known as Proxy‑CSCF, serving CSCF, and interro‑ tasks includes the fact that the orchestrator determines
gating CSCF. The communication layer carries the data if the task has to be collaboratively processed by the edge
and binds the media layer to the IPMS layer; besides, it and cloud servers or the edge server alone by computing
includes a Media Gateway Controller Function (MGCF), and choosing the environment with the smaller fuzzy
Media Resource Function Controller (MRFC) and Break‑ values of input parameters, and crisp output value; if the
out Gateway Control Function (BGCF). In addition, it crisp output value is greater than the threshold, the task
includes an energy‑ef icient detection model to ascertain is executed by the cloud server alone.
the energy of nodes when initiating the handoff process.
The energy consumed during the handover process is
proportional to the distance between the mobile device
and its access point and the time required to complete
the handover process. Thus, minimizing the distance

© International Telecommunication Union, 2021 11