Page 28 - ITU Journal Future and evolving technologies – Volume 2 (2021), Issue 2

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

Table 1 – Qualitative review of different solutions proposed to achieve seamless communication in MCC.

Proposed Approach Analysis Summary Shortcoming
Technique Vehicular Computing, V2X, FVC, Edge • Mobile vehicles cannot act as fog‑enabled
Selection, etc.
Seamless handover & devices yet.
service provision Description • The proposed solutions are mainly based on fog‑enabled vehicles, FVC, and base stations
to ensure fast and ef icient handover.
[28, 29, 30, 31, 32]
• Proposed a fog computing platform that enables the allocation and management on the
set of computational resources for executing effectively IoT tasks.
Technique Mixed‑integer quadratic program‑ •Limited to an environment with homogeneous
ming, K‑Means, Caching at edge, setting.
Proxy‑based and Greedy content
placement algorithm, etc. • Do not include load balancing management.
Placement of • Expensive solutions since caching hardware
Edge Servers & must be integrated on each edge cloud.
Base Stations Description • The proposed solutions are mainly based on mixed‑integer quadratic programming and
K‑Means algorithms to compute optimal placement locations of edges such that the work‑
[29, 37, 38, 39]
loads are balanced and the access delay reduced.
• The proxy server is selected based on four parameters, which include the type of host,
the state of the host, the hardware performance of the host and the available amount of
concurrent connections.
• Proposed a homogeneous mobile network with edge caching where the mobile device
fetches coded segments directly from candidate SBSs in ascending order of transmission
distance, if the requested content is cached.
Technique Web worker migration, machine • Solution is very limited.
learning, Gradient algorithm, Data • Mainly suitable for web applications only.
segmentation Mixed‑integer non
linear programming, Gaussian ran‑ • Intelligence in fog computing is still
domization, Subgradient algorithm, in its infancy, and the assumptions made
etc.
are not realistic yet.
Reducing Latency & • Lack of accuracy in loss.
Energy Consumption, • Latency in links is not the only major
Improvement of parameter to be considered.
handover QoS Description • A hybrid edge and central cloud computing architecture was proposed, including one
macro cell with a Macro Base Station (MBS) and multiple small cells each with an SBS,
[15, 40, 41, 42, 43, 44]
and an iterative algorithm used to solve the combinatorial mixed‑integer and non‑convex
optimization problems.
• Web worker migration techniques and machine learning were proposed to detect user’s
behaviors, ind optimal servers and make ef icient of loading decision.
• Spatial and channel‑wise reduction units were applied to create a compressed represen‑
tation of the feature tensor which is transmitted to the cloud.
• Energy consumed during the handover process can be reduced by computing the mini‑
mum distance between the UE and the handover BS.
• Subgradient algorithm was applied to compute the minimum latency between links in an
edge and perform resource allocation accordingly.
Technique Finite Horizon Markov Decision • The solution only reduces the complexity
Process, Ant Colony Optimization, of the problem.
Divide‑and‑conquer based near
optimal placement algorithms, etc. • Not appropriate for real time systems which
Data Of loading & require real time performance guarantee.
Load Balancing • Static algorithm based and suitable
[46, 47, 48, 49] for batch scheduling only.
Description • Optimal placement (EOPA) and divide‑and‑conquer based near optimal placement algo‑
rithms (DCNOPA) were proposed to ef iciently distribute virtual machine replica copies
(VRCs) of applications to the edge network to reduce high data traf ic in edge networks.

14 © International Telecommunication Union, 2021