Page 11 - ITU Journal Future and evolving technologies – Volume 2 (2021), Issue 2

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

A resource management survey for mission-critical and time-critical
applications in multiaccess edge computing

Pages 61-80
Nina Santi, Nathalie Mitton

Multiaccess Edge Computing (MEC) brings additional computing power in proximity of mobile users,
reducing latency, saving energy and alleviating the network's bandwidth. This proximity is beneficial,
especially for mission-critical applications where each second matters, such as disaster management or
military operations. Moreover, it enables MEC resources embedded on mobile units like drones or
robots that are flexible to be deployed for mission-critical applications. However, the MEC servers are
capacity-limited and thus need an acute management of their resources. The mobile resources also need
a smart deployment scheme to deliver their services efficiently. In this survey, we review mission-
critical applications, resource allocation and deployment of mobile resources techniques in the context
of the MEC. First, we introduce the technical specifics and uses of MEC in mission-critical applications
to highlight their needs and requirements. Then, we discuss the resource allocation schemes for MEC
and assess their fit depending on the application needs. In the same fashion, we finally review the
deployment of MEC mobile resources. We believe this work could serve as a helping hand to design
efficient MEC resource management schemes that respond to challenging environments such as
mission-critical applications.
View Article

ARPA: An autonomous renderer placement algorithm in distributed

multimedia fog networks with delay guarantees

Pages 81-108
Athanasios Tsipis, Konstantinos Oikonomou

Multimedia cloud computing has emerged as a popular paradigm for the support of delay-intolerable
immersive multimedia applications with high-end three-dimensional rendering. To that end, fog
computing offers distributed computational offloading solutions, by positioning rendering servers in
close proximity to end users promising in this way continuous service provision, that is otherwise not
easily attainable under the strictly centralized cloud-only model. Yet, in order to alleviate the
multimedia providers from unnecessary capital expenditure, a strategic placement approach of the
servers at the fog layer must be implemented, that can effectively cope both with the network dynamics
and the overall imposed deployment cost, and still adhere to the delay bounds set forth by the
multimedia application. In this paper, we formally formulate the problem as a facility location problem
using constrained optimization over a finite time horizon. We then theoretically analyze the minimum
acceptable conditions necessary for a decentralized location of the servers, utilizing solely local
information around their immediate neighborhood, that iteratively leads to better solutions. Based on
the analysis, we propose a distributed algorithm, namely the Autonomous Renderer Placement
Algorithm (ARPA), to address it. ARPA employs localized service relocation to shift the placement
according to simple rules that designate elastic migration, replication, and complementary consolidation
of the underlying renderers. Simulation results under diversified deployment scenarios, as well as trace-
driven comparisons against other approaches, testify to ARPA's accountability in obeying the delay
limits and fast converge in finite time slots to a placement solution that both outperforms the baseline
alternatives and is close to the optimal one, rendering it suitable for scaling up and down to meet the
current demands of the offered multimedia applications.
View Article

– ix –