Enhancement of Fog Computing based Autonomous Management Systems
1. Enhancement of Fog Computing based
Autonomous Management Systems
1
Sabelo Dlamini and 2
Neco Ventura
e-mail: 1
sabelo.dlamini@csir.co.za and 2
neco.ventura@uct.ac.za
Introduction
Fog computing aims to bring cloud computing
capabilities to the edge of the network, closer to the
end user, enabling lower latency levels, location
awareness, and mobility support among other
advantages. The combination of IoT and Fog
encompasses a highly complex scenario with a huge
amount of data and huge amount of different devices
that must cooperate with each other. This requires
effective management and orchestration mechanisms to
guarantee acceptable performance of applications and
services. Mechanisms typically applied to the Cloud,
however, cannot naturally be migrated to the Fog given
its particular characteristics. This calls for the design and
development of new management and orchestration
mechanisms for the Fog.
Proposed Scheme References
[1] Cisco Systems, Inc. “The Zettabyte Era: Trends and Analysis” June, 2017 [Online]
[2] OpenFog Consortium Architecture Working Group “OpenFog Reference Architecture
for Fog Computing.” White Paper, February 2017, Available at:
www.OpenFogConsortium.org
[3] Jose Santos, Tim Walters, Bruno Volckaeert and Filip De Turck. “Fog Computing:
Enabling the Management and Orchestration of Smart City Applications in 5G Networks”
MDPI Journal Entropy, 2017
[4] Mathias Santos de Brito, et. al. “A Service Orchestration Architecture for Fog-enabled
Infrastructures” in proc. International Conference on Fog and Mobile Edge Computing
(FMEC), 2017
[5] Karima Velasquez, et. al. “Service Orchestration in Fog Environments.” In proc.
International Conference on Future Internet of Things and Cloud, 2017
[6] ETSI, NFV: MAN 001 (v1.1.1) Network Functions Virtualisation (NFV); Management
and Orchestration, ETSI Std. GS NFV-MANO 001, 2014
Centre of Excellence
Broadband Networks and Applications 02-05 September 2018
Problem Statement
The problem that this work considers is the lack of
autonomy in management and orchestration of Fog
computing systems to enable decision making to be
made at all levels of a deployment’s hierarchy including
near the device or higher order layers. This is a vital
requirement for autonomous management systems if not
achieved they will not be able to maintain required
service level agreements, protect the execution of the
system from external attacks or prevent and recover
from failures.
University of Cape Town, Department of Electrical Engineering, South Africa
§ How can an intelligent finite state machine algorithm
improve autonomic function of Fog computing for
autonomous management system?
§ How can an intelligent decision-making be achieved
in a Fog computing network without negatively
impacting delay in decision making and energy
consumption?
§ How can current Fog computing management and
orchestration be optimised to improve autonomous
resource allocation for autonomous management
system?
Research Questions
Research Objectives
The objectives of this study are to:
§ To develop intelligent finite state machine algorithm
for Fog-based management and orchestration
§ To implement the proposed algorithms in Fog-based
networks decision-making
§ To evaluate the performance of the proposed
scheme
§ To compare the proposed scheme with the related
existing algorithms.
Fig 1: Autonomous Management and Orchestration Architecture in the Edge Node.
The proposed scheme will have a universal view of the
network under management, which will use finite state
machine to automatically switch between both centralised
and distributed mode depending on the state of the
network and the services’ requirements at the time.
Conclusion
This paper has outlined the proposed scheme for automatic
adaptation and optimisation of Fog computing based
autonomous management systems that uses finite state
machine. In addition it has also mentioned the need for
management and orchestration in Fog based networks, as
well as cost savings that can be provided by the proposed
scheme. The future work to be performed is to design and
implement the proposed scheme with the validated finite state
machine.
Fig 2: Proposed Scheme High-Level State Transition Model.
As shown in Figure 1, the proposed scheme takes into
consideration a hybrid approach including both centralised
and distributed management and orchestration models.
Cloud computing management and orchestration requires a
more centralised model, while Fog computing requires a
distributed model.