Improvement of QoS in an IoT Ecosystem by Integrating Fog Computing and SDN

Improvement of QoS in an IoT Ecosystem by Integrating Fog Computing and SDN

Ishtiaq Ahammad, Md. Ashikur Rahman Khan, Zayed Us Salehin, Main Uddin, Sultana Jahan Soheli
Copyright: © 2021 |Pages: 19
DOI: 10.4018/IJCAC.2021040104
OnDemand:
(Individual Articles)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

The internet of things (IoT) creates immense volume of objects online. But cloud computing isn't suited to environmental demands. Hence, fog computing (FC) emerged which shifts the computation load into edge fog devices. However, FC also faces some obstacles which can be mitigated by software-defined networking (SDN). By combining SDN and FC, the network form can overcome almost all cloud limitations and can boost QoS. Within this article, architecture is proposed by combining SDN and FC to improve QoS for IoT ecosystem. With the architecture, an algorithm is propounded based on virtual partition. Then a use case is presented and evaluated through iFogSim simulator. The result shows a significant improvement of several QoS parameters in the execution of fog with SDN compared to the cloud-only execution. The results also show better results for energy consumption, network use (212.21% reduction), and latency (275.9% reduction) compared with previous similar use case.
Article Preview
Top

Introduction

The IoT is now set into billions of things devices worldwide that are now connected to the internet. IoT can range from something as small as a medicine pill to something as large as a vessel. All of these devices are gathering and share data. The data won’t stay raw as it is today. Rather than the data will be personalized to the users according to their necessity and even overlap with diverse data (Lee et al., 2017). More and more devices are moving to the IoT and now it evolves into Internet of Everything. Studies show that, by 2025, there are expected to be more than 75 billion IoT devices operating and are projected to yield 79.4 ZB of data (Mass, 2020). Coping with this rapid expansion requires preparation and planning.

The gap between end devices and the data centers (DCs) is not substantially affected by current cloud services when it’s basically intended for leading web-based applications. On the contrary, numerous new IoT services needs connectivity in real-time and support for mobility; which makes network latency a significant limiting factor. When traffic load is distributed equally upon the network, the effect of delay could have been minimized (Tomovic et al., 2017). Sadly, in the existing internet architecture defined by the distributed control plane, progressive routing is seen as more risky than beneficial. Also cloud suffers from massive IoT generated data. Therefore, existing cloud computing has to expand.

Fog computing allots device, storage, power and networking resources and services anywhere towards the cloud to things continuum. So fog actually brings the cloud nearby edge devices. Fog offers local data processing and storage for things devices rather than dispatch them into cloud. Unlike cloud, fog gives quicker reaction and higher quality services. Therefore, fog can be marked as the ideal option for enabling the IoT to deliver feasible and stable services (real-time services) to plenty of IoT users in an IoT ecosystem. Fog is distinguishable from cloud by its closeness to end-users, its concentrated geographical distribution and mobility support (Stojmenovic & Sheng, 2014). Fog also narrows greatly the usage of backbone network bandwidth. Consequently, the shortened service costs will favor the users. A Fog server handles 3-D arrangements including: Storage, Networking and Computing. Fog networking principally uses regional computer resources instead of using outlying computer resources. This leads to a reduction in latency problems.

The service orchestration cites a key challenge assigned by the notion of FC. Because of models of regular or eventual data delivery, many IoT systems deal with complex workload. Ideally, applications must be transparently scaled during runtime except oversupply of the resource. SDN can be used to achieve this, instead of traditional networks. In addition, the hardware resource networking has not rise to keep pace with the cloud while computing hardware has. It can’t be scalable or quick to deploy. The challenges facing network providers are enormous. Because they are forced to grow in line with customer need. The primary challenges are meeting rising bandwidth demand and rapid introduction of new client services. This means network providers need not only a flexible network, but also a clever one and that is where SDN comes in.

The boundaries of traditional networking in terms of QoS can be mitigated through SDN. Managing QoS is one of those critical networking issues that have not yet come up with an acceptable solution. For IoT ecosystem, QoS parameters must be defined accurately. Only then consumers would be eligible to use these parameters to understand and reveal their specifications. There are many IoT devices, and they have growing functions in everyday lives. Thus IoT ecosystems QoS expectations can differ from one application to another. QoS assists to control the device functionality and resources needed to deliver IoT services. It helps service providers to give customers strong transparency of their facilities, as well as efficiency and availability of the facilities. Service Level Agreement (Singh & Baranwal, 2018) among IoT service provider & IoT consumer can be implemented by utilizing QoS. QoS parameters can help consumers figure out the ideal IoT solution for their use. QoS essentially has a lot of parameters in it. A number of QoS problems are still unanswered and others have room to change as they are conflicting to their present state. These QoS problems can be solved by integrating SDN with FC and can be improved where necessary. Four parameters of QoS (network usage, cost, latency & energy consumption) are considered and evaluated in this article through iFogSim.

Complete Article List

Search this Journal:
Reset
Volume 14: 1 Issue (2024)
Volume 13: 1 Issue (2023)
Volume 12: 4 Issues (2022): 2 Released, 2 Forthcoming
Volume 11: 4 Issues (2021)
Volume 10: 4 Issues (2020)
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing