=Paper=
{{Paper
|id=Vol-1746/paper-07
|storemode=property
|title=Energy Efficiency of 5G Mobile Networks in Hybrid Fog and Cloud Computing Environment
|pdfUrl=https://ceur-ws.org/Vol-1746/paper-07.pdf
|volume=Vol-1746
|authors=Stojan Kitanov,Toni Janevski
|dblpUrl=https://dblp.org/rec/conf/rtacsit/KitanovJ16
}}
==Energy Efficiency of 5G Mobile Networks in Hybrid Fog and Cloud Computing Environment==
https://ceur-ws.org/Vol-1746/paper-07.pdf
Energy Efficiency of 5G Mobile Networks
in Hybrid Fog and Cloud Computing Environment
Stojan Kitanov Toni Janevski
Mother Teresa University Ss Cyril and Methodius University
School of Informatics Faculty of Electrical Engineering
Skopje, and Information Technologies,
Republic of Macedonia Skopje, Republic of Macedonia
stojankitanov@hotmail.com tonij@feit.ukim.edu.mk
devices, terminals, machines, and also smart things and
robots will become innovative tools that will produce
and use applications, services and data.
Abstract 5G will have to support huge mobile traffic volumes,
1000 times larger than those today in the order of
The new emerging applications in 5G network,
multiples of gigabits per second [SKT14], [Dat13],
in the context of the Internet of Everything
[Tik15], [GSA15]. The new emerging applications in
(IoE), will introduce high mobility, high
5G network, in the context of the Internet of Everything
scalability, real-time, and low latency
(IoE) [Kal15], [Bre13], will introduce high mobility,
requirements that raise new challenges on the
high scalability, real-time, and low latency
services being provided to the users.
requirements that raise new challenges on the services
Fortunately, Fog Computing and Cloud
being provided to the users.
Computing, with their service orchestration
Fortunately, Fog Computing [Bon12], [Lua15],
mechanisms offer virtually unlimited dynamic
[Vaq14], and Cloud Computing [Arm10], [Zha10],
resources for computation, storage and service
[Kit14] with their service orchestration mechanisms
provision, that will effectively cope with the
offer virtually unlimited dynamic resources for
requirements of the forthcoming services. 5G
computation, storage and service provision, that will
will use the benefits of centralized high
effectively cope with the requirements of the
performance computing cloud centers, cloud
forthcoming services. Fog Computing extends cloud
and fog RANs and distributed peer-to-peer
computing and services to the edge of the network.
mobile cloud that will create opportunities for
With its service orchestration mechanisms, it provides
companies to deploy many new real-time
data, computing, storage, and application services to
services that cannot be delivered over current
end-users that can be hosted at the network edge or
mobile and wireless networks. This paper
even end devices such as set-top-boxes or access
evaluates a model for fog and cloud hybrid
points. The main features of Fog are its proximity to
environment service orchestration mechanisms
end-users, its dense geographical distribution, and its
for 5G network in terms of energy efficiency
support for mobility.
per user for different payloads.
5G will use the benefits of the centralized cloud,
distributed cloud and fog Radio Access Networks and
1 Introduction the distributed peer-to-peer mobile cloud among the
Mobile and wireless networks have made smart devices. This will create opportunities for
tremendous growth in the last decade. This growth is companies to deploy many new real-time services that
due to the support of a wide range of applications and cannot be delivered over the existing mobile and
services by the smart mobile devices such as laptops, wireless networks [Kit16]. The core idea is to take full
smartphones, tablets, phablets, etc. This resulted with advantages of local radio signal processing, cooperative
an increased demand for mobile broadband services radio resource management, and distributed storing
[Jan15]. capabilities in edge devices, which can decrease the
Therefore, many global research and industrial heavy burden on front haul and avoid large-scale radio
initiatives are already working on the building blocks of signal processing in the centralized baseband unit pool
the next fifth generation of mobile and wireless [Chi15].
networks, also known as 5G [Jan14], [Jan09], [Tud11]. This paper presents a further extension on the previous
5G will enable the future Internet of Services (IoSs) studies given in the conference papers [Kit16], [Kit14]. It
paradigms such as Anything as a Service (AaaS), where proposes an architecture for the hybrid cloud and fog
computing environment in 5G network. Then this legacy networks for investment protection. The future
environment is explored in terms of energy efficiency. 5G system should support different types of services.
The rest of the paper is organized as follows. Section 2 The 5G mobile network will be an open service
describes the 5G requirements from different platform to bear all kinds of mobile internet
perspectives. Section 3 describes the hybrid fog and applications and it will support more flexible model of
cloud computing environment in 5G network. Section 4 operation that will enable both network operators and
evaluates this environment in terms of energy efficiency service providers to generate their own revenue.
per user for different payloads. Finally, Section 5 Two key traffic models should be considered: high-
concludes the paper and provides future work speed video flow from the server to the subscriber and
directions. massive Machine-to-Machine (M2M), or Device-to-
Device (D2D) communications [Tik15].
5G will support a wide range of applications in the
2 Service Requirements in 5G context of Internet of Everything (IoE) [Kal15],
[Bre13], and services to satisfy the requirements of the
5G will be a multi-layered heterogeneous network information society by the year 2020 and beyond. It
that will consist of existing 2G, 3G, LTE and future will have user-centric approach, where telecom
Radio Access Technologies (RATs). It may also operators will invest in developing new applications
converge many other radio technologies like mobile that will provide ubiquitous, pervasive, seamless,
satellite system (MSS), digital video broadcasting continual and versatile mobile experience to the end-
(DVB), wireless local access network (WLAN), user [Jan09]. The applications will become more
wireless personal access network (WPAN), etc., with personalized, and more context-aware and will be able
multi-tiers coverage by macro, pico, femto, relay and to recognize user identity, user location, and user
other types of small cells [Jan14]. preferences [Kit14]. The new emerging applications in
5G requirements should be defined in multiple 5G network, will introduce high mobility, high
dimensions such as technology perspective, user scalability, real-time, and low latency requirements that
perspective, network operator perspective and traffic raise new challenges on the services being provided to
models [Dat13]. the users.
From the technology perspective, 5G will be the
continuous enhancement and evolution of the present 3 Fog and Cloud Computing Environment
radio access technologies, and also the development of
novel radio access technologies to meet the increasing in 5G network
user’s demand of future. In order to satisfy 5G Requirements it is necessary
From the user’s perspective, 5G mobile system will Full Network Function Virtualization (NFV) to take
enhance user’s experience in many aspects such as: place in 5G. Network virtualization pools the
higher demand for data rate and capacity, good underlying physical resources, or logical elements in a
performances in terms of pervasive coverage, reliable network, by using the current technologies such as
QoS and battery life of the mobile device, easy to use, cognitive and software defined radios in the 5G RAN
affordable price for subscription, safety and reliability, for fog computing, and software defined networking for
and personalization of the services. 5G should provide centralized cloud services in 5G core [Mar12]. 5G in
user-centric services, where the users can customize the hybrid fog and cloud computing environment will
subscription of services and add/remove subscriptions use the benefits of the centralized cloud, cloud RAN
at his/her own will at any time. and fog RAN and the distributed Peer-to-Peer mobile
From the network operator’s point of view 5G cloud among the devices which will create
should provide sufficient bandwidth and capacity in opportunities for companies to deploy many new real-
order to support the high data traffic volume (1000 time services that cannot be delivered over current
times greater than today in the order of multiple mobile and wireless networks.
gigabits per second and at affordable cost. 5G should An overview of such 5G network architecture in a
provide low cost, easy deployment, and simple, hybrid fog and cloud computing environment is given
scalable and flexible operation in order to decrease in Figure 1. The architecture consists of centralized
CAPEX and OPEX. 5G network should provide a cloud computing nodes in 5G core, and the fog
support for backward compatibility with current and computing nodes in the 5G RAN.
Table 1 : A Comparison between Fog and Cloud
Computing Nodes [Lua15]
Fog Computing Cloud
Nodes Computing
Nodes
Target Type Mobile users General Internet
users
Service Type Limited localized Global
information services information
related to specific collected from
deployment worldwide
locations
Hardware Limited storage, Ample and
compute power and scalable storage
wireless interface space and
compute power
Distance to In the physical Faraway from
Users proximity and users and
communicate communicate
through single hop through IP
wireless connection networks
Working Outdoor (streets, Warehouse-size
Environment parklands, etc.) or building with air
indoor (restaurants, conditioning
shopping malls, etc.) systems
Figure 1: 5G Network Architecture in a Hybrid Fog and Centralized or
Cloud Computing Environment distributed in
Centralized and
regional areas by
The centralized cloud computing nodes are maintained by
Deployment local business (local
powerful, centralized and high performance Amazon,
telco vendor,
Google, etc.
computing platforms located in 5G core. They provide shopping mall
to the smart devices ubiquitous, convenient, on- retailer, etc.)
demand network access to a shared pool of
configurable computing resources (e.g., networks, [Bon12], [Lua15], [Vaq14]. They provide applications
servers, storage, applications, and services) that can be with awareness of device geographical location and
rapidly provisioned and released with minimal device context. The fog nodes support the mobility of
devices i.e. if a device moves far away from the current
management effort or service provider interaction.
servicing FCN, the fog node can redirect the
Like that the limited data processing and storage application on the mobile device to associate with a
capabilities of the mobile devices are solved by new application instance on a fog node that is now
moving both the data storage and data processing closer to the device. [ETS15]. A comparison between
away from the mobile device to the cloud computing cloud computing nodes and the fog computing nodes is
nodes [Dih11], [Qur11], [Hua11]. given in Table 1.
Fog computing nodes (FCN) are typically located FCNs absorb the intensive mobile traffic using local
away from the main cloud data centers, at the edge of fast-rate connections and relieves the long back and
the network. They extend the cloud computing at the forth data transmissions among cloud and mobile
edge of the network. Cloud computing on fog nodes devices. This significantly reduces the service latency
enables low and predictable latency. The main features and improves the service quality perceived by mobile
of fog computing nodes are their proximity to end- users, and more importantly, greatly saves both the
users, and their dense geographical distribution bandwidth cost and energy consumptions inside the
Table 2 : Energy per bit for different RAN types and
Internet backbone. Fog computing represents a
different data file size
scalable, sustainable and efficient solution to enable the
convergence of cloud-based Internet and the mobile
RAN Type
computing. Therefore, fog paradigm is well positioned
Parameter 3G 4G 5G
for real time big data analytics, 5G network, and IoT.
Energy per bit [µJ/bit]
In this environment the distributed Peer-to-Peer 100 170 17
(Data File: 10 KB)
(P2P) mobile cloud approach among the smart devices
Energy per bit [µJ/bit]
can be used [Gup11], [Kav12]. Like that a group of 4 0.3 0.03
(Data File: 10 MB)
mobile devices acts as a cloud and provides cloud
services to other mobile devices with a guaranteed
certain level of service agreements. The peers have 4.1 Energy Efficiency per User
strong capacities such as storage space, computational
The energy efficiency per user (EE), that uses fog or
power, online time, and bandwidth. The workload of
cloud computing service is a product of the energy per
the application is managed in a distributed fashion
bit which depends from the RAN type and the size of
without any point of centralization. The lack of data file being transferred to the user:
centralization provides scalability, while exploitation of
user resources reduces the service cost. P2P
EE EranT (1)
architectures have ability to adapt to network failures
and dynamically changing network topology with a
transient population of nodes/devices, while ensuring where,
acceptable connectivity and performance. Thus, P2P Eran is the energy per bit that depends from the type of
systems exhibit a high degree of self-organization and the RAN;
fault tolerance. and T is the size of the payload 10 KB or 10 MB.
The values for the energy per bit for different types
4 Architecture Evaluation RAN networks is provided in [Hua12], and are
The performances of the hybrid fog and cloud summarized in Table 2. Here it is assumed that 5G
computing environment in 5G can be explored in many RAN will have 90% improvement in energy per bit
ways such as Round Trip Time (RTT) latency, over 4G [Tik15].
throughput, product latency – throughput, energy
efficiency and power consumption. The focus in this 4.2 Analysis of the Results
paper is the energy efficiency per user for different data Our simulation scenario consists of the following: 10
payloads: 10 KB and 10 MB. The most significant cloud computing centers, three types of RANs are
impact in the energy efficiency will have the RAN type, considered (3G, 4G, and 5G), and the number of the
while the 5G core impact on the energy efficiency can users is varied from 100 to 1000. For simplicity the
be treated as a constant, and therefore it can be impact of the distance between the smart user device
neglected. and the RAN on energy efficiency was neglected. The
The following scenario will be used. There is a simulation results are provided in Figure 2. The
region that contains a group of N users uniformly following can be noticed.
distributed, which are simultaneously covered by 3G RAN wastes a lot of energy for the transfer of
several different RANs. Each RAN is connected to big data files. 4G RAN provides much better energy
several clouds, which can be in the same or different efficiency for large data files, compared to 3G RAN.
region with the RANs. The smart user devices are On the other hand, 4G RAN wastes energy for the
assumed to be equally capable, and are located on a transfer of small data files, and 3G RAN demonstrates
different distance from the RANs. They can be better performances. Finally, 5G RAN has the best
simultaneously served by the RANs and the clouds. energy efficiency for the transfer independently from
the size of data files, that the user is requesting them
from the fog or cloud.
References
[Jan15] T. Janevski. Internet Technologies for Fixed
and Mobile Networks. Artech House, USA,
2015.
[Jan14] T. Janevski. NGN Architectures Protocols and
Services. John Wiley & Sons, UK, 2014.
[Jan09] T. Janevski. 5G Mobile Phone Concept.
Proceedings of 6th IEEE Consumer
Communications and Networking Conference
- CCNC 2009, pp. 1–2, Las Vegas, Nevada,
USA, 2009.
Figure 2: Energy Efficiency in 5G Network in the Hybrid [Tud11] A. Tudzarov and T. Janevski. Functional
Fog and Cloud Computing Environment Architecture for 5G Mobile Networks.
International Journal of Advanced Science and
In 5G network where the user device will be served Technology (IJAST), vol. 32, pp. 65–78, July
by different RANs, it has to make choice which RAN 2011.
will be the most suitable for transferring data files. The [SKT14] SK Telecom Network Technology Research
choice should be make primarily of the size of data
and Development Center 5G White Paper, SK
files being transferred, throughput, latency, energy
Telecom’s View on 5G Vision, Architecture,
efficiency, etc. The algorithm for such RAN selection
will be our future work direction. Technology, and Service and Spectrum, SK
Telecom, October 2014.
5 Conclusion [Dat13] Datang Mobile Wireless Innovation Center 5G
White Paper, Evolution, Convergence and
This paper evaluates energy efficiency per user in Innovation, Datang Telecom Technology and
different payload and networks. The results show that Industry Group, December 2013.
5G RAN has the best energy efficiency for the transfer
independently from the size of data files, compared to [Tik15] V. Tikhvinskiy and G. Bochechka. Prospects
3G and 4G RAN. and QoS Requirements in 5G Networks.
5G network will act as a nervous system of the Journal of Telecommunications and
digital society, economy, and everyday people’s life. Information Technologies, Vol. 1, No. 1, pp.
The cloud in 5G networks will be diffused among the 23 – 26, 2015.
client devices often with mobility too, i.e. the cloud will [GSA15] The Road to 5G: Drivers, Applications,
become fog. Requirements and Technical Development. A
More and more virtual network functionality will be GSA (Global mobile Suppliers Association)
executed in a fog computing environment, and it will Executive Report from Ericsson, Huawei and
provide mobiquitous service to the users. This will Qualcomm, 2015.
enable new AaaS service paradigms, where devices, [Kal15] V. L. Kalyani, and D. Sharma. IoT: Machine to
terminals, machines, and also smart things and robots Machine (M2M), Device to Device (D2D)
will become innovative tools that will produce and use Internet of Everything (IoE) and Human to
applications, services and data. Human (H2H): Future of Communication.
Finally, the choice of selecting the most suitable Journal of Management Engineering and
RAN, should be make primarily of the size of data files Information Technology (JMEIT) Volume -2,
being transferred, throughput, latency, etc. The Issue 6, pp. 17 – 23, Dec. 2015.
algorithm for such RAN selection will be our future
work direction.
[Bre13] B. Brech, J. Jamison, L. Shao, G. Wightwick. 18, Issue 2, pp. 147–164, Springer, USA,
The Interconnecting of Everything. IBM February 2012.
Redbook, 2013. [Dih11] T. H. Dihn, C. Lee, D. Niyato, and P. Wang. A
[Bon12] F. Bonomi, R. Milito, J. Zhu, S. Addepalli. Survey of Mobile Cloud Computing:
Fog Computing and its Role in the Internet of Architecture, Applications, and Approaches.
Things. Proceedings of the First Edition of the Wireless Communications and Mobile
ACM SIGCOMM Workshop on Mobile Cloud Computing, Wiley, Vol. 13, Issue 18, pp.
Computing (MCC 2012), pp. 13-16, Helsinki, 1587–1611, 2011.
Finland, 2012. [Qur11] S. S. Qureshi, T. Ahmad, K. Rafique, and
[Lua15] H. T. Luan, L. Gao, Z. Li, L. X. Y. Sun. Fog Shuja-ul-islam. Mobile Cloud Computing as
Computing: Focusing on Mobile Users at the Future for Mobile Applications –
Edge. arXiv:1502.01815[cs.NI], 2015. Implementation Methods and Challenging
[Vaq14] L. M. Vaquero, L. Rodero-Merino. Finding Issues. Proceedings of IEEE Conference on
your Way in the Fog: Towards a Cloud Computing and Intelligence Systems
Comprehensive Definition of Fog Computing. (CCIS), pp. 467–471, Beijing, China, 2011.
ACM SIGCOMM Computer Communication [Hua11] D. Huang et al. Mobile cloud computing.
Review Newsletter, Vol. 44, No. 5, pp. 27-32, IEEE COMSOC Multimedia Communications
2014. Technical Committee (MMTC) E-Letter, Vol.
[Arm10] M. Armbrust et al. A view of cloud computing. 6, No. 10, pp. 27–31, 2011.
Communications of the ACM, vol. 53, No.4, [ETS15] Fog Computing and Mobile Edge Cloud Gain
pp.50–58, April 2010. Momentum Open Fog Consortium. ETSI
[Zha10] S. Zhang, S. Zhang, X. Chen, X. Huo. Cloud MEC and Cloudlets, Version 1.1 Guenter I.
Computing Research and Development Trend. Klas, November 22, 2015.
Proceedings of the Second IEEE International [Gup11] A. Gupta, and L. K. Awasthi. Peer-to-Peer
Conference on Future Networks (ICFN 2010), Networks and Computation: Current Trends
pp. 93-97 Sanya, Hainan, 2010. and Future Perspectives. Journal of
[Kit14] S. Kitanov and T. Janevski. State of the Art: Computing and Informatics, Vol. 30, pp. 559–
Mobile Cloud Computing. Proceedings of the 594, 2011.
Sixth IEEE International Conference on [Kav12] H. Kavalionak and A. Montresor. P2P and
Computational Intelligence, Communication Cloud: A Marriage of Convenience for
Systems and Networks 2014 (CICSYN 2014), Replica Management. Proceedings of the 6th
pp. 153–158 Tetovo, Macedonia, 2014. IFIP TC 6 International Conference on Self-
[Kit16] S. Kitanov, E. Monteiro, T. Janevski. 5G and Organizing Systems, pp. 60–71, Delft,
the Fog – Survey of Related Technologies and Netherlands, 2012.
Research Directions. in Proceedings of the [Hua12] J. Huang, F. Qian, A. Gerber, Z. M. Mao, S.
18th Mediterranean IEEE Electrotechnical Sen, O. Spatscheck. A Close Examination of
Conference MELECON 2016, pp. 1-6, Performance and Power Characteristics of 4G
Limassol, Cyprus, 2016. LTE Networks. Proceedings of the 10th
[Chi15] M. Chiang: Fog Networking: An Overview on international conference on Mobile systems,
Research Opportunities, white paper, 2015. applications, and services (Mobisys 2012), pp.
225-238, 2012.
[Mar12] J. Marinho and E. Monteiro. Cognitive radio:
Survey on Communication Protocols,
Spectrum Decision Issues, and Future
Fesearch Directions. Wireless Networks, Vol.