Conference PaperPDF Available

Efficient Energy Management using Fog Computing



Smart Grid (SG) is a modern electricity network that promotes reliability, efficiency, sustainability and economic aspects of electricity services. Moreover, it plays an essential role in modern energy infrastructure. The main challenges for SG are, how can different types of front end smart devices, such as smart meters and power sources, be used efficiently and how a huge amount of data is processed from these devices. Furthermore, cloud and fog computing technology is a technology that provides computational resources on request. It is a good solution to overcome these obstacles, and it has many good features, such as cost savings, energy savings, scalability, flexibility and agility. In this paper, a cloud and fog based energy management system is proposed for the efficient energy management. This frame work provides the idea of cloud and fog computing with the SG to manage the consumers requests and energy in efficient manner. To balance load on fog and cloud a selection Base Scheduling Algorithm is used. Which assigns the tasks to VMs in efficient way.
Efficient Energy Management using
Fog Computing
Muhammad KaleemUllah Khan, Nadeem Javaid, Shakeeb Murtaza,
Maheen Zahid, Wajahat Ali Gilani, and Muhammad Junaid Ali
Smart Grid (SG) is a modern electricity network that promotes reliability,
efficiency, sustainability and economic aspects of electricity services. More-
over, it plays an essential role in modern energy infrastructure. The main
challenges for SG are, how can different types of front end smart devices,
such as smart meters and power sources, be used efficiently and how a huge
amount of data is processed from these devices. Furthermore, cloud and fog
computing technology is a technology that provides computational resources
on request. It is a good solution to overcome these obstacles, and it has many
good features, such as cost savings, energy savings, scalability, flexibility and
agility. In this paper, a cloud and fog based energy management system is
proposed for the efficient energy management. This frame work provides the
idea of cloud and fog computing with the SG to manage the consumers re-
quests and energy in efficient manner. To balance load on fog and cloud a
selection Base Scheduling Algorithm is used. Which assigns the tasks to VMs
in efficient way.
Key words: Energy Management Controller, Macro Grid, Micro Grid,
Smart Grid, Internet of Things (IoT), Fog Computing, Cloud Computing.
1 Introduction
Cloud computing is a new concept in the area of information and commu-
nication technology. Cloud computing is a globally distributed network. It is
the combination of memory, soft ware processing units, Physical Machines
(PM) and some networks. There are many VMs in PM according to cloud
COMSATS University, Islamabad 44000, Pakistan
2 Muhammad KaleemUllah Khan et al.
infrastructure [1]. Cloud computing provides some kind of services to the
consumers. Consumers get benefits from these services according to their de-
mand. Tasks are coming from the consumer side. Cloud analyzes these task
types and allocate the resources to the consumers, according to consumer
Cloud computing concept firstly introduced in 2010 [2]. It is a geographical
distributed network. It offers facilities that can be simply managed with less
effort over the Internet. We can say, cloud computing provides delivery of
hosted services on the Internet. It is merger of networks, storage and virtual
machines to compute, process, store and sharing data. Each component of
this system has its own importance. Authors in [3] work on cloud computing
is based on three main services.
Infrastructure as a Service (IaaS),
Platform as a Service (PaaS),
Software as a Service (SaaS).
IaaS helps to avoid the expense and complications of purchasing and man-
aging his physical servers such as your data center infrastructure [4]. Each
component has its own services. Customers can rent these services as needed.
Basically, this service is based on infrastructure. In this service client has its
own operating system, software and special applications.
PaaS is a complete cloud development and deployment environment [5].
Its resources allow, to provide everything from simple cloud applications to
enterprise and cloud applications. Customers can purchase cloud resources
from cloud providers as needed. Access these resources across the Internet.
The PaaS environment, including IaaS. PaaS provides all IaaS services and
some other services such as database management systems, deployment tools,
business intelligence and middleware programs.
SaaS is a complete software environment that provides users with cloud-
based applications. In this service system, all IaaS services are also included.
Customers can purchase or rent these services through agreements or licenses
within a certain period of time. The availability and security of customer
data are the responsibility of the service provider. It provides different kind
of services as Drop box, Slack and Docusign etc[6].
All smart devices connected with the Internet, these devices are called
Internet of Things (IoT) [7]. Nowadays, every person in this world is directly
or indirectly connected to IoT. Authors in [8] say cloud users are increasing
every day due to increasing no of IoT devices. These IoT devices generate
large amounts of data. These numbers of users impact the efficiency of cloud.
It is a big problem for cloud to compute, analyze and store this large amount
of data, latency and delay also increase. It becomes a big task to handle this
large amount of data.
To overcome the burden of cloud, fog computing introduced to extend
cloud computing. Fog computing is a virtualized environment also called fog-
ging or edge computing [9]. Cisco introduced fog computing. It becomes a
Efficient Energy Management using Fog Computing 3
middle layer between cloud and end user. Fog is not a centralized environ-
ment, with the merger of fog efficiency of cloud increases. Fog acts like cloud,
fog provides data, storage, analyses and application services to the users. Fog
computing is introduced to overcome the latency and delay. Some processes
need very quick response like Smart Grid (SG), software defined network and
smart light in vehicular networks. SG is an electric grid of modern era. There
are many smart meters in a SG. Smart meters are used to identify the electric-
ity usage and number of appliances in a home. The work in [10],[11]and[12]
has integrated cloud and fog computing with SG. In the real world scenario
SG provides multiple facilities to users. When a Micro Grid (MG) is attached
with this scenario, it becomes more helpful for users and utility.
we proposed cloud and fog based SG scenario. Requests are coming from
consumers end to utility. When a user sends an electricity request, it makes
a link between user, fog and cloud. Fog receives consumer request and fog
explores the nearest MG and MG fulfill the demands of the user. In this paper
users send a request to fog and fog refers this request to the MG of its region.
MG analyzes the request of user, if MG is unable to fulfill the users need, this
request is sent to a centralized cloud. Centralized cloud is attached with a
Macro Grid. Cloud fulfill the demands of users with the help of Macro Grid.
Large no of requests from the users end create a load on fog. Authors in [13]
present cloud and fog based load balancing by use two techniques round robin
and throttled algorithm for efficient load balancing. In this authors take six
regions, each region has two fogs and two clusters of buildings.
In this paper six regions are considered, each region has two fogs and
three clusters of buildings. Selection base scheduling algorithm is used for
the reduction or balance the load on fog. Basic working of this algorithm
is load balancing on fog side, by scheduling Virtual Machines (VM). This
algorithm assigns task to VM by analyzing which VM is available or with
minimum load.
1.1 Motivation
Now a day, energy becomes the basic need of the society and demand of
energy increase with the passage of time. Renewable energy resources are
used to overcome these demands [14]. Many authors work on cloud and fog
based energy management in different scenarios. To achieve better response
and processing time and efficient management. Cloud and fog computing is
also a reliable and scalable environment. As [13] authors presents a cloud
and fog based architecture for the distribution of resources in smart cities.
Authors used one fog to manage one cluster of buildings. Author also use
Round Robin Algorithm for efficient load management. Load is distributed
to all virtual machines so that the RT of cloud and fog is decreased. In this
paper two fogs manage three cluster of building. In this paper a selection base
4 Muhammad KaleemUllah Khan et al.
scheduling algorithm is used to achieve better RT and minimum cost. Rest
of the paper based on: section II is based on Literature Review, Section III is
based on System Model, Section IV is based on Simulations and Discussions
and the V portion of this paper is based Conclusion. References used in the
paper are mentioned at the end of paper.
2 Related Work
Cloud computing is a virtualized environment including virtual machines,
data centers and soft ware processing unit. Different types of processes are
processed in cloud. Increasing number of users are main cause of delay and
latency in processing and RT. To avoid delay and latency fog computing
introduced, it shears the load of cloud. There are many algorithms used to
overcome the load of cloud and fog. In this work [13] authors use Round
Robin and Throttled algorithm for efficient management of load on cloud
and fog. These algorithms schedule the tasks, requests and virtual machines
for load balancing.
The authors of [15] wants to achieve the Quality of Service (QoS). First, for
the heterogeneous cloud data-centers authors designed a QoS based energy
consumption model to achieve the quality of the service. Second, authors
purposed a consistent task scheduling technique to minimize the consumption
of energy in a data center with the help of QoS aware Physical Machines
(PM) selection method. Finally, analyze the performance of this proposed
technique. Authors also work on some mathematical model to achieve the
efficient response time and throughput. Response time is a time which is
required to send a request to the cloud and get the response from the cloud.
In [16] authors present a technique Cloud Demand Response (CDR) to
overcome the issues of Distributed Demand Response (DDR). CDR is a two
layer cloud platform and DDR is a k cluster based environment. There are
some issues in DDR, when a wireless communication network is used in com-
munication channel, it becomes a cause of bit error rate. Due to this bit
error rate they effect the overall demand response program. In clustered en-
vironment bit error rate is high. It requires high bandwidth. DDR depends
on the iteration, and each time an update message is send among all users
of the same group. After some iteration, the performance of this channel is
not unique. It becomes the cause of message loss and delay. It also requires
high bandwidth for the message delivery. Features of CDR is, it is an inde-
pendent communication channel. It gives better cost analysis. It reduces the
total communication cost, peak to average ratio, convergence time and use
of bandwidth. If user demand high efficient response time it may cause high
Author in [17] presents a distributed algorithm to respond to demands
which are made from consumer end for real time. Each utility company and
Efficient Energy Management using Fog Computing 5
local users solve the sub-problems of the allocation process. For supply and
demand balance, each competitive company evaluates clearing price.
In [18] authors want to achieve some goals. Author presents Home Energy
Management (HEM). To implement power management system, a platform is
needed, which offers interactive interoperability between devices and process
elasticity. HEM is applied over a network platform to meet these require-
ments at lowest cost. However, the operating areas of the system are: scala-
bility, heterogeneity and delay sensitive devices. Cost element not considered
Authors in [19] present a fog base model for SG management. This system
is divided in to three layers, first layer is based on smart grids, second layer
is based on fog layer and the third layer is based on a centralized cloud layer.
This system makes SG more reliable. It is a geographical distributed system.
It provides locality and reliability to smart grid. This fog system increases
the efficiency of cloud based SG system. This system increases privacy and
reduce latency. This system is also beneficial in terms of control energy flow
and balance energy load.
In this work [20] authors present a fog based energy management system.
The term fog is introduced to divide the load of cloud. Peoples are directly
or indirectly connected to the Internet through IoT. These IoT produce large
amount of data. This large amount of data, create a problem for cloud. Au-
thor introduced fog in his system to overcome the delay, processing time,
efficient sharing and management of resources and load balancing. Energy
is managed on boat user and utility side. User side Energy management is
called HEM and the utility side management is called Micro grid Level Man-
agement (MLM). In HEM energy is further managed in two different ways,
HVAC controller and EV charging controller. Smart transfer controller man-
ages MLM energy. Energy is managed on two sides. This two way energy
management, reduces the setup cost, process computation time and power
In this paper [21] authors proposed a bio inspired algorithm which is work-
ing on multi rumen. Tasks are allocated to related or suitable VM through
anti grazing principle. In multi data center environment each data center is
called one rumen. Rumens are the VMs, these VMs are used to complete
the tasks. Authors design an algorithm for multi data centers load and task
balancing in the cloud. The authors also analyze the some load balancing
techniques through Task Completion Time (TCT) and Task Execution and
Completion Time (TECT). Task completion time is a time required, when a
virtual machine is assigned to a task and meet the task requirements of the
task. TECT is a time required, when a task is coming from the consumer end
including network delay and task completion time. The authors also analyze
the randomly created datasets with this proposed algorithm. The authors
also analyze the performance of this proposed technique with some existing
6 Muhammad KaleemUllah Khan et al.
In above related work authors work on load balancing on cloud and fog
through different techniques and algorithms. Some authors work on load bal-
ancing of cloud, some authors work on efficient management and allocation of
VMs. Authors introduced some service broker polices for the efficient man-
agement of cloud resources. We are also working load balancing on cloud
through efficient allocation of VMs to the consumers requests.
3 Proposed Model
Cluster 2
Cluster 18
F 1F 2 F11 F12
Cluster 1
Macro Grid
Clusters Layer
Fog Layer
Cloud and Macro Grid Layer
BuildingsBuildings Buildings
Two way communicator
Fig. 1: Proposed System Model
This proposed system is based on three layer architecture, first layer is
based on EMC (Energy Management Controller) on buildings and this EMC
is connected with MG, second layer is based on fog and third layer is based
Efficient Energy Management using Fog Computing 7
on centralized cloud network. This centralized cloud network is connected
with macro grid and service provider. Fog becomes the middle layer between
cluster and cloud layer to operate the requests of cloud and reduce the bur-
den on cloud. In other words EMC communicates with fog network and fog
network communicates with cloud. Each cluster is comprised of 60 to 120
buildings and each building is comprised of 100 to 200 flats. Each cluster of
the buildings has EMC to manage the requests of the consumers and respond
to the requests. The demand of electricity of each flat request to EMC, then
the EMC communicates with fog and fog refers this electricity request to
MG. After this MG respond to the request and provides required electricity
to the particular flat. When MG is not capable to meet the requirement of
the customers demand, then fog communicates with cloud platform to deliver
the closer Macro Grid facility to the particular consumer. Fogs are located in
various regions of the world and the world is divided into six major regions
and these regions consist of six continents as elaborated in Table 3.
Two fogs are located in each region and these fogs are capable to entertain
needs of the three clusters located within the same region. The MGs are
attached withe the fogs, Consumers request to the fog to fulfill the energy
need. Fogs are capable to forward the requirement of energy according the
needs of the consumer. Consumers are not capable to deliver their request
to MG directly to meet their energy requirement. The fog send consumer’s
requests to MG to fulfill their needs. The MG sends energy to meet the
requirement of the consumer. When the consumer request to fog for the
energy. The fog now choose the MG that requires power to meet the consumer
demands or the fog will release a request to the neighboring MG. When a
high demand sends to all MG they will not be able to meet the requirements,
then the fog will request to the cloud, and cloud will fulfill the consumers
demand with the help of Macro Grid.
3.1 Problem Formulation
This system is based on cloud and fog based Smart Grid (SG) model. Im-
proving the efficiency of the system is the main task. Efficiency of the system
is improved by the efficient allocation of resources to the tasks. According
to this scenario there are three clusters of buildings attached with fog and
each cluster has N number of buildings and these buildings have M num-
ber of homes. These homes generate requests and sent to the fog. Here the
available requests are presented by this set T={t1, t2, ..., tn}on each fog.
V={v1, v2, ..., vm}represents the available virtual Machines (VM) on each
The total number of available tasks on each fog can be written as.
8 Muhammad KaleemUllah Khan et al.
T total =
The total number of available VMs on each fog are calculated as.
V total =
Our objective function:
Minimum: P Ttotal
Minimum: RTtotal
Minimum: T otalc.
If the value is 1 then assign the task otherwise task is not assigned according
to equation 3.
Tassing =(1; Task assing
otherwise; Task not assign (3)
VM current status will be checked as.
Vassign =(0,if the V is free
1,if the V is not free (4)
Total number of tasks divided by total available number of VM to get the
processing time according to equation 5.
P T i, j =T total
V total (5)
The total processing time formally represented as.
P Ttotal =
(P Taverage Vassig n) (6)
Response time is obtain from equation 7 by subtracting the Arrival time of a
task through the sum of task delay time and task finish time. In this equation
time is taken in seconds s represents the seconds.
RT =Delay(s) + F inish(s)Arrival(s) (7)
Equation 8 gives the total cost of the proposed system model. Here c rep-
resents the cost. Sum of data-center cost, total VM cost and total MG cost
represents total cost of the system.
T otalc=Datacenterc+T otalVc+T otalM Gc(8)
Efficient Energy Management using Fog Computing 9
4 Simulations
4.1 Setup And Parameters
Simulations of the scenario are completed by using a tool Cloud Analyst
and the system specification is given below in table 1. Six regions of the world
Table 1: System Specification
Components Power
Processor i5-4200u CPU @ 1.60GHz
RAM 4.0 GB
Hard Drive 3 500 GB
Graphic Card 4.0 Gb
are considered in this paper. Each region has two fogs and three building clus-
ters. Fogs are directly connected to a central cloud system and clusters and
fogs of a region are connected with each other. When a consumer of cluster
building sends his request of electricity to fog. At that time, many other con-
sumers also send their requests to fogs to fulfill the electricity demands. Huge
number of requests are the problem for fog. As we know fog is based on virtual
machines and virtual memory. All requests of the consumers are processed
by the virtual machines. When the number of requests are very high. It is
a difficult task for fog to decide which virtual machine is assigned to which
request. To handle this situation we implement selection base scheduling to
reduce the load of fog with the help of assigning virtual machines.
Table 2 shows the division of regions. In each region two fogs and three
clusters are placed. Showing the regions one to six respectively and fogs 1
to 12 respectively, with the 1 to 18 clusters. According to this table 2 fogs
manage 3 clusters of buildings. Each cluster is connected with each fog with
in a region. If one fog is busy with the other clusters requests, so, cluster will
communicate with the other fog.
Table 2: Regions Division
Regions Fogs Clusters
Region 1 F-1 F-2 C-1 C-2 C-3
Region 2 F-3 F-4 C-4 C-5 C-6
Region 3 F-5 F-6 C-7 C-8 C-9
Region 4 F-7 F-8 C-10 C-11 C-12
Region 5 F-9 F-10 C-13 C-14 C-15
Region 6 F-11 F-12 C-16 C-17 C-18
10 Muhammad KaleemUllah Khan et al.
Table 3 shows the minimum RT, average RT and maximum RT of each
fog. This table also show which fog requires maximum, minimum and average
response time. According to this table fog 10 has maximum response time.
Fog 1 has minimum average response time. This table shows how much time
is required by fog to respond the consumer’s request.
Table 3: Fogs Performance
Fog Average RT
Minimum RT
Maximum RT
Fog1 1.06 0.02 2.51
Fog2 1.58 0.02 4.06
Fog3 1.61 0.03 3.89
Fog4 1.63 0.03 3.99
Fog5 1.58 0.02 3.52
Fog6 1.58 0.02 4.05
Fog7 1.58 0.02 4.15
Fog8 1.56 0.02 4.08
Fog9 1.63 0.04 4.23
Fog10 1.60 0.04 4.27
Fog11 1.55 0.02 4.06
Fog12 1.58 0.02 4.07
The time required to connect is called RT. Figure 2 is a graphical rep-
resentation of 12 fogs RT. That shows the minimum RT, average RT and
maximum RT of each fog. This figure also show which fog requires maxi-
mum, minimum and average response time. According to this figure fog 10
has maximum response time. Fog 1 has minimum average response time.
Fog1 Fog2 Fog3 Fog4 Fog5 Fog6 Fog7 Fog8 Fog9 Fog10Fog11Fog12
Response Time
Average RT
Minimun RT
Maximum RT
Fig. 2: RT of each Fog
Efficient Energy Management using Fog Computing 11
Table 4 shows the minimum RT, average RT and maximum RT of each
cluster. According to this table cluster 7 has maximum response time to
the users requests. According to this table cluster 1 has minimum average
response time. This table shows how much time is required by cluster to
respond the consumer’s request.
Table 4: Clusters Performance
Cluster Average RT
Minimum RT
Maximum RT
Cluster 1 50.82 38.95 61.32
Cluster 2 50.90 39.08 63.88
Cluster 3 51.06 39.03 63.36
Cluster 4 51.18 38.65 61.81
Cluster 5 51.34 38.57 63.76
Cluster 6 51.36 38.44 64.06
Cluster 7 51.27 37.57 67.85
Cluster 8 51.23 40.16 63.34
Cluster 9 51.37 39.13 64.83
Cluster 10 51.21 36.89 62.88
Cluster 11 51.11 41.12 65.04
Cluster 12 51.28 39.85 64.92
Cluster 13 51.31 37.72 66.89
Cluster 14 51.42 41.57 65.80
Cluster 15 51.52 40.48 64.68
Cluster 16 51.30 39.00 65.86
Cluster 17 51.22 39.37 61.62
Cluster 18 51.25 39.63 65.38
Figure 3 is a graphical representation of 18 cluster of buildings RT of. The
response time of each fog is also represented in this figure. That shows the
minimum RT, average RT and maximum RT of each cluster. According to
this figure cluster 7 has maximum response time to the users requests. This
figure shows cluster 1 has minimum average response time. This figure shows
how much time is required by cluster to respond the consumer’s request. It
also shows that which fog gives better RT to consumer’s request.
Table 5 shows the required cost of each data center which is presented on a
fog. According to this table all fogs have same VMs cost. Fog 2 has minimum
MG, data transfer and total cost. Fog 4 has maximum MG, data transfer and
total cost. In this table cost of VM, MG, data transfer cost and total cost
which is required to create a data center on a fog.
Figure 4 is a graphical representation of cost of each fog, showing that
which fog gives the minimum cost. This fig shows the required cost of each
data center which is presented on a fog. According to this figure all fogs have
same VMs cost. Fog 2 has minimum MG, data transfer and total cost. Figure
4 shows fog 4 has maximum MG, data transfer and total cost. In this table
12 Muhammad KaleemUllah Khan et al.
C1 C2 C3 C4 C5 C6 C7 C8 C9
User Base
Response Time (ms)
Average Response Time (ms)
Minimun Response Time (ms)
MaximumResponse Time (ms)
Fig. 3: RT of each Cluster
Table 5: Cost of each Fog
Data Center VM Cost Microgrid Cost Data Transfer
Total Cost
Fog1 16.80 196.74 12.30 225.83
Fog2 16.80 195.66 12.23 224.69
Fog3 16.80 270.14 16.88 303.82
Fog4 16.80 293.51 18.34 328.65
Fog5 16.80 231.09 14.44 262.33
Fog6 16.80 240.53 15.03 272.36
Fog7 16.80 258.27 16.14 291.21
Fog8 16.80 247.83 15.49 280.12
Fog9 16.80 277.25 17.33 311.37
Fog10 16.80 269.82 16.86 303.49
Fog11 16.80 222.70 13.92 253.42
Fog12 16.80 227.91 14.24 258.95
cost of VM, MG, data transfer cost and total cost which is required to create
a data center on a fog.
5 Conclusion
Proposed system model is based on cloud and fog is presented to manage
the huge number of requests from the consumers on fog. For this we used six
regions of the world. The goal of the technique is to mitigate load on cloud.
An EMC is used in the cluster of buildings for the efficient management of
the energy. This EMC tells the demand of energy of each cluster to the fog.
Efficient Energy Management using Fog Computing 13
Fog1 Fog2 Fog3 Fog4 Fog5 Fog6 Fog7 Fog8 Fog9 Fog10Fog11Fog12
Cost of Each Fog
Cost ($)
Virtual Machine Cost
Migrogrid Cost
Data Transfer Cost
Total Cost
Fig. 4: Cost of each Fog
The Large Numbers of requests form the electricity consumers creates a load
on the fog and cloud. to over come this situation a selection based schedul-
ing algorithm is used. This technique is used for efficient VMs allocations
to the user’s request and balance the load on fog and cloud. After applying
this selection base scheduling algorithm on the proposed system model min-
imum RT. minimum request PT and minimum cost of the proposed system
is achieved. in short selection base scheduling algorithm is RT,PT and cost
efficient. The implementation of proposed model also provides some features
which are interoperability, connectivity and flexibility. Java platform is used
to perform simulations in eclipse platform.
1. Chris D Hyser and Bret A McKee. Starting up at least one virtual machine in a
physical machine by a load balancer, May 22 2012. US Patent 8,185,893.
2. Michael Armbrust, Armando Fox, Rean Griffith, Anthony D Joseph, Randy Katz,
Andy Konwinski, Gunho Lee, David Patterson, Ariel Rabkin, Ion Stoica, et al. A view
of cloud computing. Communications of the ACM, 53(4):50–58, 2010.
3. Santosh Kumar and R. H. Goudar. Cloud Computing Research Issues, Challenges,
Architecture, Platforms and Applications: A Survey. International Journal of Future
Computer and Communication, 1(4):356–360, 2012.
4. Prateek Sharma, Stephen Lee, Tian Guo, David Irwin, and Prashant Shenoy. Managing
risk in a derivative iaas cloud. IEEE Transactions on Parallel and Distributed Systems,
5. Vibhu Saujanya Sharma, Shubhashis Sengupta, and Annervaz Karukapadath Mohame-
drasheed. Method and system for managing user state for applications deployed on
platform as a service (paas) clouds, April 25 2017. US Patent 9,635,088.
14 Muhammad KaleemUllah Khan et al.
6. Georgios L Stavrinides and Helen D Karatza. The effect of workload computational
demand variability on the performance of a saas cloud with a multi-tier sla. In Future
Internet of Things and Cloud (FiCloud), 2017 IEEE 5th International Conference
on, pages 10–17. IEEE, 2017.
7. Ateeq Khan, Matthias Pohl, Sascha Bosse, Stefan Willi Hart, and Klaus Turowski. A
holistic view of the iot process from sensors to the business value. 2017.
8. Songqing Chen, Tao Zhang, and Weisong Shi. Fog Computing. IEEE Internet Com-
puting, 21(2):4–6, 2017.
9. Luis M Vaquero and Luis Rodero-Merino. Finding your way in the fog: Towards a com-
prehensive definition of fog computing. ACM SIGCOMM Computer Communication
Review, 44(5):27–32, 2014.
10. Anila Yasmeen and Nadeem Javaid. Exploiting Load Balancing Algorithms for Re-
source Allocation in Cloud and Fog Based Infrastructures.
11. Sakeena Javaid, Nadeem Javaid, Sheraz Aslam, Kamran Munir, and Masoom Alam.
A Cloud to Fog to Consumer based Framework for Intelligent Resource Allocation in
Smart Buildings. pages 1–10.
12. Saman Zahoor and Nadeem Javaid. A Cloud-fog based Smart Grid Model for Effective
Information Management.
13. Itrat Fatima and Nadeem Javaid. Integration of Cloud and Fog based Environment
for Effective Resource Distribution in Smart Buildings.
14. Jinsung Byun, Youngil Kim, Zion Hwang, and Sehyun Park. An intelligent cloud-
based energy management system using machine to machine communications in future
energy environments. Digest of Technical Papers - IEEE International Conference on
Consumer Electronics, pages 664–665, 2012.
15. Shengjun Xue, Yiyun Zhang, Xiaolong Xu, Guowen Xing, Haolong Xiang, and Sai
Ji. QET : a QoS-based energy-aware task scheduling method in cloud environment.
Cluster Computing, 20(4):3199–3212, 2017.
16. Mohammad Hossein Yaghmaee Moghaddam, Alberto Leon-Garcia, and Morteza
Moghaddassian. On the Performance of Distributed and Cloud-Based Demand Re-
sponse in Smart Grid. IEEE Transactions on Smart Grid, PP(99), 2017.
17. Ruilong Deng, Zaiyue Yang, Fen Hou, Mo-Yuen Chow, and Jiming Chen. Dis-
tributed Real-Time Demand Response in Multiseller–Multibuyer Smart Distribution
Grid. IEEE Transactions on Power Systems, 30(5):2364–2374, sep 2015.
18. Korosh Vatanparvar, Quan Chau, and Mohammad Abdullah Al Faruque. Home energy
management as a service over networking platforms. In 2015 IEEE Power & Energy
Society Innovative Smart Grid Technologies Conference (ISGT), pages 1–5. IEEE, feb
19. Feyza Yildirim Okay and Suat Ozdemir. A fog computing based smart grid model. In
2016 International Symposium on Networks, Computers and Communications (IS-
NCC), 2016.
20. Mohammad Abdullah Al Faruque and Korosh Vatanparvar. Energy Management-as-a-
Service over Fog Computing Platform. IEEE Internet of Things Journal, 3(2):161–169,
21. Sumanta Chandra Mishra Sharma and Amiya Kumar Rath. Multi-Rumen Anti-
Grazing approach of load balancing in cloud network. International Journal of Infor-
mation Technology, 9(2):129–138, 2017.
... For example, the energy cloud can be seen as a platform with economic and technical requirements for fusing distributed renewable energy systems with intelligent technologies (such as micro grid, smart instruments, storage facilities and IoT) [97] . To better manage the energy system using cloud computing, Schaefer et al. formulated the fundamental requirements for an energy cloud and its management, and discussed the major challenges as well as opportunities as the technology evolves [98] . Using fog computing to relieve the burden of data analysis, processing, and storage, Kaleemullah proposed an efficient energy management system based on the cloud [99] . ...
Digitalization and decarbonization are projected to be two major trends in the coming decades. As the already widespread process of digitalization continues to progress, especially in energy and transportation systems, massive data will be produced, and how these data could support and promote decarbonization has become a pressing concern. This paper presents a comprehensive review of digital technologies and their potential applications in low-carbon energy and transportation systems from the perspectives of infrastructure, common mechanisms and algorithms, and system-level impacts, as well as the application of digital technologies to coupled energy and transportation systems with electric vehicles. This paper also identifies corresponding challenges and future research directions, such as in the field of blockchain, digital twin, vehicle-to-grid, low-carbon computing, and data security and privacy, especially in the context of integrated energy and transportation systems.
... These improvements should also lead to lower energy consumption [10,51,[102][103][104][105][106], and also decrease maintenance costs [52,102,103,105,106]. Other technological challenges cited by the authors that can contribute to the improvement of activities related to the cloud are the improvement of data collection and acquisition, and scheduling tasks to reduce data traffic in the cloud [16,98,107], increase storage capacity, increase the number of users connected [29,108], reduce cloud saturation bottlenecks [33,53,59,60,69,72,98], improve data processing algorithms [33,46,63], decrease server failures and loss of user data [29,63], and integrate distributed data storage systems [16,63]. ...
Full-text available
The use of emerging technologies such as cloud computing, Internet of Things, and Big Data, is increasing as tools to assist the management of data and information related to energy systems grow. This allows for greater flexibility, scalability of solutions, optimization of energy use, and management of energy devices. In this sense, the objective of this research is to present the basic elements and requirements for the energy cloud and its management and discuss the main management challenges and opportunities for the development and diffusion of the energy cloud. This study was based on a systematic review carried out to identify the elements that compose the energy cloud and what is necessary for its management, and to list the challenges and opportunities that may be explored by researchers and practitioners. The results show that the layout for the energy cloud and its management can be structured in layers and management support blocks' format. It was found that 70 basic elements make up the main layers and 36 basic elements make up the management support blocks. The findings of this article also provide insights into the technical, scientific, and management development necessary for the evolution of energy systems toward the cloud computing environment.
Energy management systems, both at the level of end-users and companies and organizations, are migrating from a centralized way to cloud-based environments, thus generating virtual environments for the users to manage their energy generation, storage, and consumption. This energy management system, also called Energy Cloud (EC), is driven by the distributed generation of renewable energies, electric vehicles, and new energy storage technologies, thus providing ample flexibility and autonomy to users. In this sense, the objective of this article is to identify, evaluate, and discuss the Critical Success Factors (CSFs) that impact the implementation and management of EC environments. For this, an approach based on a survey with experts and the Multi-attribute Utility Theory multicriteria method was used, evaluating the impact levels of CSFs for EC by grouping them according to business areas and structuring them hierarchically in a decision tree. Thus, this research shows that the implementation of EC environments depends more heavily on the energy infrastructure, data management, and the computational systems that process this data. This shows the need for investments in infrastructure renewal and also in the development of new and innovative technological solutions for energy management. Another important result was the evidence of the centralizing role that the users who make up the EC environments have, demanding easier access to the technologies that make up this energy management trend, as well as greater freedoms and decision autonomy, establishing themselves in the role of free users in a dynamic and flexible energy market.
Conference Paper
Full-text available
Traditional electric generation based on fossil fuel consumption threatens the humanity with global warming, climate change, and increased carbon emission. Renewable resources such as wind or solar power are the solution to these problems. The smart grid is the only choice to integrate green power resources into the energy distribution system, control power usage, and balance energy load. Smart grids employ smart meters which are responsible for two-way flows of electricity information to monitor and manage the electricity consumption. In a large smart grid, smart meters produce tremendous amount of data that are hard to process, analyze and store even with cloud computing. Fog computing is an environment that offers a place for collecting, computing and storing smart meter data before transmitting them to the cloud. This environment acts as a bridge in the middle of the smart grid and the cloud. It is geographically distributed and overhauls cloud computing via additional capabilities including reduced latency, increased privacy and locality for smart grids. This study overviews fog computing in smart grids by analyzing its capabilities and issues. It presents the state-of-the-art in area, defines a fog computing based smart grid and, gives a use case scenario for the proposed model.
Conference Paper
Fog computing concept is introduced to reduce the load on cloud and provide similar services as cloud. However, fog covers small area rather than cloud by storing the data temporarily and sends data to cloud for permanent storage. In this paper, an integrated fog and cloud based environment for effective energy management of buildings is proposed. So, the load on cloud and fog should be balanced. Various load balancing algorithms are used to manage the load among virtual machines (VMs). In this scenario, algorithm used for load balancing among VMs is round robin (RR). Service broker policies considered in this paper are; dynamically reconfigure with load (DR) and the proposed policy. New dynamic service proximity (DSP) service broker policy is proposed for fog selection and results of DSP policy are compared with DR policy. Therefore, a tradeoff is observed between cost and response time.
Currently, energy consumption for cloud data centers has attracted much attention from both industry and academia. Meanwhile, it is also important to satisfy the customers’ quality of service (QoS) for cloud service providers. However, it is still a challenge to achieve energy savings based on QoS during task scheduling. In this paper, a QoS-based energy-aware task scheduling method, named QET, in cloud environment is proposed to address the above challenge. Technically, an energy consumption model based on QoS is proposed for heterogeneous cloud environment. And a corresponding task scheduling method is designed to minimize the energy consumption through QoS-aware PM selection. Comprehensive experimental analysis is conducted to evaluate the efficiency and effectiveness of our proposed method.
In this globalized world with the advancement of technology the use of computation and simulation gradually increases. To fulfill the increased user demand cloud network provides its ubiquitous service in rent basis. The augmented ultimatum of cloud service increase loads on virtual machines and fallouts load imbalance in cloud system. There are many challenges associated with the cloud system, load balancing is one of them. Proper resource utilization and minimization of makespan is the basic motive of load balancing. This paper describes a multi-datacenter load adjustment technique called Multi-Rumen Anti-Grazing algorithm for assigning tasks to virtual machines of different datacenters. Our proposed mechanism is a static load balancing strategy that concerns about minimization of makespan and it gives better result than the existing ones. The simulation is carried out with different randomly generated datasets and result is compared with static Min–Min and ELBMM algorithm. In each time the proposed multi-datacenter method gives better performance and makespan as compare to the traditional intra datacenter Min–Min and ELBMM technique.
By locally solving an optimization problem and broadcasting an update message over the underlying communication infrastructure, demand response program based on the distributed optimization model encourage all users to participate in the program. However, some challenging issues present themselves, such as the existence of an ideal communication network, especially when utilizing wireless communication, and the effects of communication channel properties, like the bit error rate, on the overall performance of the demand response program. To address the issues, this paper first defines a Cloud-based Demand Response (CDR) model, which is implemented as a two-tier cloud computing platform. Then a communication model is proposed to evaluate the communication performance of both the CDR and DDR (Distributed Demand Response) models. The present study shows that when users are finely clustered, the channel bit error rate is high and the User Datagram Protocol (UDP) is leveraged to broadcast the update messages, making the optimal solution unachievable. Contradictory to UDP, the Transmission Control Protocol (TCP) will be caught up with a higher bandwidth and increase the delay in the convergence time. Finally, the current work presents a cost-effectiveness analysis which confirms that achieving higher demand response performance incurs a higher communication cost.
Fog computing could alleviate many of the Internet of Things' unique challenges. This special issue explores fog computing's opportunities and challenges to form a distributed and virtualized platform, supporting computation-intensive tasks and distributing advanced computing, storage, networking, and management services to the edge of the network. From reducing latency to enhancing security, this special issue delivers novel solutions to an exciting frontier.
Infrastructure-as-a-Service (IaaS) cloud platforms rent computing resources with different cost and availability tradeoffs. For example, users may acquire virtual machines (VMs) in the spot market that are cheap, but can be unilaterally terminated by the cloud operator. Because of this revocation risk, spot servers have been conventionally used for delay and risk tolerant batch jobs. In this paper, we develop risk mitigation policies which allow even interactive applications to run on spot servers. Our System, SpotCheck is a derivative cloud platform, and provides the illusion of an IaaS platform that offers always-available VMs on demand for a cost near that of spot servers, and supports unmodified applications. SpotCheck’s design combines virtualization-based mechanisms for fault-tolerance, and bidding and server selection policies for managing the risk and cost. We implement SpotCheck on EC2 and show that it i) provides nested VMs with 99.9989% availability, ii) achieves upto 2-5 cost savings compared to using on-demand VMs, and iii) eliminates any risk of losing VM state.
The cloud is migrating to the edge of the network, where routers themselves may become the virtualisation infrastructure, in an evolution labelled as \the fog". However, many other complementary technologies are reaching a high level of maturity. Their interplay may dramatically shift the information and communication technology landscape in the following years, bringing separate technologies into a common ground. This paper o ers a comprehensive definition \the fog", comprehending technologies as diverse as cloud, sensor networks, peer-to-peer networks, network virtualisation functions or configuration management techniques. We highlight the main challenges faced by this potentially break-through technology amalgamation.