Professional Documents
Culture Documents
Iccmc48092 2020 Iccmc-0009
Iccmc48092 2020 Iccmc-0009
Iccmc48092 2020 Iccmc-0009
2) Cloudlet
In 2009, M.Satyanarayana introduced the cloudlet concept
with his teammate and implemented it at Carnegie Mellon
University. Cloudlet mainly works to provide cloud
capabilities at the edge of a network by covering a vast area
network. It required a continuous internet connection. The
small data center is deploying as a node in the cloudlet [8].
server. The selected chunks are the small part of the code. The assignment) algorithm proposed in mobile edge computing that
partial offload, also known as fine-grained offloading. The addresses the joint problem of caching as well as offloading
several algorithms proposed to overcome the challenge of [13]. The algorithm proposed in [14] considers energy
offloading. DDLO (Distributed deep learning-based efficiency, as well as resource allocation during task offloading
Offloading) algorithm proposed to make the fast and efficient decisions using Time and Frequency division multiple access
offloading decisions using a deep learning model[12]. An (TDMA, FDMA). The linear regression is used for prediction
adaptive framework proposed that adopts different network based offloading in MEC [15].Table 2 shows the comparison
situations and makes an offloading decision accordingly [11]. of different task offloading methodologies and their limitations.
The COSTA (Cost aware service caching and task offloading
2016 [14] TDMA, Multi-user, Channel gain To allocate resources while Resource allocation based on
FDMA single edge cloud considering energy during threshold structure
offloading
2018 [12] Deep Neural Multi-device, Multi-task, Link information To generate an offloading Unable to apply for real-time
Network single access point decision offloading
2019 [11] Adaptive Mukti-device, Multi-task, Proxy, Current To support a mobile application Used for Object-oriented
framework Multi-node location of a device with offloading capabilities applications only
2019 [13] COSTA Multi-task, Multi-base station Input data of task To generate offloading decision The only local optimal solution
considering service caching provided
2019 [15] Linear Single user, Input task data To make an offloading decision Used for LAN network only
regression Multi-task, Multi-node based on prediction
2016 [19] SCA (Successive Convex Approximation) Relaxed delay AR(Augmented Reality) apps To reduce energy consumption when
allocating resources
2018 [20] JCC(Joint Communication and Computation) Delay-aware General To allocate resources on-demand
2019 [17] Deep Reinforcement Learning Routing delay, SDN (Software Defined To cut-down service time and balance
Processing delay Network) resource allocation
2019 [18] Stackelberg Game Normal Smart Home To improve resource allocation efficiency,
minimizing MEC server’s utility
D. Security and Privacy issue common issue that needs to be solved. When the user/device
Security and privacy of data is a significant challenge in the moves from one location to another it enters into the different
MEC network. The applications from user's devices are territory each time, because of that performance decrease.
transfer to MEC via a wireless medium. At that time data are Many researchers proposed different approaches to solve the
disclosed to a trespasser and converted into data stealing. It mobility issue effectively. The user-centric approach is
reduces by using different encryption techniques that take proposed in [23] to optimize delay in an ultra-dense network
more time and degrades the performance of the application by considering energy efficiency during mobility
[21]. When some tasks are migrated to the cloud from the management. The proposed algorithm in [24] predicts a user’s
edge at that time, data security is a vital issue. Certain movement in order to take an offloading decision. The
privacies one needs to keep in mind like infrastructure frameworks proposed in [25] work well with IoT devices with
security, Edge server security, Edge Network security, Edge limited mobility. The GAMEC approach used to reduce the
device security [22]. response time of service invocation and for the optimization
purpose. GAMEC is a combination of a genetic algorithm plus
E. Mobility a simulated annealing algorithm in the MEC network [26].
The mobile users are generally moving, which results in Table 4 shows methodologies used for mobility and its
connection lost with the edge devices. In MEC, mobility is a limitations.
TABLE 4: MOBILITY COMPARISON
Year Methodology Network/Scenario Objective Limitations
2017 [23] EMM (Energy-aware mobility UDN (Ultra-Dense To optimize delay User movement is limited
management) Network)
2018 [24] DMPO (Dynamic mobility aware partial SDN( Software To offload task by predicting Limited to one dimension
offloading) Defined Network) the device's mobility
2019 [25] Framework for IoT application General Network, Security and efficiency Limited mobility
IoT device
2019 [26] GAMEC (Genetic Algorithm and the General To reduce the response time Data transmission speed is predefined.
simulated annealing algorithm for service of service invocation
selection in MEC)
F. Standardization
There are no standard protocols or frameworks available helpful while selecting the simulator. These features are
for the MEC network. The researchers always assume required for the MEC environment.
scenarios whenever they are working with MEC due to a lack 1) Virtualize Resources: To build virtualized resources
of environment. So, the standardization of MEC is the biggest and its arrangement in a real-time environment is a big
issue. If this issue resolved, then all assumptions of challenge[28]. Challenges may include CPU, storage,
researchers will cleared and uniformity is maintained in memory. Somehow this challenge can overcome because
MEC[3]. cloud computing works in virtual funds, but the scenario is
different when it comes to edge computing because the
IV. SIMULATION TOOLS application that runs on the edge is separate. Suppose the
Edge computing is the latest trend in the market for application runs on the edge is a sensor device app. So we
researchers but due to a lack of standardizes protocol, the main need to allocate VM (Virtual Machine) and also need to
challenge is proposing a standard for edge computing. It is combine computation capability as well. That’s how it
possible in cloud computing to create a real time environment becomes more complex[27].
or implementation. But it is not possible in edge computing 2) Mobility: To build a movement in real-time is an
due to reasons, which is discussed in this section[27]. The expensive and time-consuming task. So, most of the
researchers mainly depend on the simulation tool when it researchers work on simulation[26]. In the cloud computing
comes to edge computing. There are many simulation tools mobility not considered generally, but in edge computing
available for edge computing, each has its pros & cons mobility is one of the essential aspects that need to be
discussed in this section. The top-rated simulation tool that considered [25].
satisfies almost all criteria for mobile edge computing is the 3) Network designing: In edge computing the different
"Edgecloudsim" tool. The Edgecloudsim explained in this
access channels like LAN, WLAN, LTE, is used for testing
section.
purposes[16]. Channels may vary based on the requirements
A. Challenges for Real-Time Implementation because several users are significant, so fixing
The following problems researchers are facing in the real- channel/transmission is a challenging task[3].
time application of MEC. The below mentioned problems are
4) Scalability: It is challenging to decide the level of because of the missing mobility module [29]. The network
scalability. If too many aspects consider, then scalability is design is provided in Table 6.
poor. If fewer aspects consider, then scalability is high [27]. It 3) IOTSim: It mainly used for IoT application processing.
generally shows how to handle increasing demand. IoT has much data so it moves towards the processing of big
5) Extensibility: Researches generally works on modified data also. It includes storage and processing into cloudsim. For
architecture according to their requirement. So, whatever big data processing, the map-reduced technique used along
model is designed should have to be altered or extended with the batch paradigm [30]. The network design for IoTSim
according to the researcher's obligation [28]. described in Table 6.
6) Cost: One of the biggest challenge is cost. To overcome 4) EdgeCloudsim: It is an architecture based on different
all the problems listed above, it surely requires a considerable modules [27] which is shown in Table 5. The modules that are
amount [22]. Generally, the cost includes resource cost, missing in IOTSim and iFogsim are added in Edgecloudsim
workforce cost, and material cost. [28]. Table 6 shows a comparison based on the challenges
listed above. Based on the comparison, we conclude that
B. Edge Computing Simulation Aspects
Edgecloudsim almost full fills all the requirements, so it is a
Edgecloudsim tool [27] proposed in 2018 to mimic the suitable simulation tool for edge computing.
environment of MEC and develop the algorithms for
communication. TABLE 5: EDGECLOUSIM MODULES
Module Description
Mobility Module The mobility module manages the location of edge
devices and clients. The Hash table contained all the
data about the mobile device's current location
(i.e.X, Y component). Whenever the device location
changed, the Hash table gets updated.
Load Generator Based on the user's application, it generates/develop
Module tasks.
Networking Whenever users upload or download data at that
Module time, the transmission delay is generated, which is
handled by the Networking module.
Edge Orchestrator It handles all the client requests by collecting data
Module from the rest of the modules and make decisions.
Core Simulation The module uses to run the different configuration
Module files provided for creating an edge scenario.
lightweight virtualization through gateways. The gateway is to produces results of patients’ health and suggest the
responsible for running services. It uses WLAN, LAN channel precautions [32].
for accessing. The purposes for inside usages are [4]
D. Self-driving vehicle
x To examine activities like air conditioning, The vehicle learns from previous data using different
temperature, access control, by joining different machine learning algorithms. But that data is stored in the
sensors.
cloud, so every time vehicles have to request to cloud for data
x To deliver fast services to customers because it's
and latency is increased. MEC approach used to provide data
within their territory.
and reduces the latency using the nearest edge device [3].
x In theatres, station, or any big campus it is used to
provide security by delivering live content in E. Connected house & working areas
station/airport. It also guides passengers through It is the scenario of inside usage where MEC Server
AR (Augmented Reality) services.
installed at the campus and user access through WLAN or
B. Outside Usage LAN channel[3]. Different sensors of smart homes are
If MEC uses at the mobile base station then it will cover a connected with a gateway, and the gateway is operated by a
wide area, and it falls under outside usage. For external usage, network service operator. MEC works as a node, and it allows
macro-cell dealer enclosed the computation capabilities and local execution for IoT applications [33].
virtualization features to RAN [4]. Because of this joint F. Video monitoring
venture, it allows applications to deliver content over the
network with high and fast service. It uses LTE for accessing. The video surveillance used for security purposes and to
The purposes of outside mobile usage are [5] find criminals using face recognition from live surveillance. It
x To give better QoS (Quality of service) to the user. needs central processing like a cloud, but it suffers in terms of
x To cut down latency. speed and latency. The issues can be resolved using MEC [1].
x Provide personalized services. G. Blockchain
x To improve the capabilities of the infrastructure.
It is a ledger technology for creating a blockchain. Each
x Because it blends with radio machinery tightly, it
node works as a computation unit that follows a decentralized
provides benefits like finding device location,
providing network traffic information symbols. manner. Therefore, MEC is suitable for such type of
x Because it blends with radio machinery tightly, it processing [22].
provides benefits like finding device location, VII. CONCLUSION
providing network traffic information symbols.
The excessive growth of resource-intensive applications
VI. APPLICATIONS certainly is a challenge to provide more reliable service to users
by retaining limited mobile device resources in mind. MEC has
A. Content caching & content delivery the potential to be the future technology that offers bandwidth,
Caching is used to keep a copy of the content for further storage, and battery power to a mobile device to perform
distribution to the client. When a user requests a first time for computationally intensive tasks. The direct wireless
content at the edge, and the content is not available. In such a communication among mobile devices and edge servers cause
the possibility of enabling applications that demand ultra-low
case, it gets content from the core network and caching keeps a
latency. This paper presents a thorough survey of
copy of that content. Thus when a user requests another time, it
contemporary work in the field of MEC that covers the MEC
directly provides from an edge but first-time core connection is concept, various edge paradigm, and architecture. The study
necessary [3]. also incorporates challenge, simulation tools, and applications
B. Augmented Reality(AR) which helps and motivates the researchers to work in the
dynamic and evolving research area of MEC.
AR is a technology that enables a user to feel a realistic
scenario using virtualization. But to allow this type of scene at
the mobile devices is quite tricky due to resource limitation & REFERENCES
energy consumption. The use of MEC technology enables
computation-intensive tasks to be offloaded to nearest MEC [1] Abbas, N., et al., Mobile edge computing: A survey. IEEE Internet
Server. It reduces delay and gives a fast response to the user of Things Journal, 2017. 5(1): p. 450-465.
[31]. [2] Kumar, T.S., EFFICIENT RESOURCE ALLOCATION AND QOS
C. Health Sector ENHANCEMENTS OF IOT WITH FOG NETWORK. Journal of
ISMAC, 2019. 1(02): p. 101-110.
In the health sector, the patient data and its processing are
[3] Safavat, S., N.N. Sapavath, and D.B. Rawat, Recent advances in
important aspects. Patient data is available locally and process
mobile edge computing and content caching. Digital
it using MEC gives a fast response. For e.g., Health application
Communications and Networks, 2019.
in a wearable device, it tracks user's activity and generates data
[4] Hu, Y.C., et al., Mobile edge computing—A key technology
like a number of footsteps during a day, water consumption
record. Based on the collected data, the MEC made processing towards 5G. ETSI white paper, 2015. 11(11): p. 1-16.
[5] Patel, M., et al., Mobile-edge computing introductory technical [22] Zhang, J., et al., Data security and privacy-preserving in edge
white paper. White paper, mobile-edge computing (MEC) industry computing paradigm: Survey and open issues. IEEE Access, 2018.
initiative, 2014: p. 1089-7801. 6: p. 18209-18237.
[6] Hakak, S., S.A. Latif, and G. Amin, A review on mobile cloud [23] Sun, Y., S. Zhou, and J. Xu, EMM: Energy-aware mobility
computing and issues in it. International Journal of Computer management for mobile edge computing in ultra dense networks.
Applications, 2013. 75(11). IEEE Journal on Selected Areas in Communications, 2017. 35(11):
[7] Bhalaji, N. "Efficient and Secure Data Utilization in Mobile Edge p. 2637-2646.
Computing By Data Replication." Journal of ISMAC 2, no. 01 [24] Yu, F., H. Chen, and J. Xu, DMPO: Dynamic mobility-aware
(2020): 1-12. partial offloading in mobile edge computing. Future Generation
[8] Dolui, K. and S.K. Datta. Comparison of edge computing Computer Systems, 2018. 89: p. 722-735.
implementations: Fog computing, cloudlet and mobile edge [25] Almajali, S., et al. A framework for efficient and secured mobility
computing. in 2017 Global Internet of Things Summit (GIoTS). of IoT devices in mobile edge computing. in 2018 Third
2017. IEEE. International Conference on Fog and Mobile Edge Computing
[9] Sabella, D., et al., Mobile-edge computing architecture: The role of (FMEC). 2018. IEEE.
MEC in the Internet of Things. IEEE Consumer Electronics [26] Wu, H., et al. Mobility-Aware Service Selection in Mobile Edge
Magazine, 2016. 5(4): p. 84-91. Computing Systems. in 2019 IEEE International Conference on
[10] Roman, R., J. Lopez, and M. Mambo, Mobile edge computing, fog Web Services (ICWS). 2019. IEEE.
et al.: A survey and analysis of security threats and challenges. [27] Sonmez, C., A. Ozgovde, and C. Ersoy, Edgecloudsim: An
Future Generation Computer Systems, 2018. 78: p. 680-698. environment for performance evaluation of edge computing
[11] CHEN, X., et al., An adaptive offloading framework for Android systems. Transactions on Emerging Telecommunications
applications in mobile edge computing. SCIENCE CHINA Technologies, 2018. 29(11): p. e3493.
Information Sciences, 2019. 62(8): p. 082102. [28] Svorobej, S., et al., Simulating fog and edge computing scenarios:
[12] Huang, L., et al., Distributed Deep Learning-based Offloading for An overview and research challenges. Future Internet, 2019. 11(3):
Mobile Edge Computing Networks. Mobile Networks and p. 55.
Applications, 2018: p. 1-8. [29] Gupta, H., et al., iFogSim: A toolkit for modeling and simulation of
[13] Tran, T.X., K. Chan, and D. Pompili. COSTA: Cost-aware Service resource management techniques in the Internet of Things, Edge
Caching and Task Offloading Assignment in Mobile-Edge and Fog computing environments. Software: Practice and
Computing. in 2019 16th Annual IEEE International Conference Experience, 2017. 47(9): p. 1275-1296.
on Sensing, Communication, and Networking (SECON). 2019. [30] Zeng, X., et al., IOTSim: A simulator for analysing IoT
IEEE. applications. Journal of Systems Architecture, 2017. 72: p. 93-107.
[14] You, C., et al., Energy-efficient resource allocation for mobile- [31] Ren, J., et al., An edge-computing based architecture for mobile
edge computation offloading. IEEE Transactions on Wireless augmented reality. IEEE Network, 2019.
Communications, 2016. 16(3): p. 1397-1411. [32] Oueida, S., et al., An edge computing based smart healthcare
[15] Kim, K., et al. Prediction Based Sub-Task Offloading in Mobile framework for resource management. Sensors, 2018. 18(12): p.
Edge Computing. in 2019 International Conference on Information 4307.
Networking (ICOIN). 2019. IEEE. [33] Vallati, C., et al., Mobile-edge computing come home connecting
[16] Nasimi, M., et al. Edge-Assisted Congestion Control Mechanism things in future smart homes using LTE device-to-device
for 5G Network Using Software-Defined Networking. in 2018 15th communications. IEEE Consumer Electronics Magazine, 2016.
International Symposium on Wireless Communication Systems 5(4): p. 77-83.
(ISWCS). 2018. IEEE.
[17] Wang, J., et al., Smart Resource Allocation for Mobile Edge
Computing: A Deep Reinforcement Learning Approach. IEEE
Transactions on Emerging Topics in Computing, 2019.
[18] Guo, S., et al., Mobile edge computing resource allocation: A joint
Stackelberg game and matching strategy. International Journal of
Distributed Sensor Networks, 2019. 15(7): p. 1550147719861556.
[19] Al-Shuwaili, A. and O. Simeone, Optimal resource allocation for
mobile edge computing-based augmented reality applications.
arXiv preprint arXiv:1611.09243, 2016.
[20] Chen, X., et al., Efficient resource allocation for on-demand
mobile-edge cloud computing. IEEE Transactions on Vehicular
Technology, 2018. 67(9): p. 8769-8780.
[21] Ahmed, E. and M.H. Rehmani, Mobile edge computing:
opportunities, solutions, and challenges. 2017, Elsevier.