(PDF) Comparison of Edge Computing Implementations - Fog Computing, Cloudlet and Mobile Edge Computing

10/7/2019 (PDF) Comparison of edge computing implementations: Fog computing, cloudlet and mobile edge computing
See all › See all › See all ›

Download citation Share
Recruit researchers
Download
Join
full-text
for free
PDF Login
82 Citations 29 References 2 Figures
uting implementations: Fog computing, cloudlet and mobile edge computing
Available) · June 2017 with 1,496 Reads

.8016213
nternet of Things Summit (GIoTS)
Dolui Soumya Kanti Datta

Leuven 17.89 · EURECOM
research
bers
cations
rojects
stabh Dolui Author content

copyright.
Edge Computing Paradigm. Decision Tree for Edge Computing implementations.
r Edge Computing Paradigm. Decision Tree for Edge Computing implementations.

https://www.researchgate.net/publication/319284011_Comparison_of_edge_computing_implementations_Fog_computing_cloudlet_and_mobile_edg… 1/14

Download citation Share Download full-text PDF
Content uploaded by Koustabh Dolui Author content

DF
Content may be subject to copyright.
mparison of Edge Computing Implement

Fog Computing, Cloudlet and Mobile E
Computing
Koustabh Dolui Soumya Kanti Datta
Fondazione Bruno Kessler, Trento, Italy EURECOM, Sophia Antipolis, Fran
Email: k.dolui@fbk.eu Email: dattas@eurecom.fr
—When it comes to storage and computation of large while provisioning resources to an applicat
ta, Cloud Computing has acted as the de-facto solution Moreover, there is also a need for mobility
past decade. However, with the massive growth in agile nature of the end devices in various ap
and mobile devices coupled with technologies like
f Things (IoT), V2X Communications, Augmented emerging technologies. In parallel, the num
R), the focus has shifted towards gaining real-time devices is estimated to reach 30-50 Bill
along with support for context-awareness and mobility. routing of massive scales of network traffi
delays induced on the Wide Area Network (WAN) and can prove to be a bottleneck degrading the
nostic provisioning of resources on the cloud, there is turn the Quality of Service (QoS) and Qua
bring the features of the cloud closer to the consumer
his led to the birth of the Edge Computing paradigm (QoE). The vast number of requests to the
s to provide context aware storage and distributed to the operation of the DCNs at a high duty
at the edge of the networks. In this paper, we discuss in emissions of harmful greenhouse gases
different implementations of Edge Computing namely effect on the environment.
uting,
SeeCloudlet
all › andSee
Mobile
all › Edge Computing
See all › in based
detail The Edge Computing (EC) attempts to
re their features. We define a set of parameters Download citation
scribed challenges. The EC Share
leverages the Download full-text PDF
ne 82
of Citations 29 References
these implementations can be 2chosen
Figuresoptimally
rticular use-case or application and present a decision cessing capacities of a large number of
e selection of the optimal implementation. nected to the Internet deployed for the p
rms—Cloud Computing; Cloudlet; Edge Computing; an intermediate layer between the end devi
uting; IoT; Mobile Edge Computing. With the presence of these ”Edge devices
load at the data centers are reduced by han
I. INTRODUCTION
requests directed to the cloud, locally, wh
ent growth in services and applications leveraging intervention from the cloud. This in turn, r
et has contributed to a steep rise in data storage in resolving the requests and allows real-t
sing requirements. They are diverse in terms of the subset of requests. Edge devices also suppo
required by different applications and thus, often the abundant availability and geo-distribute
or-made solutions. Cloud Computing provides as a The Edge layer between the end devi
lution in in this context by leveraging the advance- are implemented in different ways in ter
omputing and network technologies. The backbone which act as the intermediate edge nodes, t
ud Computing paradigm is based on the data centers protocols and networks used by the Edg
capable of handling storage and processing of large the services offered by the Edge layer. T
data. These data centers are often connected with of the edge layer can be classified into th
over optical networks to form data center networks Edge Computing (MEC), Fog Computing
ppearing as a singular resource to the end user, with Computing (CC). Fog Computing presents
y communication among the data centers. However, leveraging devices like M2M gateways an
f Things (IoT) systems have presented a new set These are called Fog Computing Nodes (F
ments the well established CC based solutions. to compute and store data from end devi
ns especially connected vehicles require near real- forwarding to the Cloud. On the other han
ssing of sensor data to take decisions and perform deployment of intermediate nodes with stor
Even though the communication inside the DCNs capabilities in the base stations of cellu
le for low-latency communication, the latency of offering Cloud Computing capabilities insi
ation between the end devices and the DCNs prove Network (RAN). The Cloudlets are based on
tleneck. This owes to the lack of location awareness with capacities similar to a data center bu
873-0/17/$31.00 2017
c IEEE present in logical vicinity to the consume
emerging technologies and applications ran

tational offloading to edge content deliver
Surveys on Mobile Edge Computing [7]
research efforts including MEC architectura
models. The discussions include FemtoCl
and CloudAware while also highlighting o
this area.
In comparison to MEC and FC, Cloudl
newer paradigm. Satyanarayanan et. al pro
of virtual machine based Cloudlets as one
[9] on Cloudlet Computing. Li et. al prese
1. N-tier Architecture for Edge Computing Paradigm.
correlation of user mobility and how it a
in Cloudlets [10]. Authors of [11] propose
d devices to offload Computing to the Cloudlet architecture which attempts to leverage
ith resource provisioning similar to that of a data abilities of devices in a LAN network by ha
See all › See all › See all › at a component level.

he context of EC implementations,
82 Citations 29 References our contributions
2 Figures ExistingDownload citation
work concerning Share implem Download full-text PDF
multiple
per are manifold. We critically examine the differ- Computing include that of Jararweh et. al [
mentations and present a taxonomy to define and a study on the integration of MEC and C
he features of these implementations. We propose a framework and performance evaluation
ork for selecting a specific implementation given a architecture. However, in the existing litera
uirements from a use case. The framework, based of studies which directly compare these thre
ion tree, is devised to match the requirements of a under the umbrella of Edge Computing. W
application or use-case with the features offered by through our contribution in this paper.
ementations to make an optimal selection for the
case or application. B. Edge Computing based applications
II. STATE- OF- THE-ART The distributed computing philosophy h

various emerging technologies and IoT ap
d growth of emerging applications like autonomous advantage of different features offered by E
smart cities, e-health monitoring among others, has have developed an emergency alert servic
e case for Edge Computing as one of the major phones by computation offloading and pre
or these applications. We study the related works received from an emergency to Fog nodes
ts. In the first part we discuss the related work on proposes a fall-detection algorithm and an
ctural aspects while the second part explains several the same using the computation capabilitie
ns and use-case scenarios developed leveraging the Fog nodes. Authors of [15] have used the s
gm. layer to implement a parking service whi
aggregate data collected from various Fog
cture and Definition optimal parking spot, thus leveraging the d
et. al presented one of the first works on Fog Fog nodes. Similarly authors of [16] lever
g [1] assessing the suitability of Fog Computing for nature of Fog nodes to improve location awa
This paper presents the requirements of emerging and apply their model to a lane changing ass
ns in terms of location awareness, real time inter- Fog Computing architecture devised with
d need for geo-distributed end-points and how Fog approach is presented in [17] with a use
g addresses these issues. The authors provide further connected vehicles.
o the suitability of FC for IoT applications with a On the other hand, for MEC based use c
ses including a smart traffic light system and a smart is exposed to more information in terms of
in the following paper [2]. On the other hand, the for the consumer due to the association of th
[3], [4] provide an overview of the Fog architecture the Radio Access Network (RAN). Thus, p
mprehensive definition of the Fog nodes and their the MEC nodes co-located with the base s
The authors of [5] provide an application oriented major improvement for performance of we
tion of FC along with the security issues involved of [18], [19] address this issue, pointing o
se-case scenarios. of using proactive caching to avoid acces
. al presented the taxonomy and an architectural networks while also improving Quality of
of Mobile Edge Computing [6] along with the for the end-user. Resource virtualization
t would serve with respect to the requirements of pect of Mobile Edge Computing as addre
MEC nodes to run applications in containers offering the end users and allocates resources in
s-a-Service (PaaS). requirements of the requests.
rayanan in his work [22] proposes use cases for 2) Mobile Edge Computing: MEC can
Computing in Vehicle to Vehicle (V2V) communica- implementation of Edge Computing to br
ng different
See all › architectures
See all ›for different
Seeuse-cases.
all › For and storage capacities to the edge of the n
n the case of an on-vehicle video game application, Download
Radio Access citation
Network Sharelatency an Download full-text PDF
to reduce
et is more suitably co-located with the device, also awareness. The MEC nodes or servers are
g on-device processing from the sensor data while, with the Radio Network Controller or a m
oring of road and traffic conditions an ad-hoc set The servers run multiple instances of ME
ts deployed in an area appears to be more suitable. the capabilities to perform computation an
also discusses the suitability of Cloudlet Computing tualized interface. The MEC hosts are overl
text of handling privacy and security measures as a Edge Orchestrator which handles informati
p between end-devices and the cloud. The authors of offered by each host, the resources availab
ss the role of Cloudlets in improving crowd-sourcing topology while also managing the Mobile
ns in terms of scaling and pre-processing of massive The MEC servers offer real time informati
data generated from crowd-sourcing applications. itself including the load and capacity of
ve studies help us extract the important trends in the also offering information on the end device
nts of the emerging EC applications and we discuss servers including their location and networ
ds in the following sections. 3) Cloudlet Computing: A Cloudlet can
trusted cluster of computers, well connect
COMPARISON AMONG EDGE COMPUTING with resources available to use for nearby m
IMPLEMENTATIONS
A Cloudlet can be treated as ”data center
ection, we address the gap in existing literature by a virtual machine capable of provisioning
a detailed analysis of the implementations in the devices and users in real time over a WL
subsections. In the first part of the section we look services are Cloudlets are provided over a o
itecture of the implementations individually while in high bandwidth, thus offering low latency fo
part we study the work flow through the different architecture proposed in [11] for Cloudlets
the architecture while handling requests from the layers, the component layer,the node laye
in the final part, we compare the Edge Computing layer. The component layer offers a set of s
ations and present our study. ing interfaces to the higher layers overloo
tion Environment. One or multiple Executi
lementation architectures running on top of an OS form a Node, m
mpare the three EC implementations in terms of Agent. A group of co-located nodes form
ecture they follow, the function and location of managed by a Cloudlet Agent. Satyanaran
s serving as the intermediate layer between the end an architecture for cognitive assistance ap
d the cloud, their offered services as well as their which involves a primary virtual machin
lications in the following subsection. The N-tier cognitive functionalities offered by other v
e involving the cloud platforms, the end devices and the Cloudlet to serve a request. The data
nt implementations of Edge Computing is portrayed VMs is gathered on a user guidance VM wh
to the end user.
Computing: The Fog Computing implementation is
lized Computing infrastructure based on Fog Com- B. EC implementation request handling
des (FCNs) placed at any point of the architecture In this section, we study how requests
he end devices and the cloud. The FCNs are hetero- different Edge Computing implementation
n nature and thus can be based on different kinds consider a use-case which requires offload
ts including but not limited to routers, switches, tasks to the Edge, for example, an IoT base
nts, IoT gateways as well as set-top boxes. The where processing of data from car sensors
eity of FCNs paves the way for supporting devices at processing of the data is offloaded to t
rotocol layers as well as support for non-IP based Computing implementations.
hnologies to communicate between the FCN and 1) Fog Computing: In the Fog Computin
vice. The heterogeneity of the nodes is hidden from Fog Orchestrator presides over the underlyi
vices by exposing a uniform Fog abstraction layer municating with the nodes through the fun
oses a set of functions to perform resource allocation the Fog abstraction layer. The requests from
oring, security and device management along with at the Fog Orchestrator with a set of require
d compute services. These functions are utilized by policies. The required policy may include pa
e Orchestration Layer which receives requests from minimal Fog node configuration, load balan

TABLE I
C OMPARISON OF E DGE C OMPUTING IMPLEMENTATIONS
Fog Computing Mobile-Edge Computing Cloudlet Computing

es Routers, Switches, Access Points, Gateways Servers running in base stations Data Center in a box
on Varying between End Devices and Cloud Radio Network Controller/Macro Base Station Local/Outdoor installation
rchitecture Fog Abstraction Layer based Mobile Orchestrator based Cloudlet Agent based
areness Medium High Low
One or Multiple Hops One Hop One Hop
chanisms Bluetooth, Wi-Fi, Mobile Networks Mobile Networks Wi-Fi
Communication Supported Partial Partial
Orchestrator matches the policies with the services In terms of proximity to the edge, in case
y each of the nodes and returns an ordered list of the FCN may not be the first hop access
erms of suitability against the requested policy. The device due the leveraging of legacy devi
ch are most suitable are chosen based on availability. example, the first router connected to the e
le Edge Computing: The MEC servers co-located be resourceful enough to run an FCN frame
ase stations, receive the requests at the Mobile Edge closest FCN may be present multiple hop
or from the end user. The orchestrator maintains a to the support for inter-node communicatio
applications that are running on the underlying ME Computing as well. However, for Cloudlet
receives updates on the available resources from the Computing, the devices connect directly to
rm Manager. If an application is already running, the Fi and mobile network at the base station
redirected to the application while if an application Computing leverages gateways as devices fo
running state but is supported by the platform, the support for non-IP based protocols like BLE
n is instantiated if resources are available and the allows Fog Computing to connect to a w
accepted. Otherwise, the request is passed on to be devices and also offer protocol translation.
the cloud passing through the core of the network. If we analyze the request handling m
dlet Computing: In case of Cloudlet Computing, services offered by each of these implement
et Agent overlooks the Cloudlets and the underlying a similar hierarchical approach, where the
Cloudlet Agent communicates with the underlying entity overlooking the underlying nodes wh
ts through the Node Agent and the Execution Envi- with them to gather information on the r
Policy violations in the components are passed on to availability. However, the diversity and he
et Agent from the components hierarchically. This the Fog devices invoke the need for an abst
Cloudlet Agent to make an optimized choice for in the other implementations it is not necess
ing node when a request is received such that more devices are used as nodes. MEC, on the oth
ueries are handled by nodes with higher processing advantage of having fine grained informatio
The Cloudlet Agent can also provision and allocate location and network load for improved con
urces by instantiating new VMs if necessary to Cloudlets, the data stored and processed on
received requests. a soft state, i.e. the data is already backe
and is updated once the processing is finis
rison advantage offered by Cloudlets is that the
a fresh Virtual Machine on the Cloudlet ev
n the features mentioned in the previous subsection, Cloudlet performs a pre-use customized res
t a comparative study of the implementations in and a post-use cleanup by backing up the p
ince the Fog Computing implementation proposes the cloud
ce of FCNs anywhere between the end devices and
DCNs, Fog Computing offers more flexibility in IV. USE C ASE BASED DECISION MA
e of devices for using them as FCNs. However, The Edge Computing paradigm offers s
Ns leverage legacy devices by adding storage and variety of end devices, applications and u
to them, the computation and storage capacities These use cases and applications have
See allthan
y lesser › thatSee
of allCloudlets
› SeeMobile
and all › Edge requirements and trade-offs which determin
n the82other hand, due to the requirement of dedicated Download
Edge Computing citation Share is suitabl
implementations Download full-text PDF
Citations 29 References 2 Figures
r Mobile Edge Computing and Cloudlet Computing In this section, we define a taxonomy to ev
ation of these implementations are slower than that case and build a decision tree based on the
mputing. However, these devices can be reused as available implementations to make an optim
servers as well as Cloudlets. implementations of Edge Computing.
devices but lack of support for constrained

3) Context awareness: Context awarenes
eter for applications and use cases where
the network and surrounding devices is ex
nodes. In this context, the MEC servers pr
geous since they are placed in the RNCs,
information about the device location, load
also the capacity of the network. However,
Fog Computing are usually devices with a
the network, like routers or switches, the co
lesser than that of MEC. However, the abili
among the nodes themselves offers mitig
some extent. On the other hand, for Cloud
Cloudlets are designed to be standalone de
the cloud, thus offering minimal context aw
2. Decision Tree for Edge Computing implementations.
4) Power consumption: The power consu
ters for implementation selection devices is a major contributing factor if th
resource constrained. The authors of [25] h
section we present a set of parameters based on energy consumption with the use of LTE a
er can compare the features and the performance of is much higher than the energy consumptio
e implementations of the Edge Computing paradigm the power consumption while accessing M
articular application or use case. We study each is higher than that of Cloudlets. On the
arameters comprehensively and illustrate how these Computing allows access to its nodes throug
affect the Edge Computing paradigm. which consume lower energy like 802.15.4
imity: Proximity between the Edge Computing 5) Computation time: Computation time
the end devices can be defined in two ways. The the time required by the Edge layer to
tion is that of logical proximity where the proximity assigned and responding to the end user
by the number of hops between the end device results. In terms of computation time, M
dge layer. The higher the number of hops, higher Computing prove to be advantageous due
es of encountering queues along the path and thus nature of resources along with dynamic reso
f increased latency. Thus, the importance of logical On the other hand, since the Fog devices
lies in the reduction of possibilities of encountering devices, the processing and storage capaciti
the back haul network. On the other hand physical are lower, leading to a higher computation
is defined by the actual distance of the end device
igher layer of computation. For example, in case B. Decision Management
Computing, physical proximity plays a major role In this subsection, we present a multi-e
device is located on one continent while the DC based on the parameters defined in the pre
on another continent, delay becomes an important goal of this tree is to consider the desira
ysical proximity might also affect the performance a particular use case or application and
Computing
See all › when aSee
single
all › RNC handles devices
See all › over decide on a particular implementation of
a, physical proximity also comes into play for MEC Download
The parameters andcitation Share charact Download full-text PDF
their possible
me and delay sensitive applications. are presented as entry points to the tree. The
ss mediums: The connection to the Edge layer use case needs to be presented in the form o
be established by end-devices using different medi- and a choice is to be made as an entry p
Wi-Fi, Bluetooth, Zigbee, Mobile Radio Networks Following the requirements of the use ca
hers. Access mechanisms are important in more parameters specified, the path has to be fo
one as it determines the bandwidth available to the optimal choice. We illustrate the use of the
es, the range of connectivity and also support for the following use case.
ypes of devices. Fog Computing offers support for If we consider a use-case for a V2X com
ay of mechanisms including Bluetooth and Zigbee cation where the Vehicle offers video stre
ows constrained devices with not enough memory streaming video from a list of videos cache
HTTP stack to connect to the FCNs and offload node. For this use case, we can classify th
on and storage. Moreover, these devices can also be logical proximity to avoid the need to
r resources from the cloud through the FCNs in networks, need for IP based access for hi
ayer. However, Cloudlets only support Wi-Fi as an need for physical proximity and high contex
chanism which offers high bandwidth to the end for information on the network. Since, thi
Load more
References (29)
BIoT application is detailed in [70], where it is presented the architecture of a blockchain-based platform for
ion medicine. It is also worth mentioning the work described in [71], which presents a generic smart
makes use of IoT devices, cloud and fog computing [72] , a blockchain, Tor [73] and message brokers. IoT
lso be enhanced by blockchain technology. ...
SBCs are the different versions of Raspberry Pi [84] or Beagle Bone [85]. Fog computing is actually
edge computing [72] , which has recently been presented as a valid architecture for supporting blockchain
G IoT applications [86]. As it can be observed in Figure 5, in the Edge Computing Layer, besides fog
udlet, which in practice consists in one or more high-end computers that act like a reduced version of a cloud.
Block chain for the Internet of Things
· Surendar Singh
ues occur because SecLaaS allows investigators to access and read user logs. Log-chain technology
quentially accumulating log entries, which can cause performance degradation during large log-entry creation
vide multiple log sources, such as a mobile edge cloud [12][13] [14] [15][16] and fog computing [13,17].
e [8] has a low probability of false positives, which are inherent in search failure. ...
g Assuring-Secrecy
See all › Scheme
See all › for DigitalSee
Forensics
all ›
le
am Huh
he network infrastructure components. The nodes close to the users have a smaller computational capacity,
utational capacity that MEC servers, but the capacity increases towards the cloud in a hierarchy [13] , [12],
ed Nanoservices for Local IoT Edge Networks

-text available
Harjula · Tanesh Kumar · Mika Ylianttila
e data storage in Cloud is studied by [4], where a review of acquisition, management, processing and mining
resented. The authors of [8] discuss a comparison of Edge computing implementations, Fog computing,
dge computing. The focus is also on a comparative analysis of the three implementations together with the
that affect nodes communication (e.g., physical proximity, access mediums, context awareness, power
tion time). ...
mputation Model for Supporting Analytics at the Edge
Panagiota Papadopoulou · Christos Anagnostopoulos · Stathes Hadjiefthymiades
een a lot of discussions about Fog computing or Fog networks, which may lead to some confusion regarding
he term [4] . A brief history shows that the term was first used by Cisco Systems, Inc. in describing the
Cloud towards the edge of the network [5], of which has seen wide adoption of the concept. ...
n Fog Computing: Review
eco Ventura
s includes smart city devices, smart city alarm devices, and privacy in smart city. The privacy in smart city
acy objectives, context privacy and content privacy in the process of interactions or interoperations, which can
gy by specifying two sub-classes, Data TABLE 1: Comparison between fog computing, multi-access edge
t computing [25] . privacy and Privacy Rules. ...
g Architecture for Privacy-Preserving in IoT-based Smart City

le
uoc-Viet Pham · Mamoun Alazab · Gautam Srivastava

r of the two methods took joint consideration of energy consumption and transmission delay. In the realm of
25] , Dolui et al. discussed and compared the multiple features of the novel paradigms, including fog, cloudlet
abled Computation Offloading for IoT in Mobile Edge Computing
g · Honghao Gao · Wanchun Dou
d Smart Cities: A Comprehensive Survey

able
qoob · Nguyen H. Tran · Choong Seon Hong
omputing System, Communication Technology and Protocols in IoT system
ent and delay-aware offloading scheme based on device to device collaboration in mobile edge
N SYST
ohammed Mansoor · Asmiza Abdul Sani
Show more
Recommendations
Discover more publications, questions and projects in Fog Computing
https://www.researchgate.net/publication/319284011_Comparison_of_edge_computing_implementations_Fog_computing_cloudlet_and_mobile_ed… 10/14

82 Project
AGILE: an Adaptive & Modular Gateway for the IoT
Csaba Kiraly · Koustabh Dolui · Roman Sosa
AGILE (Adaptive Gateways for dIverse muLtiple Environments) builds a modular hardware and software gateway for the Internet of
Things with support for protocol interoperability, device and data man ... [more]
View project
Project
AGILE: Adaptive and Modular Gateway for IoT
Koustabh Dolui · Csaba Kiraly · Alexander Felfernig · [...] · Fabio Antonelli
View project
Project
DataTweet
Soumya Kanti Datta
This project investigates the data-centric approaches in IoT.
View project

82 Project
High precision positioning for cooperative ITS applications (HIGHTS)

Jérôme Härri · Dirk T. M. Slock · Irfan Khan · [...] · Benoit Denis
Cooperative intelligent transport system (C-ITS) applications rely on knowledge of the geographical positions of vehicles.
Unfortunately, satellite-based positioning systems (e.g., GPS and Galileo) ... [more]
View project
Article
The Random Trip Model, Part II: Stationary Regime and Perfect Simulation1
Jean-Yves Le Boudec · Milan Vojnovic
Abstract— In part I, we defined the random trip class of mobility models, and showed that a necessary and sufficient condition for
existence (and convergence to) a stationary regime is that the mean,trip duration is finite. In this part, using Palm calculus, we
establish properties of the stationary regime for the examples considered in the first part, when this condition holds. Then we,provide
... [Show full abstract]
Read more
Conference Paper Full-text available
Implementation of a Remote Mobile Thin-Client Lab Solution Incorporating Virtual Machine-based Techn...
October 2014
Charles J Lesko
This effort describes the background, design, implementation, and early impacts of fielding a remote mobile thin-client lab solution to
support online course delivery to remote students. Specifically, the effort focuses on the incorporation of virtual machine-based
technologies to provide a flexible operating environment for remote online students to interact in while minimizing the need for a ...
[Show full abstract]
View full-text

82 Citations
Article 29 References 2 Figures
Proposing a novel adaptive rate control algorithm for application to forward link of high rate packe...
December 2006
X. Wen · H. Zhang · J. Xu
IS-856 standards stipulates minimum performance requirements that HRPD mobile terminal must satisfy. We now propose a novel
adaptive control algorithm that appears to be a good way of meeting these minimum requirements. We explain our novel algorithm in
detail. Adaptive rate control algorithm is presented including the fixed margin data rate selection method, optimal rate control
algorithm, and ... [Show full abstract]
Read more
Article
Softnet supports for extended mobile IP protocol

December 2000
Y. Lu · D. Qian · B. Xu · L. Wang
In order to solve the problems such as low performance, security limitation, and deployment-difficulty of new protocols in
implementing the basic mobile IP protocol, an extended mobile IP protocol is implemented on an active network prototype-Softnet.
The Softnet has the feature of dynamic loading of the mobile agents that dynamically extends functions of the active node. Routing
tables for the ... [Show full abstract]
Read more
Discover more
About Support Business solutions
News Help center Recruiting

Company FAQ Advertising
Careers
© ResearchGate 2019. All rights reserved. Imprint · Terms · Privacy

(PDF) Comparison of Edge Computing Implementations - Fog Computing, Cloudlet and Mobile Edge Computing

Uploaded by

Copyright:

Available Formats

(PDF) Comparison of Edge Computing Implementations - Fog Computing, Cloudlet and Mobile Edge Computing

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(PDF) Comparison of Edge Computing Implementations - Fog Computing, Cloudlet and Mobile Edge Computing

Uploaded by

Copyright:

Available Formats

10/7/2019 (PDF) Comparison of edge computing implementations: Fog computing, cloudlet and mobile edge computing

See all › See all › See all ›

uting implementations: Fog computing, cloudlet and mobile edge computing

Available) · June 2017 with 1,496 Reads

Dolui Soumya Kanti Datta

stabh Dolui Author content

Edge Computing Paradigm. Decision Tree for Edge Computing implementations.

r Edge Computing Paradigm. Decision Tree for Edge Computing implementations.

See all › See all › See all ›

Content uploaded by Koustabh Dolui Author content

mparison of Edge Computing Implement

emerging technologies and applications ran

See all › See all › See all › at a component level.

II. STATE- OF- THE-ART The distributed computing philosophy h

See all › See all › See all ›

Fog Computing Mobile-Edge Computing Cloudlet Computing

devices but lack of support for constrained

Block chain for the Internet of Things

ed Nanoservices for Local IoT Edge Networks

Harjula · Tanesh Kumar · Mika Ylianttila

mputation Model for Supporting Analytics at the Edge

Panagiota Papadopoulou · Christos Anagnostopoulos · Stathes Hadjiefthymiades

n Fog Computing: Review

g Architecture for Privacy-Preserving in IoT-based Smart City

uoc-Viet Pham · Mamoun Alazab · Gautam Srivastava

See all › See all › See all ›

abled Computation Offloading for IoT in Mobile Edge Computing

g · Honghao Gao · Wanchun Dou

d Smart Cities: A Comprehensive Survey

qoob · Nguyen H. Tran · Choong Seon Hong

omputing System, Communication Technology and Protocols in IoT system

See all › See all › See all ›

This project investigates the data-centric approaches in IoT.

See all › See all › See all ›

High precision positioning for cooperative ITS applications (HIGHTS)

Conference Paper Full-text available

See all › See all › See all ›

Softnet supports for extended mobile IP protocol

About Support Business solutions

News Help center Recruiting

© ResearchGate 2019. All rights reserved. Imprint · Terms · Privacy

You might also like