Toward a real time framework in cloudlet-based architecture

TSINGHUA SCIENCE AND TECHNOLOGY
ISSNll1007-0214ll07/10llpp80-88
Volume 21, Number 1, February 2016
Toward a Real-Time Framework in Cloudlet-Based Architecture
O. Kotevska , A. Lbath, and S. Bouzefrane
Abstract: In this study, we present a framework based on a prediction model that facilitates user access to a
number of services in a smart living environment. Users must be able to access all available services continuously
equipped with mobile devices or smart objects without being impacted by technical constraints such as performance
or memory issues, regardless of their physical location and mobility. To achieve this goal, we propose the use of
cloudlet-based architecture that serves as distributed cloud resources with specific ranges of influence and a real-
time processing framework that tracks events and preferences of the end consumers, predicts their requirements,
and recommends services to optimize resource utilization and service response time.
Key words: smart city; cloudlet; prediction; recommendations; framework; real-time
1 Introduction
The concept of smart environments evolves from the
definition of ubiquitous computing, which, according to
Mark Weiser, promotes the ideas of “a physical world
that is richly and invisibly interwoven with sensors,
actuators, displays, and computational elements,
embedded seamlessly in the everyday objects of our
lives, and connected through a continuous network”[1]
.
This is a general idea from which the smart city concept
has evolved. The vision of the future smart city is
of a city with a pervasive overlay of Information and
Communication Technologies (ICT) connecting things,
organizations, and people (e.g., sensors that connect
cars to transportation management centers that analyze
day-to-day traffic flow data and provide solutions to
what-if scenarios in case of events or accidents[2]
); this
is known as the Internet of Things (IoT).
The IoT connects various objects, such as people,
data, smart objects, and processes in networks
O. Kotevska and A. Lbath are with the CNRS, LIG,
University Grenoble Alpes, Grenoble F-38000, France. E-
mail: fkotevksa.olivera, ahmed.lbathg@imag.fr.
S. Bouzefrane is with the CEDRIC Lab, Conservatoire
National des Arts et Metiers CNAM, Paris 75003, France. E-
mail: samia.bouzefrane@cnam.fr.
To whom correspondence should be addressed.
Manuscript received: 2015-09-15; accepted: 2015-10-13.
of billions or even trillions of connections. These
connections create vast amounts of data, some of which
we have never been accessible before[3]
. Because these
objects have embedded intelligence, the device itself
can filter out relevant information and even apply
analytics to enable real-time decision making. Thanks
to these enabling capabilities, the IoT will completely
change the type of services that are offered and also
how they are offered[2]
. Services will be available
to interact with these objects over the Internet, and
query their state and any information associated with
them. These new services provided over the Internet
are collectively known as the Internet of Services
(IoS)[3]
. The main goal of IoS is to present everything
on the Internet as a service, including software
applications, platforms for developing and delivering
these applications, and underlying infrastructures. In
this scenario, cloud technology plays an important
role in enabling IoS deployment, because it comprises
different provisioning models for on-demand access
to applications, platforms on which developers can
build services and applications, and elastic computing
infrastructures.
The main goal of cloud computing is to make
a better use of distributed resources, combine them
to achieve higher throughput, and solve large-scale
computation problems. This means that a mobile device
can execute a resource intensive application on a distant
www.redpel.com +917620593389

O. Kotevska et al.: Toward a Real-Time Framework in Cloudlet-Based Architecture 81
high-performance server or computational cluster and
support thin client user interactions with the application
over the Internet. The mobile device functions as a thin
client, with all complex computations being conducted
in a nearby cloudlet, which is a trusted, resource-rich
computer or cluster of computers that is well-connected
to the Internet and available for use by nearby mobile
devices[4]
. Figure 1 shows the general overview of
cloudlet-based architecture.
At present, the use, prioritization, and selection of
resources is a major concern in the design of solutions
based on emerging technologies. The focus of this study
is to create a framework based on cloudlet architecture
for a dynamic and real-time predictive model for mobile
user potential requirements.
The rest of the paper is organized as follows.
Section 2 describes related works. Section 3 presents
our proposed framework for smart cities based on
cloudlet architecture, while Section 4 illustrates the
proposed solution under different scenarios. Section 5 is
dedicated to a detailed case study and Section 6 contains
the conclusions and plans for future work.
2 Related Work
Recently, the concept of Smart Cities has become
increasingly attractive for research, academic, and
industry communities. There are factors that determine
the different approaches to solve some of the challenges
posed by Smart Cities, e.g., how to handle and analyze
Big Data from social and physical sensors, how to make
life in Smart Cities easier for users, or how users can
select services that will be useful for them.
Large companies in the ICT sector, such as Fujitsu,
Cisco, and IBM, are strongly involved in and have been
sharing their vision for the Smart City[2,5,6]
. Following
are some related works in this domain.
Similar to participatory sensing, Mobile Crowd
Sensing (MCS)[7]
uses the power of citizens for
Fig. 1 General architecture.
large-scale sensing. It goes beyond participatory
sensing by having implicit and explicit participations
and collects crowdsourced data from both mobile
sensing and mobile social network services. The
authors characterized the key features of MCS,
including crowd-powered data collection, cross-space
data mining, and low-quality data analysis. To facilitate
the development of MCS apps, they proposed a
framework and discussed the efforts of balancing
human and machine intelligence in MCS systems.
A semantic service composition architecture,
method, and algorithm has been presented in the
context of Web of Objects (WoO)[8]
. Ontological
relations among objects, services, and rules are
described to create new services dynamically. Services
are described in the knowledge-base of the ontology
layer, and rules are generated to perform reasoning
tasks for automatic service discovery and processing.
The system is capable of natural language processing,
which allows users to determine their requirements for
inspection and repair. The authors also implemented an
object assembler and a service composer for application
to match services.
Anteater[9]
is a service-oriented architecture for
data mining that relies on Web services to achieve
extensibility and inter-operability. It offers simple
abstractions for users and supports computationally
intensive processing on large amounts of data through
massive parallelism.
All these approaches provide similar solutions, i.e.,
they process the received data ofﬂine. In this work,
we expand this approach by adding a framework
for a real-time prediction model. We consider user
content, proﬁles, and trends from social sensors (social
media) to create timely predictions and context-aware
recommendations using machine learning and data
mining techniques.
3 Proposed Approach
In order to overcome the limitations of existing
recommendations, we propose a framework for a Smart
Environment aimed at allowing unlimited connectivity
for end consumers to their context aware services
through any accessible device as well as allowing real-
time prediction of user requirements.
3.1 Actors in the system
The actors in the proposed system are cloud, cloudlet,

82 Tsinghua Science and Technology, February 2016, 21(1): 80-88
end consumer, and user, defined as follows:
The cloud acts as a federated cloud for a defined set
of cloudlets. It offers better performance than cloudlets
and can be rapidly provisioned and integrated with the
cloudlets with minimal management effort.
A cloudlet is the middle tier in our three-tier
architecture model. It uses less memory but offers
less performance than the cloud, and its function is
to act as a local cloud for some territory (this means
each cloudlet has an area of influence). This approach
decreases requests and accesses to the federated cloud.
End consumers are devices receiving output results
from the processing units. The definition includes smart
objects and Cyber-Physical Systems (CPS). Smart
objects are any type of user interface with the capability
of receiving commands via voice or touch pad and
returning the response. Examples of Smart Objects
include mobile devices, Google Glass, smart watches,
and smart cars. Cyber-physical systems are devices
that have the capability to receive commands such as
billboards and alarm systems.
User is a person who uses smart objects for their
activities.
3.2 System architecture
In our proposed framework, we consider a collection
of interconnected smart objects within a smart city.
Based on the scope introduced in Section 1, we
use a reference system architecture[10]
that connects
smart objects through wireless technology using cloud
principles as in Fig. 2.
To meet the requirement of real-time interactive
response (i.e., low latency) for the services that require
it, instead of relying on a distant cloud (as measured by
the latency), we might be able to address the limitations
of a mobile device’s resource using a nearby, resource-
rich cloudlet. If no nearby cloudlet is available, the
Fig. 2 System architecture.
mobile device can gracefully degrade to a fallback
mode that involves a distant cloud or, in the worst
case, rely solely on its own resources providing limited
functionality (with full functionality and performance
returning when the device subsequently discovers an
available nearby cloudlet).
This system architecture is divided into four levels of
communication between the actors in our system.
The first level is the communication between
user(s) and smart object(s) by employing a user
interface for interaction.
The second level comprises the communication
between all types of end consumers (smart objects,
CPS, and physical sensors) with various cloudlets.
This level uses LAN and Wi-Fi technologies for
communication.
The third level of communication happens
between cloudlets and the Federated cloud and
between cloudlets. As we see, only cloudlets can
communicate directly with the cloud. This level
uses Internet protocols for communication.
The fourth level of communication takes place
between “social sensors” and the federated cloud,
using Internet protocols. The federated cloud
contains the processing unit, data management
systems and mechanisms, analytic engines, and
service deployment and composition.
Cloudlets, originally motivated by narrow end-to-
end latency, can play a much broader foundation
role in mobile computing. A cloudlet can be viewed
as a surrogate or proxy of the real cloud, located
as the middle tier of a three-tier hierarchy: mobile
device, cloudlet, and cloud. To meet the need for
real-time interactive response with low latency, instead
of relying on a distant cloud, data is transferred
to a nearby resource-rich cloudlet[4]
. Cloudlets are
interconnected in a similar way to the topologies used
in wireless sensor networks, namely mesh, star, or
hub. Each cloudlet contains a service repository
and outgoing logic, an end consumer register, Virtual
Objects (VO) representation, processing unit(s), and a
data management mechanism.
In turn, smart objects contain the user interface for
interaction with users and available services.
Cloudlets, smart objects, and CPS devices are
considered to be trust worthy, and the network
connection between them is of sufficient bandwidth
and low latency. All subjects that are included
and participate use the common platform to enable

interoperability among the service distributors.
Finally, an important aspect of this architecture is that
from the perspective of the end consumer everything is
considered as a cloud service. This means that users
do not need to be aware of or notified whether they are
connected to different cloudlets, or where the services
that they are using are executed (in their mobile device,
in a cloudlet or in the cloud). They just need to have
the opportunity to access and use their services without
performance or memory constrains or limitations on
time and place, and to have a quick response time.
This whole process should be automatic, and complete
without user intervention.
3.3 Prediction model for user requirements
As mentioned, a smart city is a complex network
of people and things all interdependent, increasingly
instrumented, and interconnected. In this system, this
means that users will always be connected to a cloudlet,
have interaction with IoT, social sensors, and use
different types of services. Therefore, there will be
significant amounts of data streams, which leads to
questions such as:
How to extract knowledge from data collected
from physical and social sensors?
How to identify the right services to be deployed
on the cloudlets based on recommendation?
How to predict service deployment on the
cloudlets in an optimal way?
These questions are the guidelines from which
this smart framework for real-time processing is
derived. By using new technologies that allow the
collection and analysis of data in real-time, we can
make rapid decisions and smart predictions. This
proposed approach for a smart framework for real-time
processing provides intelligent integration, aggregation,
and processing of real-time data streams based on
clouds.
An important source of information for our proposed
framework is social sensors where people share their
opinions, experiences, events, etc., on a daily basis.
Analyzing this data can give us insight into current and
future actions. Especially important is the detection of
current situations on social sensors: what is happening
in some regions, or what are the people usually doing
during a particular time of the day or year? Is there
some unexpected event that people going in a given
direction need to be notified? Based on these questions,
our framework provides answers by collecting social
sensor data streams, analyzing them in real-time, and
detecting “significant trends”. In this way, when an
event or situation occurs, all users affected by this event
will be notified rapidly. This is especially important
in the case of natural disasters (flash flood, fire, etc.)
or traffic jams which require preventive action by
the users. There will be knowledge-bases with known
patterns classified into different categories (such as
sport events, art performances, safety, etc.), but it
will also be possible to identify patterns generated by
unknown events or situations. Knowledge discovery in
order to obtain timely information is a very important
component of this proposed framework. Figure 3 shows
the complete architecture of the smart framework
for real-time event processing based on cloudlet
architecture.
The information generated by the Pattern Discovery
unit is one of the parameters used for predicting user
actions and creating service recommendations.
The Recommendation unit utilizes user profiles,
geo-location history, and contextual information to
understand what the user is interested in. When a
relevant event takes place in the city (based on the
information received from the Pattern Discovery unit),
the user will receive recommendation notifications
about it.
The Prediction unit uses the previously mentioned
user data and knowledge discovery information to
forecast what will be the next user action, for instance,
which path they will take if there is a traffic jam,
which concert they will attend, when will they continue
watching an online movie, etc. Prediction actions are
used to predict what the next user steps will be and to
discover how the user is deciding which action to take.
The idea of this proposed framework is to be
independent from communication technologies and
protocols. As previously mentioned, these units have
Fig. 3 Framework architecture for real-time processing.

a service-oriented access, and can be approached as a
single service.
These will be deployed to the cloudlets, and be
accessible to every user.
4 General Scenarios
In this section we describe cases that would benefit from
this proposed framework and explain how the user can
interact and benefit from it.
4.1 Scenario 1: Event detection
Information collected from social sensors can
significantly contribute to timely inform about
disruptive and preventive actions in people’s
routines. Social media is a source of information
for situations and facts related to the users and their
social environment. When urgent situations (like a
natural disaster) arise, this information needs to be
collected, analyzed, and mined as soon as possible, in
order to inform the population in a timely manner. In
a smart city, the real-time event detection processing
engine is listening for data streams gathered from
different geo-locations in the city, and using this
information the processing unit can then provide
information about what is going on in each part of the
city in real-time. If there is some social event like a
concert or an art event, people with compatible interests
(based on the user’s context, profile, and history)
will be informed, so they can attend the events or
follow the associated social media streams. Similarly,
if some natural disaster occurs (like fire, earthquake,
etc.), users in that area or heading in that direction
(as gathered from their location and context) will be
timely informed. This kind of event detection acts as
preventive action for user’s safety. Figure 4 depicts the
process step actions.
For this scenario, as already described, streams of
data are collected in real-time from different areas
and sources, then filtered, cleaned, and normalized
resulting in a standard form ready for analysis. If
the data shows a known pattern, it will be processed
immediately. However, if it doesn’t, it will be processed
by unknown pattern discovery algorithms that attempt
to extract patterns. This process is integrated with
a learning system that takes into account previously
unknown events. Output data received from this phase
is ready for semantic analysis which makes sense of the
received data stream. The semantic analysis phase is
very important in this step because its output is used to
Fig. 4 Complex event processing.
determine if some event happened or not in a particular
area. In the case of a positive output, users affected or
interested in it will be informed. A concrete example of
this process would be real-time streams of Twitter data
from New York City. The streams are pre-processed
and forwarded to the processing phase. If this phase
identifies a fire at 7th Avenue and 52nd Street, the
process of notifying users in or heading to that area
will be triggered. This information is later used by the
recommendation and prediction units.
4.2 Scenario 2: Predicting user actions and content
recommendation
In the smart city environment, where users are accessing
multiple services simultaneously, it is very important to
predict their actions. The system can be pre-populated
with information collected by other means, such as
habits, preferences, context, locations, service usage
history, and event detection data. Building on this, our
proposal uses these sources of information to predict the
user’s actions. This functionality is part of the real-time
event detection unit, shown in Fig. 5. The user profile
information such as age, education, hobbies, favorite
books, artists, movies, and habits are saved in a database
system with respect to privacy. This information is
used to construct a baseline for determining what the
user preferences are. This baseline is complemented by
other information sources, such as service registration
(containing all the existing services and their meta-
data), or the previously mentioned social sensors. Based
on this information (the user’s movements, time of year,
day, or week), the prediction unit uses data mining
algorithms to predict the next activity or event that
the user may want to take part in. For example, if
a given user in the New York City area frequently
reads news about tennis and the event detection unit

Fig. 5 Data stream processing.
identifies that people are attending the US Open Tennis
Championship, the system determines that the user
will be interested in attending. Recommendations are
especially important in this Smart City scenario where a
large amount of very different services exist as it may be
very difficult for end users to find appropriate services,
so mechanisms are needed to help users find suitable
cloud services. The Recommendation unit is also part
of the real-time processing unit, shown in Fig. 5.
It uses the same source of information as the
Prediction unit (user profile data and information
obtained from the knowledge unit), but also takes
into consideration the user’s context before providing
recommendations and tuning the output scenario.
Context-aware recommendations are designed to react
to a user’s context without their intervention. It consists
of two parts: sensing a context scenario and adapting
the system to a changing scenario by providing desired
services for the user. As an example let’s say that on
weekdays the user requires lunch recommendations for
places with quick service as the time for the lunch break
is limited, but on weekends the suggestions provided
should be for family friendly restaurants.
4.3 Scenario 3: Service continuity, distribution,
and accessibility on cloudlets
Adding continuous data providers as services provides
value-added information, with various levels of
granularity, that best suit the users’ requirements
and interests: concert events, sport events, transport
providers, local and regional planning authorities, etc.
Users are offered the option to pause the services
then continue to use them later, resuming from the
previous state. While this characteristic is more useful
for services that provide video or music streaming or
file downloads, it can also help users to monitor the
progress of events, such as utility restoration after a
severe thunderstorm or traffic congestion on certain
roads so that alternative routes can be planned (e.g.,
someone going to the airport on a Sunday).
Services are distributed among cloudlets depending
on the frequency of usage, the region specification,
users’ preferences, and/or time of year. Some of the
services are more attractive for some region(s) of the
city during a particular time of the year. For example,
if there is a sporting event, such as the aforementioned
US Tennis Open in New York at the end of August
and beginning of September, during this period there
is a possibility that more users will use services related
to the event (ranging from information about the event
to traffic reports). When the event is over, some of
the services will be no longer distributed on all the
cloudlets, but the users will still be able to find those
services.
Service accessibility, as shown in Fig. 6, means
allowing access to all the services when the user
requires them without constraint. Accessibility is also
concerned with the mechanism of finding the right
services with ease. As there are multiple services
available, users need to be provided with search
mechanisms that use a simple interface and timely
results. By doing this, our system acknowledges and
uses the inherent granularity of the architecture to
optimize the interaction with the end user, allowing,
among others, approximate searches, spatial synonyms,
and complex information retrieval, thereby surpassing
the limitations of traditional, query-based systems.
5 A Case Study: Cloud Connected Vehicles
As stated in Ref. [11], among the research
work regarding smart cities, we took note of
smart transportation that relies on sophisticated
infrastructures to overcome situations like traffic
congestion’s, emergencies, and accidents. In this paper,
Fig. 6 Service accessibility.

we consider a vehicular infrastructure built on mobile
cloud computing in order to achieve more efficient
transportation.
Connected vehicles can be viewed as entities that
have not only access to the cloud or cloudlet but are
also equipped with sensors and actuators capable of
collecting data and actuating some vehicle components.
As shown in Fig. 7, each cloudlet covers an area defined
by a limited distance traveled by vehicles.
In addition to the data collected by vehicle sensors
being used for machine-to-machine purposes, such as
maintaining a minimal distance between the vehicles
in order to avoid accidents, some of this data (such
as speed, location, etc.) can be sent by the vehicles
to the cloudlet for it to build a view of the local
traffic in real time. These data and others gathered
by the cloud generate a global view of traffic in order
to provide recommendations to drivers regarding the
weather, congested routes, etc. In addition to the data
provided by physical sensors, helpful when making
recommendations to the drivers, other data may be
supplied by social sensors such as providing relevant
information on events related to the covered area.
Mobile devices are also useful entities that enable
drivers not only to get data from the cloudlet or
cloud but also to externalize services and data on the
cloudlet. Because of the displacement of vehicles
moving from one cloudlet coverage to another, data
must be processed in real time. Hence, the data, as
well as the tasks that process it, must meet temporal
constraints. In fact, data are collected periodically and
each must be consumed/processed by a task before
generating new data, otherwise the preceding collected
data becomes obsolete.
In order to manage the services and data of the
vehicular infrastructure in real-time, our scenario is
based on the following features.
5.1 Data collection by physical sensors
The data collected by the physical sensors are processed
Fig. 7 Vehicle infrastructures.
by application tasks executed within smart objects such
as user mobiles or vehicle devices. Because of their
limited resources, smart objects are mono-processor
devices that host a real-time kernel to periodically
execute application tasks because of periodic data
sensing. These tasks are characterized according to
temporal parameters as defined by Ref. [12]. In fact,
each periodic task Ti in real-time systems is defined
with four parameters (ri ; Ci ; Pi ; Di ) where:
ri is the arrival time of Ti .
Ci is the execution time of Ti , i.e., the time used to
process the collected data.
Pi is called the period, and it is the time interval
that separates two successive data collections,
hence two instances of Ti occur.
Di is the hard deadline of Ti : dDi D ri CDi is the
absolute deadline, a time after which the execution
of Ti is obsolete. Generally, Di is equal to Pi .
Each smart object that collects distinct types of sensor
data can schedule its corresponding tasks using a fixed-
priority scheduling algorithm called Rate Monotonic, in
which the priority is calculated using the tasks’ periods.
Hence, the higher the priority of a task, the lower is its
period.
According to Ref. [12], each configuration of n
periodic tasks, with Pi D Di .Vi /, is schedulable using
Rate Monotonic, if the following condition is satisfied:
nX
1
Ci
Pi
n.2
1
n 1/ (1)
In other words, to process data in real-time, the
corresponding tasks must meet the temporal constraints
dictated by Eq. (1). Another algorithm proved to be
optimal is the Earliest Deadline First (EDF) which
schedules real-time tasks according to their absolute
deadline. In this case, the highest priority is assigned
to the task with the nearest deadline. Reference [12]
proved that a configuration of n periodic tasks, with
Pi D Di .Vi /, is schedulable using EDF, if and only if,
the utilization factor U satisfies the following condition:
U D
nX
1
Ci
Pi
1 (2)
5.2 Data aggregation in the cloudlet
Because of the movement of vehicles, data must be
processed in real-time to provide recommendations in
real-time, such as suggesting an efficient direction for a
vehicle regarding its desired destination.
To supply such recommendations to the driver, either
tasks or data can be offloaded from vehicles to the
cloudlet in order to be aggregated with other data from

other vehicles, thereby building a local view of the
covered area in terms of vehicle-to-vehicle interactions
and moving-vehicle states. Data issued by the in-vehicle
sensors can be enriched with social data provided by the
cloud. To manage these distinct data types, the cloudlet
needs to host an ordinary guest operating systems such
as Windows or Linux, as well as a real-time guest
OS. The real-time VM hosts a Real-Time Operating
System (RTOS) to handle real-time tasks while other
VMs are dedicated to general-purpose services such
as those related to social networks. Hence, tasks
that process the sensing data must be scheduled and
executed by the real-time VM to allow the tasks to meet
their timing constraints. The real-time VM handles
real-time applications by executing the application tasks
according to a real-time strategy such as EDF described
above. However, the deadline cannot be guaranteed to
be met unless the cloudlet hosts a real-time hypervisor.
The use of this type of hypervisor is motivated by
the fact it schedules VMs according to a real-time
scheduling algorithm based on a fixed priority. The
highest priority is assigned to the VM that hosts a real-
time guest OS while lower priorities are assigned to the
other VMs. Hence, an ordinary VM is launched by the
real-time hypervisor only if there are no real-time tasks
to execute within the real-time VM.
Consequently, there are two scheduling levels:
The first level concerns the real-time hypervisor
that schedules VMs and determines the privileges
of the real-time VM (that hosts RToS) among the
others. The scheduling strategy of the hypervisor
can be based on a fixed-priority scheduler.
The second level is related to the scheduling
algorithm provided by RTOS that can use EDF
to schedule tasks. According to the criticality of
the manipulated data, we can distinguish two
categories of real time tasks:
Hard real-time tasks that must process data on time
because a missed deadline could trigger a system
disaster such as an accident.
Firm real-time tasks that can miss their deadline
infrequently, but degrade the system’s quality of
service.
Regarding this classification, we can consider all
the tasks that deal with in-vehicle sensing data as
hard tasks while tasks related to social sensors can be
viewed as firm tasks. The hard tasks can be either
executed periodically by the vehicle’s smart objects
or externalized to the nearest cloudlet. Whichever
execution environment is used, these tasks must be
handled using a real-time scheduling strategy. In
contrast, the firm real-time tasks can be defined
according to an aperiodic task model which processes
social data provided by the cloud to the cloudlet.
Each aperiodic task Ai is modeled thanks to a triple
(ri ; Ci ; Di ) where ri is the time that Ai is ready to
execute, Ci is the execution time of Ai , and Di is its
deadline.
Aperiodic tasks handled by the cloudlet can either
be executed within the real-time VM (RTOS) during
processor slack time or within an ordinary VM.
To verify if RTOS can execute the periodic and
the aperiodic tasks while meeting their temporal
constraints, the utilization rate of the processor must be
lower than 1 as in Eq. (3).
U D
nX
1
Ci
Pi
1; 8Ti (3)
where Ti is a periodic or aperiodic task.
5.3 Data mining and learning methods on the cloud
The cloud can gather data from distinct cloudlets
regarding vehicle movement and other events like
congestion and accidents to design a global view of
the traffic. This view is also enriched with other data
obtained from social networks. Learning methods, such
as context recognition, can be used to compare current
situations with previous ones and help the system be
more adaptive and reactive.
5.4 Recommendations
The system, based on its distributed infrastructure
as well as its cooperative tasks for getting accurate
information on traffic, can issue some directives to
drivers in terms of the most efficient direction to
take or a travel route to avoid social events around a
recommended itinerary, etc.
6 Conclusions and Future Work
In the Smart City context, integrating information from
different sources is a necessity in real-time to make
smarter decisions for individual citizens and for the city
as a whole.
In this paper, we propose an approach for a
framework, based on prediction models, that provides
integration, aggregation, and processing of real-
time tracking data based on cloudlet architecture
and cloud principles. By adopting a service-oriented
approach when collecting, integrating, storing, and
intelligently analyzing social and user data, we achieve
better granularity, accessibility, recommendation, and
prediction of user requirements. In addition, our

approach provides service continuity capabilities that
allow the system to provide value added information
that suits the users’ requirements at any time and place.
The next steps involve the implementation of
the proposed framework by realizing the scenarios
presented here and using existing standards in the field.
In order to evaluate this implementation, we will build
a prototype to conduct user performance requirements,
using multiple metrics that measure request-response
time, performance, processing time, and other relevant
measures.
Acknowledgements
This work was supported by the National Institute of
Standards and Technologies (NIST), and conducted within
a collaboration under Information Technology Laboratory,
Advanced Network Technologies Division (ANTD) and
the Universities of Grenoble and CNAM. Our special
thanks go to Dr. Abdella Battou, ANTD division chief
for his support and advises.
References
[1] M. Weiser, R. Gold, and J. S. Brown, The origins of
ubiquitous computing research at PARC in the late 1980s,
IBM Systems Journal, vol. 38, no. 4, pp. 693–696, 1999.
[2] R. Y. Clarke, Smart cities and the internet of everything:
The foundation for delivering next-generation citizen
services, IDC Government Insights #GI243955, 2013.
[3] L. Heuser, C. Alsdorf, and D. Woods, eds. International
Research Forum 2008. Evolved Technology Press, 2009.
[4] M. Satyanarayanan, G. Lewis, E. Morris, S. Simanta,
J. Boleng, and K. Ha, The role of cloudlets in hostile
environments, IEEE Pervasive Comput., vol. 12, no. 4, pp.
40–49, 2013.
[5] H. Tamai, Fujitsu’s approach to smart cities, vol. 50, no. 2,
pp. 3–10, 2014.
[6] S. Dirks, C. Gurdgiev, and M. Keeling, Smarter cities for
smarter growth, IBM Global Business Services, 2010.
[7] B. Guo, Z. Yu, X. Zhou, and D. Zhang, From participatory
sensing to mobile crowd sensing, in 2014 IEEE
International Conference on Pervasive Computing and
Communication Workshops (PERCOM WORKSHOPS),
2014, pp. 593–598.
[8] S. S. Ara, Z. U. Shamszaman, and I. Chong, Web-of-
objects based user-centric semantic service composition
methodology in the internet of things, International
Journal of Distributed Sensor Networks, vol. 2014, p.
482873, 2014.
[9] D. Guedes, W. Meira, and R. Ferreira, Anteater: A service-
oriented architecture for high-performance data mining,
IEEE Internet Computing, vol. 10, pp. 36–43, 2006.
[10] M. Satyanarayanan, P. Bahl, R. Cceres, and N. Davies, The
case for VM-based cloudlets in mobile computing, IEEE
Pervasive Computing, vol. 8, no. 4, pp. 14–23, 2009.
[11] G. Dimitrakopoulos, Intelligent transportation systems
based on internet-connected vehicles: Fundamental
Research areas and challenges, in the 11th IEEE
International Conference on ITS Telecommunications,
2011, pp. 145–151.
[12] C. L. Liu and J. W. Layland, Scheduling algorithms
for multiprogramming in a hard real-time environment,
Journal of the ACM, vol. 20, no. 1, pp. 46–61, 1973.
O. Kotevska is a PhD student at Joseph
Fourier University (UJF) in Grenoble,
France and Guest Researcher at ITL Lab
NIST, Washington DC Metro, USA. She
received the MSc and BSc degrees from
the University “Ss Cyril and Methodius”
in Skopje, Macedonia in 2013 and 2008,
respectively. Her research is in the field
of text mining, machine learning, cyber physical systems, and
software design.
S. Bouzefrane is an associate professor
and has the accreditation to conduct
research (HDR) at the Conservatoire
National des Arts et Metiers of Paris.
She received her PhD degree in computer
science from the University of Poitiers
(France) in 1998. After four years at the
University of Le Havre (France), she joined
in 2002 the CEDRIC Lab of CNAM where she worked on real-
time and embedded systems. She is the co-author of many books
(Operating Systems, Smart Cards, and Identity Management
Systems). Her current research areas cover service architecture
and security in mobile cloud computing, internet of things, and
virtualization.
Ahmed Lbath is a full professor of
computer science at University of
Grenoble1 (MRIM/LIG Laboratory),
and he is also conducting research in
collaboration with ITL Lab NIST in
Washington DC metro area where he
carried out research activities as visiting
professor. He is the IUT Deputy Director,
former Head of CNS Department, and former Director of R&D
in a French company. He received his PhD degree in computer
science from the University of Lyon and hold an “Habilitation
Diriger des Recherches”. He is currently acting as project
manager coordinating scientific collaborations in the domain of
Cyber Physical Systems and smart cities. His research interests
include cyber physical human systems, smart cities, mobile
cloud computing, recommendation systems, web services, GIS,
and software design. He published several patents, papers in
books, journals, and conferences and he regularly serves as
co-chair and/or member of several committees of International
conferences, journals, and research programs.

Toward a real time framework in cloudlet-based architecture

More Related Content

What's hot (17)

Similar to Toward a real time framework in cloudlet-based architecture (20)

More from redpel dot com (20)

Recently uploaded (20)

Toward a real time framework in cloudlet-based architecture