Smart Cities Architectures

A Systematic Review

Gustavo H. R. P. Tomas

1,2

, Welington M. da Silva

1,2

, Paulo A. da M. S. Neto

, Vinicius C. Garcia

Alexandre Alvaro

and Kiev Gama

Informatics Center, Federal University of Pernambuco (UFPE), Recife, Brazil

Recife Center for Advanced Studies and Systems (C.E.S.A.R), Sorocaba, Brazil

Federal University of S

ao Carlos (UFSCar), Campus Sorocaba, Sorocaba, Brazil

Keywords:

Smart Cities, Internet of Things, Architecture.

Abstract:

The smart cities concept arises from the necessity of managing several problems caused by the unbridled

population growth at urban centers. To make a city become “smart” it is needed to employ Information and

Communication Technologies (ICT) to access, process and deliver information according to the urban context.

This information can be employed to mitigate several urban issues, such as trafﬁc jams, high natural resource

consumption, epidemias, sustainability, waste management, low quality and life expectancy of citizens, among

others. Thus, the increasing need to create architectures that are able to interact with the Internet of Things,

i.e., several built-in devices, appliances, sensors and actuators embedded in each urban context. This work is a

systematic review regarding proposals for such architectures. After selecting the relevant approaches, we have

identiﬁed a set of issues that these approaches aim to solve and some architectural patterns employed.

1 INTRODUCTION

The most accepted deﬁnition for the term City is de-

scribed in Kuper (Kuper, 1995): a relatively large and

permanent settlement. Usually, big cities have high

population density, with its citizens living in constant

interaction with industries, business and services.

According to a UNESCO report released in 1950

(Nations, 2007), 30% of world population lived in ur-

ban areas. This number grew to 50% in 2010 and

it is estimated that in 2050 the percentage of peo-

ple living in large urban centers will be around 70%.

Consequently, several issues are easily identiﬁed in

urban centers, such as trafﬁc jams, high natural re-

source consumption, epidemias, sustainability, waste

management, low life expectancy of citizens, among

others.

This unbridled populational growth sets a chal-

lenge of combining city services with Information and

Communication Technologies (ICT) in order to miti-

gate these urban issues and promote better living con-

ditions for citizens. In other words, the challenge

is how to turn a city into a Smart City and it has

been widely discussed both in academic and indus-

try, from projects and initiatives considering several

viewpoints.

To achieve this goal, it is needed the utilize of

some sort of sensing capability in the targeted ob-

jects (vehicles, people, home, etc.) and ensuring con-

nectivity. It is from this unitary monitoring mecha-

nism that a holistic view of the city is built, dedi-

cated to efﬁcient maintenance of its services. This

vision can be specialized to a scenario, where each

object is equipped with enough technology and intel-

ligence to transform it in a data supplier/consumer. It

is reasonable to picture this distribution and integra-

tion with each object owning part of the needed data

to some computation executed in a centralizing en-

tity, responsible for the processing and management.

This set of objects acting collaboratively in pursuit of

a well deﬁned common purpose is called Internet of

Things(IoT) (Atzori et al., 2010), and it constitutes

the essential technological foundation for implement-

ing smart cities.

Smart Cities offer a new approach to optimize ser-

vices, reducing costs and improving citizens life qual-

ity. This way, it is needed that sensors become smart,

since they represent the peripheral elements of a com-

plex future ICT world. However, due to the speciﬁc

application ﬁeld, smart sensors are very heteroge-

neous in terms of communication technologies, sens-

ing features and elaboration capabilities(Fazio et al.,

410

H. R. P. Tomas G., M. da Silva W., A. da M. S. Neto P., C. Garcia V., Alvaro A. and Gama K..

Smart Cities Architectures - A Systematic Review.

DOI: 10.5220/0004417204100417

In Proceedings of the 15th International Conference on Enterprise Information Systems (ICEIS-2013), pages 410-417

ISBN: 978-989-8565-60-0

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

2012).

Thus, it arises the need of establishing a smart city

architecture able to store, combine, process and de-

livery contextualized information for the Internet of

Things (Sanchez et al., 2011) (Fazio et al., 2012).

These architectures must enable the integration be-

tween different urban contexts, e.g., using smart city

architecture is possible to analyze the impact of every

city event in different contexts.

In this context, this Systematic Literature Review

(SLR) aims to analyze the main approaches that de-

scribe the architectures of smart cities based on the

Internet of Things. According to Kitchenham’s guide-

lines (Kitchenham and Charters, 2007), the aim of

an SLR is intended to support the development of

evidence-based guidelines through selection and syn-

thesize the most relevant research. Many architectural

approaches have been proposed in the literature, and

this paper will focus on architectures that combine in-

ternet of things with smart city. This way, this litera-

ture review will be based on three questions:

Q1. What architectures have been proposed combin-

ing the internet of things with smart cities?

Q2. What are the main goals and issues addressed by

these architectures?

Q3. What are the patterns used by the existing pro-

posals?

The remainder of this paper is structured as fol-

lows. Section 2 describes the applied research

methodology. Section 3 presents and analyzes the

results related to review questions, while Section 4

closes the paper by describing the main conclusions.

2 RESEARCH METHODOLOGY

The systematic review is a means of identifying, eval-

uating, interpreting and comparing all available that is

research relevant to a particular question (Kitchenham

and Charters, 2007). Thus, for review purposes, only

approaches that combine smart cities with the internet

of things were considered.

This way, this section aims to describe the search

strategy and all steps followed to ﬁlter and extract

relevant data from approaches found. Our research

methodology followed all the stages and steps recom-

mended in Kitchenham’s guidelines (Kitchenham and

Charters, 2007).

2.1 Search Strategy and Data Sources

The strategy used to construct the search terms fol-

lowed the same approach used in (Kitchenham and

Charters, 2007) (Chen et al., 2009) (Khurum and

Gorschek, 2009):

• Derive main terms based on the research question

and the topics being researched;

• Determine and include synonyms, related terms,

and alternative spelling for major terms;

• Incorporate alternative spellings and synonyms

using boolean operator “OR”;

• Link main terms using boolean operator “AND”.

Following this strategy, we constructed the search

string as bellow:

(smart city OR intelligent city OR digital city OR

urban environment) AND (internet of things OR het-

erogeneous sensors) AND (architecture OR middle-

ware OR platform)

Due to the variation of the search features pro-

vided by the main digital sources of literature (such

as IEEExplore, Springer Link, and ACM Digital Li-

brary), it was not possible to use a single search string

for all the digital sources (Chen et al., 2009). Thus,

we made a signiﬁcant effort to ensure that the search

strings used were logically and semantically equiva-

lent.

After the search string deﬁnition, we searched the

primary studies in these digital sources (1. IEEEx-

plore; 2. Science Direct; 3. ACM Digital Library;

4. Springer Link; 5. CiteSeerX; 6. Academia.edu;

and 7. ISI Web of Science). Furthermore, consid-

ering that smart city involves business concepts, we

also searched patents on the World Intellectual Prop-

erty Organization (WIPO)

About the manual search, we performed it in

these conferences (1. International Conference on

Computational Intelligence, Modeling and Simula-

tion (IJCCI); 2. International Conference on Intel-

ligent Environments (IE); 3. Multimedia Informa-

tion Networking and Security (MINES); 4. Emerging

Technologies for a Smarter World (CEWIT); 5. Inter-

national Conference on Innovative Mobile and Inter-

net Services in Ubiquitous Computing (IMIS)).

The quality of search engines may have inﬂuenced

the completeness of the identiﬁed primary studies.

That means our search may have missed those studies

whose authors would have used other terms to spec-

ify smart cities architectures associated to internet of

things.

2.2 Study Selection

We selected only works that propose architectures

or frameworks to centralize the several contexts and

www.wipo.int

SmartCitiesArchitectures-ASystematicReview

411

technologies involving the urban environment. Thus,

aiming to increase analysis reliability, all the authors

were involved in source selection process.

After deﬁning the search string and data sources,

the ﬁlters to reﬁne the work found were deﬁned. Fig-

ure 1 illustrates the results of each step with the cor-

responding ﬁlters.

Step 1

Step 3

Step 2

Identify relevante studies in

conference proceedings and

relevant databases

6435

Exclude studies on the basis

of titles

Remove duplicates

Exclude studies outside the

range 2008-2012 year

Step 4

4310

Exclude studies on the basis

of abstracts

Step 5 33

Exclude studies on the basis

of relevance

Step 6 11

Figure 1: Steps of the search strategy.

The goals of the ﬁlters were to select the main

approaches that describe architecture of smart cities

based on the internet of things. Thus, the ﬁrst ﬁlter

matches all work found in relevant databases, previ-

ously described, obtained from the search string.

The second ﬁlter was applied to exclude the work

in which the publication year is outside the range

2008-2012. This interval was chosen after analysis

and veriﬁcation that the work published before 2008

dealt with the internet of things applied to smart city

in isolation, i.e., the work proposed to solve problems

in a speciﬁc scenario using only a technology, with-

out analyzing the connection between different urban

contexts.

Regarding the inclusion criteria, the work was in-

cluded if:

• it proposed an architecture or framework to cen-

tralize the information from several urban con-

texts or;

• it described an IoT middleware design that met

more than one technology and context and;

• the approach has not been addressed in earlier

studies analysed;

During the review several papers were found de-

scribing the same architecture. In this case, only the

most complete work was included. After this stage,

there were 11 approaches left.

This high discrepancy between the initial amount

of approaches and the amount of resulting primary

studies is discussed and explained in Kitchenham et

al. (Kitchenham et al., 2009).

Regarding patents, we did not ﬁnd any patent re-

lated to integration of different city contexts. Usu-

ally, the patents that are closer to the context of this

systematic review generally describe a speciﬁc algo-

rithm or technique optimization in a controlled envi-

ronment.

3 RESULTS AND ANALYSIS

After performing the search on relevant databases,

and ﬁltering them according to the criteria described

in the previous section, 11 approaches remained.

These approaches were read and discussed among

the authors, aiming to highlight the topics related to

the review questions. This section aims to discuss

these approaches, starting with an approach review,

describing addressed issues, applied patterns, and ﬁ-

nally, analyzing all the results.

If an approach had a name, we used that name.

Otherwise, we created a name using the ﬁrst authors

surname followed by the publication year. Table 1

lists each of the 11 approaches in chronological order.

Table 1: List of approaches reviewed.

ID Approach Year Reference

1 Al-Hader’2009 2009 (Al-Hader et al., 2009)

2 Anthopoulos’2010 2010 (Anthopoulos and Fitsilis, 2010)

3 SOFIA 2010 (Filipponi et al., 2010)

4 EcoCity/ISMP-UC 2011 (Lee et al., 2011)

5 CPAF 2011 (Mostashari et al., 2011)

6 SmartSantander 2011 (Sanchez et al., 2011)

7 IMS 2011 (Shao, 2011)

8 USN 2011 (Hern

andez-Mu

noz et al., 2011)

9 Wu’2011 2011 (Wu et al., 2011)

10 Fazio’2012 2012 (Fazio et al., 2012)

11 S

OiA 2012 (Vega-Barbas et al., 2012)

3.1 What Architectures have been

proposed Combining the Internet of

Things with Smart Cities?

Architectures for smart cities can be designed start-

ing from different areas of scientiﬁc knowledge. This

subsection aims to brieﬂy discuss architectures found,

ordered by the publication date, highlighting its goals

and features proposed to solve the issues they address.

Starting with Al-Hader’2009 (Al-Hader et al.,

2009), it proposes an architecture supported by four

principles: applications, business, process manage-

ment and network infrastructure. The ﬁrst principle

corresponds to applications which uses different tech-

nologies for monitoring sensors, such as Global Posi-

tioning System (GPS). The vast majority of these ap-

plications meet cities operational demand, however,

ICEIS2013-15thInternationalConferenceonEnterpriseInformationSystems

412

using the rules deﬁned in the second principle - busi-

ness - they can aggregate economically viable solu-

tions. The third principle is the processes manage-

ment in which relationships, rules, strategies and poli-

cies between smart cities applications and business

units are deﬁned. Finally, the last principle is a net-

work infrastructure, responsible for connecting the

other three principles.

Anthopoulos et al. (Anthopoulos and Fitsilis,

2010) proposed an architecture based on the analy-

sis of different initiatives already implemented, mod-

eled applying the principles of Enterprise Architec-

ture, which sets up a mission, information, technolo-

gies, speciﬁc processes and framework required to

implement it. Anthopoulos et al. highlight that the

construction of a smart city must forecast legacy sys-

tems integration to the new infrastructure, migration

and reuse of existing data, simpliﬁcation of urban pro-

cesses through participatory performance, resource

utilization optimization, systems and equipment inter-

operability, and providing tools for monitoring, man-

agement and analysis.

The SOFIA (Filipponi et al., 2010) project is cen-

tered on the concept of smart environment as an

ecosystem of interacting objects, such as sensors, de-

vices, appliances and embedded systems in general,

self-organized, providing services and custom pro-

cessing. The SOFIA architecture is event driven, in

which an event is an observable change in the state

of an Information Technology (IT) system; it can be

triggered by real world events, such as presence de-

tection, timeouts etc., by internal events like the re-

ception of a message (e.g., a command) or the com-

pletion of a task.

Sensors monitoring and management are also the

goal of the ISMP-UC (Lee et al., 2011) project. The

EcoCity architecture has three basic layers. The bot-

tom layer consists of different types of sensors, actu-

ators, and other devices distributed accross the city.

On the top layer there is a range of U-eco City Ser-

vices. Above these layers lies the middleware which

collects and processes data and contextual informa-

tion. Its service-oriented architecture enables services

to be developed independently and invoked through

standardized Web services interfaces.

Moreover, Mostashari et al. (Mostashari et al.,

2011) proposes a framework, called Cognitive Pro-

cess Architecture Framework (CPAF), which allows

different urban processes to be designed with cogni-

tive abilities. In this context, cognition is the ability

of a system to learn from previous experiences and

adapt its behavior based on them. A cognitive system

is able to sense, perceive and respond to changes in

the environment, and can therefore improve a systems

performance by increasing its adaptive capacity.

Another approach with several sensors embedded

in an urban environment is SmartSantander (Sanchez

et al., 2011). The number of devices to be deployed

in Santander and its surroundings is foreseen to climb

up to 12,000 devices, thus creating the basis for devel-

opment of a future Smart City. The SmartSantander

aims to produce the following key outcomes: i) An

architectural reference model for open real-world In-

ternet of Things (IoT) experimentation facilities; ii) A

scalable, heterogeneous and trustable large-scale real-

world experimental facility; iii) A representative set

of implemented use cases for the experimental facil-

ity; and iv) A large set of Future Internet experiments

and results.

As SmartSantander focuses on interoperability of

objects, the IMS (Shao, 2011) proposes an approach

that combine IoT with citizens. According to the au-

thors, the development of ICT is related to the prox-

imity with people. For this, IMS is based on three

layers: access, session and application. The access

layer is the lowest layer and provides capability to ac-

cess IMS network from different terminals. The ses-

sion layer provides session establishment, modiﬁca-

tion, and termination; providing session management

to the upper layer. Finally, the application layer, the

highest one of IMS network, allows application de-

ployment.

The interoperability of objects also is explored by

Hernandez et al. (Hern

andez-Mu

noz et al., 2011),

that proposed an architecture called Ubiquitous Sen-

sor Network (USN). The goal was to provide an in-

frastructure that enabled the integration of heteroge-

neous and geographically dispersed sensors in a cen-

tralized technological base, in which services could

be developed at minimal cost; to this end, the project

based itself on the integration of Internet of Things

and Internet of Services. Additionally, the architec-

ture included a module known as USN-Gateway that

enabled interoperability between sensor and the IP

network.

Like the USN, Wu’2011 (Wu et al., 2011) pro-

poses a middleware to manage the spatial information

from multiple sources always has different formats,

structures, and spatial database engines. The mid-

dleware builds a bridge to heterogeneous spatial and

application by the support of Multi-Agent and Web

Service. This platform follows the Service-Oriented

Architecture (SOA), and it is consisted of two main

parts: Contract-First model and message conversion

agent model.

Fazio’2012 (Fazio et al., 2012) proposes an ar-

chitecture which allows the aggregation of different

information types from several sensors embedded on

SmartCitiesArchitectures-ASystematicReview

413

different urban contexts. The main goal of this archi-

tecture is to provide contextualized data, combining

several data sources. To this end, the architecture con-

sists of four levels: I) REST APIs, which allow on-

demand interactions with clients, applications and/or

other services; II) The Sensor Observation Service

Agent (SOS Agent), which supports all functionalities

for describing sensors and observations; III) Sensor

Manager, able to interact with sensors, it coordinates

their activities and collect data for the upper layers.

It provides a uniform management of heterogeneous

sensors; IV) the Sensing Infrastructure (SI), which is

composed of several different sensors and sensing de-

vices.

The architecture called S

OiA, described in

(Vega-Barbas et al., 2012), also manage the different

information types and interoperability issues among

such a large number of heterogeneous actors. The

OiA architecture is syntactic and semantic Service-

Oriented Architecture that allows the integration of

any type of object or device on the Internet of Things.

It allows an ad-hoc dynamic application composition

in cooperating and distributed environments which

are uniformly described. To such extent it has been

deﬁned within the architecture design a set of seman-

tic dependency management modules which track ser-

vices and resources allowing the already created ap-

plications to continue running despite changes of the

context.

3.2 What are the Main Goal and Issues

Addressed by these Architectures?

These architectural approaches are proposed to meet

several issues on urban systems. As the analysis ap-

proaches were not proposed for speciﬁc purposes,

the techniques used to solve these issues are easily

adapted to different urban contexts. Table 2 summa-

rizes the architectures IDs with the issues addressed.

One of the most discussed and studied issue on

IoT projects is the interoperability of objects, where

the object is an abstraction of a sensor, actuator or any

device able to perform some sort of computation (At-

zori et al., 2010). In fact, this is a critical requirement

to the consolidation of any platform that uses a range

of objects with different technical speciﬁcations and

communication protocols. The majority of architec-

tures explicitly designate a module or layer to meet

this requirement, as in SOFIA (Filipponi et al., 2010),

USN (Hern

andez-Mu

noz et al., 2011), Wu’2011 (Wu

et al., 2011) and S

OiA (Vega-Barbas et al., 2012).

In particular, the SOFIA project developed an inter-

operable platform and select a set of vertical appli-

cations to compose this smart environment based on

embedded systems. Regarding to USN (Hern

andez-

noz et al., 2011), this module - known as USN-

Gateway - besides being responsible for the interop-

erability of objects on the platform, it also implements

mechanisms that allow interoperability between sen-

sors network and the IP network. In case of Wu’2011,

the interoperability is implemented by two main parts:

Contract-First model and message conversion agent

model. i) Contract-First agent model integrates dis-

tributed and heterogeneous spatial information. ii)

Message conversion agent model resolves the proto-

col conﬂict between different applications problem.

Finally, in S

OiA the integration of heterogeneous ob-

jects (legacy and future created) is made from exten-

sible mechanism using OSGi

platforms.

Another important feature inherent to the smart

cities context is continuous real-time monitoring.

The real-time monitoring is an important require-

ment for keeping city services constantly updated,

due to an event triggered in a city domain can in-

ﬂuence the decisions-making of others. Furthermore,

the real-time monitoring is the most valuable instru-

ment to provide relevant information that will be used

to predict phenomena. In this context, the main ar-

chitectures which implement this feature built into

the structure are Al-Hader’2009 (Al-Hader et al.,

2009), SOFIA (Filipponi et al., 2010), SmartSan-

tander (Sanchez et al., 2011) and USN (Hern

andez-

noz et al., 2011).

This diversity of objects capturing data and oper-

ating over several city domains combined with real-

time monitoring raises a great opportunity for em-

ploying sustainable policies on the data analysis pro-

cess. Due to the high coverage of all city areas, archi-

tectures must include, since its conception, sustain-

able policies. These policies are related to environ-

mental, economic and social aspects of each domain.

The architectures that seek to integrate sustainability

are EcoCity/ISMP-UC (Lee et al., 2011) and Smart-

Santander (Sanchez et al., 2011), however they only

address the environment issue.

Aiming to adopt these sustainable policies, a

smart city architecture must provide mechanisms to

improve social aspects for citizens to be involved ef-

fectively; otherwise the entire investment will not be

effective. One example is the Digital City of Trikala,

Greece, after ﬁve million euro spent on infrastructure

maintenance and 6 years of operation, the population

did not use it and was not even aware of the digital

services availability (Anthopoulos and Fitsilis, 2010).

The only architecture that contains a citizen’s involve-

ment principle is IMS (Shao, 2011), in which the ar-

chitecture provides apparatus for citizens to interact

www.osgi.org

ICEIS2013-15thInternationalConferenceonEnterpriseInformationSystems

414

with the things anywhere.

This proximity between devices and people, usu-

ally with access to the Internet, originated the term

ubiquity (Sp

ınola and Travassos, 2012). Ubiquity

comprehends any mobile technology feat capture in-

formation about the environment and citizens or act

over the same (Sanchez et al., 2011). Ubiquity is an-

other key requirement that must be explored in smart

cities. The ubiquity concept is an essential key for the

implementation of real-time monitoring requirement,

that was explored on SmartSantander (Sanchez et al.,

2011) and USN (Hern

andez-Mu

noz et al., 2011) ap-

proaches.

This ubiquity eventually produces a large amount

of data (Filipponi et al., 2010), generating the phe-

nomenon known as Big Data (Gopalkrishnan et al.,

2012). The smart city space can be considered an

ubiquitous computing environment where sensors are

installed in many different areas and provide infor-

mation about the environment to interested devices,

applications and users (Sanchez et al., 2011). All

these high data volumes need to be protected, in

compliance with security policies, both in relation

to the availability and storage. These goals are dis-

cussed in SmartSantander (Sanchez et al., 2011) and

Fazio’2012 (Fazio et al., 2012).

Besides, to improve the data analysis process, sen-

sors location is needed (Shao, 2011) (Hern

andez-

noz et al., 2011) (Vega-Barbas et al., 2012). From

the geographic location of the sensors it is possi-

ble to perform actions according to the environment

(Jauregui-Ortiz et al., 2012). This feature is ad-

dressed by the approaches IMS (Shao, 2011), USN

(Hern

andez-Mu

noz et al., 2011) and S

OiA (Vega-

Barbas et al., 2012). In IMS, location is an important

information in the investigation the citizens behavior;

in USN, the location discovery is considered as a es-

sential requirement of architecture. S

OiA aims to

unify the countless device discovery protocols that are

currently used. One of the ideas issued to overcome it

aims to create a new standardized IoT common pro-

tocol or interoperable architectures that abstract these

inconsistencies.

Finally, urban environments are essentially a set

of complex systems available to meet the needs of its

citizens. Architectures that are willing to give support

to these systems should consider them as complemen-

tary in the search for effective urban management,

rather than treating them isolated. For this end, the

Anthopoulos’2010 (Anthopoulos and Fitsilis, 2010),

CPAF (Mostashari et al., 2011), IMS (Shao, 2011)

and Fazio’2012 (Fazio et al., 2012) approaches im-

plemented some service composition techniques.

3.3 What are the Patterns used by the

Existing Proposals?

Regarding the addressed patterns it was difﬁcult to

identify which patterns each architecture used due to

the lack of design and implementation details. Nev-

ertheless, some patterns were identiﬁed, specially the

patterns regarding to communication and messaging

between components.

Among these communication patterns, the most

employed is the publisher-subscriber pattern

(Buschmann et al., 1996). In the publisher-subscriber

pattern, publishers send messages to a speciﬁc

channel; the subscriber listening to that speciﬁc

channel receives all messages. Among the analyzed

approaches, four of them explicitly described the use

of this pattern: SOFIA (Filipponi et al., 2010), USN

(Hern

andez-Mu

noz et al., 2011), SmartSantander

(Sanchez et al., 2011), Wu’2011 (Wu et al., 2011)

and Fazio’2012 (Fazio et al., 2012).

The main reason is the real world natural model-

ing (Filipponi et al., 2010), since several events can

be triggered simultaneously and can be interpreted by

the same component. Additionally, this pattern en-

ables service composition since each service can re-

ceive information from other services and combine

them appropriately. Moreover, in USN this pattern

is used to constantly monitor the sensors, appliances

and devices.

Besides the publisher-subscriber pattern,

Fazio’2012 (Fazio et al., 2012) employs another

two patterns: facade and abstract factory (Gamma

et al., 2001).

By analyzing the approaches, we noticed that the

architectural model based on layers (Buschmann

et al., 1996) is more employed, due to the facility in

compose and expand components (Hern

andez-Mu

noz

et al., 2011) (Fazio et al., 2012) (Vega-Barbas et al.,

2012), being used by 8 studied approaches. How-

ever, two approaches are not arranged in layers: Al-

Hader’2009 (Al-Hader et al., 2009) and SOFIA (Fil-

ipponi et al., 2010). The SOFIA architecture is based

on event driven and Al-Hader’2009 is based on prin-

ciples.

Due to the difﬁculty of identifying information

concerning to the architectural patterns of each ap-

proach, it was not possible to identify the patterns

used in Al-Hader’2009, CPAF and IMS.

3.4 Analysis

After answering the questions, others factors were an-

alyzed. The ﬁrst factor is the adhesion to the layered

model for most of the studied approaches, primarily

SmartCitiesArchitectures-ASystematicReview

415

Table 2: Approaches reviewed summary.

Architecture ID 1 2 3 4 5 6 7 8 9 10 11

Issue

Object interoperability

Real-time monitoring

Sustainability

Social aspects

Ubiquity

Security

Sensor discovery

Service composition

Pattern

Publisher-Subscriber

Facade

Abstract factory

Layers

due to the simple composition and expansion of the

components.

Regarding the issues from this systematic review

may be noted that no architecture met all the issues

surrounding smart city architecture. We believe that

usually the architectures are proposed for speciﬁc pur-

poses, to solve problems of small niches in the smart

city context.

It may also be noted that for implementation of

these smart city architecture heterogeneity is essen-

tial, due to factors, such as the high range of devices

and proximity to citizens.

Another relevant point is the proximity with the

citizens. To turn a city into smart city is a social rather

than a technological challenge. So, all technological

mechanism should be developed based on the citizens

needs, so that they feel involved in the process.

Furthermore, a smart city should monitor events

and take preventive decisions in real time. This pre-

vention must be done examining several urban con-

texts. Thus, the architectures are focusing on the com-

position and aggregation of services, data, informa-

tion and contexts.

Finally, the approaches analyzed are in differ-

ent stages of validation. Some approaches, such as

SmartSantander and SOFIA, already have some sta-

tistical data and case studies that demonstrate the

architecture behavior in some contexts. Moreover,

some approaches, such as CPAF, have not presented

any practical validation to the proposed design.

4 CONCLUSIONS

Throughout this systematic review several smart city

architectures based on the internet of things were an-

alyzed. The section 2 described the research method-

ology used and the results were analyzed on section

3 related to the reviewed questions.

At the end of this review, we identiﬁed a set of

issues that these architectures aim to solve and that

can be considered as minimum requirements for im-

plementing a smart city architecture. Each issue rep-

resents a core set of features that are critical for the

deployment and adoption of a smart city.

Regarding the patterns, we conclude that the lay-

ered model seems to be the most promising, since

several approaches have successfully implemented it.

Moreover, this pattern provides some interesting ben-

eﬁts in relation to the smart city context.

Furthermore, the analyzed approaches are in dif-

ferent validation stages. Some approaches, such as

Smart Santander and SOFIA, have already presented

some statistical data and case studies that demonstrate

the architecture behavior in some contexts. Moreover,

some approaches, such as CPAF, were designed with-

out practical validations.

From the analyzed studies we noted that no archi-

tectures meets all the issues of a smart city. Thus,

for future work, we will design and implement a IoT-

based smart city architecture. As a case study, it will

be considered essential services in Brazilian context

as a means to validate the architecture in a real sce-

nario.

ACKNOWLEDGEMENTS

This work was partially supported by the National In-

stitute of Science and Technology for Software engi-

ICEIS2013-15thInternationalConferenceonEnterpriseInformationSystems

416

neering (INES)

, FAPESP

, grant 2012/10157-1,

FACEPE

, grants 573964/2008-4, APQ-1037-1.03/

08 and CESAR

REFERENCES

Al-Hader, M., Rodzi, A., Sharif, A., and Ahmad, N. (2009).

Smart city components architicture. In Computational

Intelligence, Modelling and Simulation, pages 93–97.

Anthopoulos, L. and Fitsilis, P. (2010). From digital to ubiq-

uitous cities: Deﬁning a common architecture for ur-

ban development. In Proceedings of the 6th Interna-

tional Conference on Intelligent Environments, pages

19–21.

Atzori, L., Iera, A., and Morabito, G. (’2010’). The internet

of things: A survey. Comput. Netw., 54(15):2787–

2805.

Buschmann, F., Meunier, R., Rohnert, H., Sommerlad, P.,

and Stal, M. (1996). Pattern-oriented software archi-

tecture: a system of patterns, vol.1. Wiley.

Chen, L., Ali Babar, M., and Ali, N. (2009). Variability

management in software product lines: a systematic

review. In Proceedings of the 13th International Soft-

ware Product Line Conference, SPLC ’09, pages 81–

90, Pittsburgh, PA, USA. Carnegie Mellon University.

Fazio, M., Paone, M., Puliaﬁto, A., and Villari, M. (2012).

Heterogeneous sensors become homogeneous things

in smart cities. In Sixth International Conference on

Innovative Mobile and Internet Services in Ubiquitous

Computing (IMIS), pages 775 –780.

Filipponi, L., Vitaletti, A., Landi, G., Memeo, V., Laura,

G., and Pucci, P. (2010). Smart city: An event driven

architecture for monitoring public spaces with het-

erogeneous sensors. In 4th International Conference

on Sensor Technologies and Applications, pages 281–

286, Washington, DC, USA. IEEE Computer Society.

Gamma, E., Helm, R., Johnson, R., and Vlissides, J. (2001).

Design patterns: Abstraction and reuse of object-

oriented design. Springer.

Gopalkrishnan, V., Steier, D., Lewis, H., and Guszcza, J.

(2012). Big data, big business: bridging the gap. In

Proceedings of the 1st International Workshop on Big

Data, Streams and Heterogeneous Source Mining: Al-

gorithms, Systems, Programming Models and Appli-

cations, BigMine ’12, pages 7–11, New York, NY,

USA. ACM.

Hern

andez-Mu

noz, J. M., Vercher, J. B., Mu

noz, L.,

Galache, J. A., Presser, M., G

omez, L. A. H., and Pet-

tersson, J. (2011). Smart cities at the forefront of the

future internet. In Domingue, J., Galis, A., Gavras,

A., Zahariadis, T., and Lambert, D., editors, The fu-

ture internet, pages 447–462. Springer-Verlag, Berlin,

Heidelberg.

http://www.ines.org.br

http://www.fapesp.br

http://www.facepe.br

http://www.cesar.org.br

Jauregui-Ortiz, S., Siller, M., Ramos, F., and Scalabrin, E.

(2012). Smart environmental architecture for node lo-

calization in a wireless sensor network. In Intelligent

Environments (IE), 2012 8th International Conference

on, pages 222 –227.

Khurum, M. and Gorschek, T. (2009). A systematic review

of domain analysis solutions for product lines. Journal

of Systems and Software, 82(12):1982 – 2003.

Kitchenham, B. and Charters, S. (2007). Guidelines for per-

forming systematic literature reviews in software en-

gineering. Technical Report EBSE 2007-001, Keele

University and Durham University Joint Report.

Kitchenham, B., Pearl Brereton, O., Budgen, D., Turner,

M., Bailey, J., and Linkman, S. (2009). Systematic

literature reviews in software engineering - a system-

atic literature review. Inf. Softw. Technol., 51(1):7–15.

Kuper, A. (1995). The Social Science Encyclopedia. Rout-

ledge world reference series. Taylor & Francis.

Lee, J., Baik, S., and Choonhwa Lee, C. (2011). Building

an integrated service management platform for ubiq-

uitous cities. Computer, 44:56–63.

Mostashari, A., Arnold, F., Maurer, M., and Wade, J.

(2011). Citizens as sensors: The cognitive city

paradigm. In Emerging Technologies for a Smarter

World (CEWIT), 2011 8th International Conference &

Expo on, pages 1–5. IEEE.

Nations, U. (2007). World population prospects: The 2006

revision and world urbanization prospects. Technical

report, United Nations, New York.

Sanchez, L., Galache, J., Gutierrez, V., Hernandez, J.,

Bernat, J., Gluhak, A., and Garcia, T. (2011). Smart-

santander: The meeting point between future inter-

net research and experimentation and the smart cities.

In Future Network & Mobile Summit (FutureNetw),

pages 1–8. IEEE.

Shao, C. (2011). An internet of things application with lo-

cation perception based on ims. In Third International

Conference on Multimedia Information Networking

and Security (MINES), pages 163–166.

ınola, R. O. and Travassos, G. H. (2012). To-

wards a framework to characterize ubiquitous soft-

ware projects. Inf. Softw. Technol., 54(7):759–785.

Vega-Barbas, M., Casado-Mansilla, D., Valero, M., Lopez-

de Ipina, D., Bravo, J., and Florez, F. (2012). Smart

spaces and smart objects interoperability architecture

(s3oia). In Innovative Mobile and Internet Services

in Ubiquitous Computing (IMIS), 2012 Sixth Interna-

tional Conference on, pages 725 –730.

Wu, H., He, C., Liao, A., and Peng, S. (2011). A framework

for integrating heterogeneous spatial information and

applications adaptively based on multi-agent and web

service. In Multimedia Information Networking and

Security (MINES), 2011 Third International Confer-

ence on, pages 253 –257.

SmartCitiesArchitectures-ASystematicReview

417