TOWARDS THE DEFINITION OF A FRAMEWORK FOR SERVICE

DEVELOPMENT IN THE AGROFOOD DOMAIN

A Conceptual Model

Donato Barbagallo

, Cinzia Cappiello

, Alberto Coen Porisini

, Pietro Colombo

Marco Comerio

, Flavio De Paoli

, Chiara Francalanci

and Sabrina Sicari

Dipartimento di Elettronica e Informazione, Politecnico di Milano, Milano, Italy

Dipartimento di Scienze Teoriche e Applicate, Universit

a degli Studi dell’Insubria, Varese, Italy

Dipartimento di Informatica Sistemistica e Comunicazione, Universit

a degli Studi di Milano-Bicocca, Milano, Italy

Keywords:

Non-functional Properties, Privacy & Security, Quality of Service, Quality of Data, Service Contract.

Abstract:

This paper describes the work done in the project MoseForAgrofood (Ms4A) whose goal is to take advantage

of mobile technologies to develop augmented Web services for the agrofood domain. The use of personal

devices, such as smart phones and tablets, brings in speciﬁc issues such as the need to consider quality of

protection, quality of services, quality of data at any stage of the service life cycle. In this paper we propose a

conceptual model that aims to support the service development and provisioning addressing the above issues

by managing suitable non-functional properties.

1 INTRODUCTION

Information Technology has already met the food in-

dustry and various applications have been developed

for supporting products traceability and offering ad-

vanced information services to the ﬁnal users. Re-

cently, some contact-less services accessible via mo-

bile phone have been proposed in the agrofood do-

main. They mainly supply users with augmented

information on products (e.g., certiﬁcations, product

provenance and description). An example is the Dy-

namic Wine labels service provided by the Adegga

social network

that allows users to get more infor-

mation and recommendations starting from the QR-

Code on the labels attached to the bottles of wine. A

limitation of such innovative services is that they pro-

vide the same information to any user, often with an

unknown quality level. Services could be more ef-

fective if they would be able to offer costumized re-

sults according to user preferences and context. Cus-

tomization requires the management of user personal

data (e.g., preferences, allergies) and consequently it

introduces the issue of Quality of Protection (QoP)

to ensure a satisfactory level of privacy and security.

Moreover, the fact that data are often collected from

http://www.adegga.com/

different sources introduces the issue of Quality of

Data (QoD) that needs to be addressed in order to

make the user aware of the dependability of the re-

trieved information. Finally, services are provided via

technological channels and it is necessary to consider

the issue of Quality of Service (QoS), to let the users

evaluate if the service fulﬁlls the desired requirements

and constraints.

In this paper, we address these issues and we refer

to QoP, QoD, and QoS with the comprehensive term

of non-functional properties (NFPs). In particular, we

describe the work done in the project MoseForAgro-

food (Ms4A) whose goal is to deliver a framework

for the development of personalized and integrated

services that are accessible via mobile devices with

contact-less interaction style. The paper provides a

comprehensive conceptual model as a foundation for

handling NFPs in a single framework. This approach

considers the quality aspects since the beginning of

the development process and incorporates such issues

in the development framework and related platform.

The adoption of the proposed model supports the au-

tomatization and generalization of discovery, selec-

tion and ﬁltering activities, which is a pre-requisite

to support new types of interaction with minimized

user involvement, as in the case of contact-less inter-

264

Barbagallo D., Cappiello C., Coen Porisini A., Colombo P., Comerio M., De Paoli F., Francalanci C. and Sicari S..

TOWARDS THE DEFINITION OF A FRAMEWORK FOR SERVICE DEVELOPMENT IN THE AGROFOOD DOMAIN - A Conceptual Model.

DOI: 10.5220/0003935102640267

In Proceedings of the 8th International Conference on Web Information Systems and Technologies (WEBIST-2012), pages 264-267

ISBN: 978-989-8565-08-2

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

action with mobile devices. In fact, the model will

drive the design of the MS4A platform that will be

able to (i) handle the contracts (i.e., agreements on

NFPs) deﬁned with the users; (ii) select the proper

services to be provided in order to both satisfy the

quality requirements and enforce the privacy and se-

curity policies speciﬁed in the contracts; (iii) monitor

the NF conditions under which the services are pro-

vided at run time; (iv) enforce proper actions in case

the agreements are violated.

The rest of the paper is organized as follows: Sec-

tion 2 proposes the comprehensive conceptual model.

Section 3 contains a short state of the art of NFPs

management in service-based solutions. Conclusions

and future work end the paper in Section 4.

2 THE CONCEPTUAL MODEL

The model proposed in this paper introduces the con-

cepts needed to drive the development of a platform

that enforces the NFPs discussed so far. Figure 1

shows the model by means of a UML Class diagram.

Each concept has been detailed to give evidence to the

involved elements.

Abstract class Service models the evident concept

of service. Service providers and service users repre-

sent the stakeholders involved in the service domain.

Abstract class ServiceProvider represents a generic

entity that supports the provisioning of services. Class

Ms4APlatform extends ServiceProvider and is aimed

at modeling the provider of the services deﬁned for

the Ms4A Project. Similarly, ExternalProvider mod-

els the provider of additional services required by the

Ms4APlatform to provide its services.

In order to enable the personalization of Ms4A

services, users have to provide information related to

their identity, life style and interests. The consumer of

the services is modeled by the ServiceUser class and

the associated data by the abstract class UserData.

The classes Identiﬁable, Sensitive, and Generic. in-

troduce speciﬁc categories of UserData. Identiﬁable

deals with data referred to the identity of a user (such

as family name). Sensitive deals with attributes re-

lated to the user’s private life such as health conditions

and religious creed. Generic deﬁnes general informa-

tion that do not belong to previous classes.

Considering Ms4A platform, class Frontend mod-

els the client side of a Service, while class Backend

the server side. The client side manages (i) the mobile

device used by the user; (ii) the part of the service ap-

plication that is executed on this device; and (iii) the

user’s proﬁle data that need to be forwarded to the

back-end for the execution of the requested service.

The abstract class FrontendData models generic

data that are handled by the frontend. Proﬁle extends

FrontendData by deﬁning data types required at ser-

vice execution time. Particularly, a proﬁle includes

1) identiﬁable, sensitive and generic data concerning

personal characteristics of the user, 2) the user’s pref-

erences on the use of the service, 3) the acknowledged

contract, and 4) the encryption keys that are used to

address protection issues. Moreover, a service may

provide various functionalities, but the user might be

interested only in part of them. Therefore, class Pref-

erenceOnService is used to customize the service use

on the basis of the users’ preferences. In particular,

customization regards both the preferred usability set-

tings and the selection of the basic functionalities.

Contract speciﬁes the NFPs the service provider

considers to provide the service and the actions that

will be executed on the users data. The user must con-

sent to the contract in order to use the service. Class

ConsentData models the acceptance of the agreement

that is established with the service provider, that in-

cludes also the consent to handle his/her data.

Due to the sensitive data that can be exchanged

between user and service provider, encryption keys

are used to protect the transmitted information. Class

Key models the encryption keys that are used 1) to ex-

change messages and 2) to encrypt the sensitive data

that refer to instances of identiﬁable data.

Some basic functionalities need to be supported

by the service client side. These functionalities in-

clude: 1) to access data stored on the mobile device,

2) to handle the interaction with the server-side via

message passing, 3) to encrypt data, and 4) to han-

dle visualization and rendering activities. Class Fron-

tendActivity models the basic functionalities of the

client-side service. Advanced client-side activities are

deﬁned on top of these basic functionalities. For in-

stance, ConsentAcquisition models the acceptance of

the agreement that enables the user to use the service,

and the provider to handle users data according to

the policy speciﬁed in the contract. ProﬁleDeﬁnition

models the acquisition of the user’s data required for

the execution of the service, and his/her preferences

with respect to the enjoyment of the service function-

ality. This activity generates data of type Proﬁle.

The Backend side collects all the services used to

provide the desired functionalities, including the con-

trol activities that are modeled by class BackendActiv-

ity. A control activity introduces the actions that are

taken to monitor, correct and report of violations of

conditions during and after the execution of a service.

Class Obligation models the actions that the proces-

sor guarantees to perform, after the data have been

processed in case particular conditions are veriﬁed.

TOWARDSTHEDEFINITIONOFAFRAMEWORKFORSERVICEDEVELOPMENTINTHEAGROFOOD

DOMAIN-AConceptualModel

265

Service

Frontend Backend

Service

User

Data

Identifiable

Generic

Sensitive

Message

Obligation

Backend

Activity

Supporting

Data

Dependency

Type

Metric

NonFunctional

Properties

QualityOf

Service

QualityOf

Protection

Cost

QualityOf

Data

Functional

Dependency

Control

Dependency

Data

Dependency

measuredBy

Contract

Generic

Personal

Purpose

required

responsableOf

Applicability

Frontend

Activity

Frontend

Data

Consent

Acquisition

Profile

Definition

ProfileKey

Consent

Data

Service

Provider

External

Provider

Ms4A

Platform

Preference

On Service

Backend

Data

Encryption

Key

Historical

Interaction

Figure 1: The main concepts of the conceptual model.

The abstract class BackendData models generic

data used during the execution of a service by the ser-

vice provider. This generic data type is reﬁned into

some concrete classes. Class HistoricalInteraction

models information about the execution of the ser-

vice. Class SupportingData models data owned by a

service provider that are required to support the exe-

cution of a service. Finally, class EncryptionKey mod-

els the encryption keys that are used to satisfy protec-

tion requirements included in the contract.

Both frontend and backend services may depend

on each other. We can identify three types of depen-

dency: i) Functional Dependency models the depen-

dencies in the execution of services; ii) Data Depen-

dency models situations in which there is a data ﬂow

between services and therefore a service relies on data

generated by another service; iii) Control Dependency

models the situation in which a service is the supervi-

sor of another one and has to check the satisfaction of

NF and functional constraints. This is the case of the

Ms4A platform that governs and monitors the execu-

tion of the external services.

Class Contract models the agreement that is es-

tablished between the platform, which is responsi-

ble for providing the services, and the user(s). The

agreement is said Generic when it is offered to a

wide group of potential users, and Personal when it

is signed between a provider and a single user that

also speciﬁes preferences on the provisioning. The

agreement speciﬁes the NFPs the Ms4A platform un-

dertakes to provide the service. Abstract class Non-

FunctionalProperty models a generic NFP. Concrete

extensions of NonFunctionalProperty are introduced

to model properties belonging to different NF cat-

egories. More speciﬁcally, class QualityOfService

models provisioning requirements that concern per-

formance and use of resources. QualityOfProtection

models security and privacy requirements. Quality-

OfData models the state of completeness, validity,

consistency, timeliness and accuracy of the data man-

aged by the service. Finally, Cost models the ﬁnancial

terms and conditions (e.g., price, insurance and com-

pensation agreements) associated with the service.

All the NF dimensions are evaluated by means of met-

rics. Class Metrics has been introduced to specify the

assessment algorithms used to measure the NFPs.

Besides all these aspects, the contract speciﬁes the

general service aims and functionalities. Class Pur-

pose models the service aims including the descrip-

tion of 1) the service functionalities, 2) the user’s data

that are required during the service execution and 3)

the activities that will be performed on such data.

Moreover, if a service includes an Obligation, Pur-

pose will describe it together with the conditions un-

der which the functionality is executed. Finally, Pur-

pose speciﬁes the service provider that is in charge to

execute the service. Finally, a service can be thought

for a speciﬁc category of users. Class Applicability

models the conditions under which the contract can

be offered to a speciﬁc class of users.

3 RELATED WORK

In the literature, the mutual understanding between

providers and consumers is typically established by

specifying policies (Bajaj et al., 2006) and Service

Level Agreements (Keller and Ludwig, 2003). Poli-

cies and SLAs are commonly referred to as Service

Contracts (Comerio et al., 2009). Besides a func-

tional description of the service, a service contract in-

cludes the speciﬁcation of contractual terms that are

constraints on NFPs that can include QoS, QoD, and

WEBIST2012-8thInternationalConferenceonWebInformationSystemsandTechnologies

266

QoP aspects. However, QoS, QoD, and QoP are not

typically considered together. The literature provides

a variety of QoS models and QoS-based Web service

discovery and selection approaches. Proposals such

as (Wang et al., 2006) tackle the discovery by deﬁning

speciﬁc sets of quality dimensions to consider. Due to

the fact that the authenticity of the advertised QoS in-

formation may be questionable, some papers in the

literature (e.g., (Xu et al., 2007)) propose a model of

reputation-based Web services discovery. The QoD is

another aspect that must be checked during the service

selection. Data quality literature provides a thorough

classiﬁcation of data quality dimensions, e.g., (Batini

and Scannapieco, 2006). Several architectures for the

data quality management have also been proposed. In

particular, in (Scannapieco et al., 2004), a scenario

similar to the one proposed in this paper is consid-

ered: user needs are satisﬁed by integrating data gath-

ered from heterogeneous sources.

Finally, users want to subscribe a service also tak-

ing into account the satisfaction of their QoP needs.

As regard the research efforts in the QoP ﬁeld, re-

quirement engineering methodologies, such as Kaos

(Lamsweerde et al., 2000) and Tropos (Liu et al.,

2002), can be used for the analysis and speciﬁca-

tion of security & privacy requirements. Moreover,

several techniques such as anonymizing mechanisms

based on data suppression or randomization (Mieliki-

nen, 2004; Narayanan and Shmatikov, 2005) have

been proposed to protect private data from unautho-

rized accesses. All these techniques do not require

the deﬁnition of any privacy policies; rather they can

be used as building blocks for realizing them. Finally,

papers such as (Coen-Porisini et al., 2010) presents a

conceptual model aiming at supporting the speciﬁca-

tion of privacy policies. In the present work, we adopt

a similar approach to (Coen-Porisini et al., 2010) to

deﬁne our conceptual model but we want to capture

not only privacy but different NFPs, such as QoS and

QoD.

4 CONCLUSIONS AND FUTURE

WORK

This paper proposes a comprehensive conceptual

model to handle multiple NFPs in order to support

design, selection and composition of added-value ser-

vices in the Ms4A scenario. An innovative feature of

our model is the capability to capture several types

of NFPs, i.e., QoS, QoD, QoP, and the relationships

among them. Future work focuses on the investiga-

tion of the requirements for the service design and on

the development of a set of tools for the management

of service contracts and for the description of NFPs

associated with the services and the platform.

ACKNOWLEDGEMENTS

This work is partially funded by the regional (Lom-

bardy) MoseForAgroFood project.

REFERENCES

Bajaj, S., Box, D., and et Al., D. C. (2006). Web

Service Policy 1.2 - Framework. Available

at: http://www.w3.org/Submission/2006/SUBM-WS-

Policy-20060425/.

Batini, C. and Scannapieco, M. (2006). Data Quality: Con-

cepts, Methodologies and Techniques. Data-Centric

Systems and Applications. Springer.

Coen-Porisini, A., Colombo, P., and Sicari, S. (2010). Deal-

ing with Anonymity in Wireless Sensor Networks. In

Proc. of ACM SAC.

Comerio, M., Truong, H.-L., De Paoli, F., and Dustdar, S.

(2009). Evaluating contract compatibility for service

composition in the seco2 framework. In Proc. of Inter-

national Conference on Service Oriented Computing

(ICSOC), Stockholm, Sweden.

Keller, A. and Ludwig, H. (2003). The WSLA framework:

Specifying and monitoring service level agreements

for web services. Journal of Network and Systems

Management, 11(1).

Lamsweerde, A., Letier, and E.Handling (2000). Obstacles

in Goal-Oriented Requirement Enginnering. IEEE

TSE, 26:978–1005.

Liu, L., Yu, E., and Mylopoulos, J. (2002). Analyzing Se-

curity Requirements as Relationships among Strategic

Actors. In SREIS 2002).

Mielikinen, T. (2004). Privacy Problems with Anonymized

Transaction Databases. In Int. Conf. on Discovery Sci-

ence (DS 2004), Vol 3245 of Lecture Notes, Springer.

Narayanan, A. and Shmatikov, V. (2005). Obfuscated

databases and group privacy. In CCS, pages 102–111.

Scannapieco, M., Virgillito, A., Marchetti, C., Mecella, M.,

and Baldoni, R. (2004). The architecture: a platform

for exchanging and improving data quality in cooper-

ative information systems. Inf. Syst., 29(7):551–582.

Wang, X., Vitvar, T., Kerrigan, M., and Toma, I. (2006). A

qos-aware selection model for semantic web services.

In ICSOC, pages 390–401.

Xu, Z., Martin, P., Powley, W., and Zulkernine, F. (2007).

Reputation-enhanced qos-based web services discov-

ery. In ICWS 2007, pages 249 –256.

TOWARDSTHEDEFINITIONOFAFRAMEWORKFORSERVICEDEVELOPMENTINTHEAGROFOOD

DOMAIN-AConceptualModel

267