Using DEMO-based SLAs for Improving City Council Services

Carlos Mendes

, Mário Almeida

, Nuno Salvador

and Miguel Mira da Silva

Instituto Superior Técnico, Avenida Rovisco Pais, Lisboa, Portugal

Câmara Municipal de Pombal, Pombal, Portugal

Keywords: Service Catalogue, Service Level Agreements, Demo, Enterprise Ontology.

Abstract: According to data from the Portuguese Association of City Councils (DGAL – Portuguese abbreviation) the

308 city councils in Portugal employ about 135 000 people and spend about 3.8 billion euros a year

(Ministry of Finance and Public Administration, 2010). In this paper we describe the implementation of our

most recent proposal to specify the services quality in the city council of Pombal (CMP) in Portugal. This

proposal is a new version of the DEMO-based SLAs with a more complex structure of Service Level

Agreement (SLA) attributes. This proposal is based on the Enterprise Ontology theory and identifies

services that do not create new original results, therefore services with great potential to be automated. We

evaluated the new proposal’s version by collecting feedback from CMP employees and customers.

Moreover, we found some possible improvements that could save millions of euros to the Portuguese state.

1 INTRODUCTION

In previous research we focused on closing the gap

between customers’ expectations and the perceived

service (Parasuraman et al., 1985) by formally

specifying the customers’ expectations into Service

Level Agreements (SLAs) (Mendes et al., 2012)

(Mendes et al., 2011) (Mendes and Mira da Silva,

2012), using as a foundation the DEMO

methodology and respective Enterprise Ontology

theory (Dietz, 2006).

DEMO (Design & Engineering Methodology for

Organizations) is a methodology for modelling,

(re)designing and (re)engineering organizations and

networks of organizations. The theory that underlies

this methodology is called Enterprise Ontology (EO)

that by itself is based on the speech act theory. We

have chosen EO because this theory can help us

expand the expressiveness of the service

descriptions and, consequently, allow a better

alignment between expectations and perceptions

(Mendes and Mira da Silva, 2012).

In this paper we propose a new version of the

DEMO-based Service Level Agreement (SLA)

(Mendes and Mira da Silva, 2012) that has a new

structure with more attributes and results from a

deeper state of art analysis. We applied this proposal

in the Portuguese city council of Pombal (CMP -

Portuguese abbreviation), i.e. we specified the

service catalogue of CMP including the services

CMP provides to Pombal residents and the SLAs

they use to comply with the residents’ expectations.

To evaluate the new SLA version we collected

feedback from 23 employees of CMP and 7

customers. The majority of the proposal attributes

(14 in 16) was classified as important since they

received a minimum score of 7.6 in 10 while the

remaining two received an score of 5.9. Besides

these findings, as our proposal identifies services

that do not create new original results, we also

identified services with great potential to be

automated. We present an example of a possible

improvement that could save millions of euros to the

Portuguese state.

Our study was conducted using the Design

Science Research Methodology (DSRM) that aims

at creating and evaluating IT artefacts intended to

solve identified organizational problems (Hevner et

al., 2004). These artefacts include constructs

(vocabulary and symbols), models (abstractions and

representations), methods (algorithms and practices)

and

instantiations (implemented and prototype

systems). This research method comprises the

following phases (Peffers, 2008): problem

identification, objectives definition, design and

development, demonstration, evaluation and

communication.

The paper is structured as follows. We will start

431

Mendes C., Almeida M., Salvador N. and Mira da Silva M..

Using DEMO-based SLAs for Improving City Council Services.

DOI: 10.5220/0004110404310440

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (SSEO-2012), pages 431-440

ISBN: 978-989-8565-30-3

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

by providing a brief overview of the literature on the

problem area (Section 2). In Section 3, we introduce

the theoretical background of this research, the

Enterprise Ontology theory. Afterwards, we

introduce the new version of the DEMO-based

solution to specify the services quality (Section 4).

In Section 5, we describe an experiment at CMP. In

Section 6, we explain the evaluation process, which

uses data from the experiment, and specify the

lessons learned. Finally, we present our conclusions

(Section 7).

This section corresponds to the problem

identification and motivation phase of DSRM. It

also corresponds to the objectives definition phase.

2 RELATED WORK

This section describes the current solutions for

specifying services quality and explains why these

solutions do not solve the gaps problem

(Parasuraman, Zeithaml, & Berry, 1985).

We analysed several solutions to specify the

Service Quality: Service Level Management best

practices, web services based solutions and the

Generic Service Specification Framework (GSSF).

In spite of the different backgrounds, all contributed

to the service quality specification. The first solution

is proposed by many best practices frameworks,

such as ITIL (Office of Government Commerce,

2007) or CMMI (CMMI for Services, Version 1.3,

2010), the second represents the solutions focused

on web services and the third is an Enterprise

Ontology-based approach (even though the main

goal of the GSSF was to specify the services and not

the service quality itself, this framework also

contributed to the problem area).

Service Level Management is one of the key

processes by which organizations manage their

services, because it acts as the interface between the

customer and the provider. At its most basic level,

Service Level Management is involved in the

following activities: define, agree, record and

manage levels of service. There are a number of key

elements required to ensure that services are fit for

purpose and use, and remain so throughout their

lifetime: service level requirements, targets and

agreements (Office of Government Commerce,

2007).

Basically, to understand the Service Level

Requirements (SLR) means that the customers’

needs and wants are understood, i.e. an SLR is a

customer requirement for an aspect of a service.

SLRs are based on business objectives and are used

to negotiate Service Level Targets (SLT) which are

commitments documented in Service Level

Agreements (SLAs). SLTs are based on SLRs and

are needed to ensure that the service is fit for

purpose. SLTs should be SMART: specific,

measurable, attainable, realistic and timely. Finally,

SLA is an agreement between a provider and a

customer that describes the service; it documents the

SLTs and specifies the responsibilities of the

provider and customer. Over the years it has also

been the chosen concept to specify services quality

(Office of Government Commerce, 2007).

Regarding Service Level Management solutions,

current approaches have two main flaws. First, they

lack a strong conceptual foundation because they

were derived from best practices of several years of

implementations - not from a well-founded theory.

Consequently, the inexistence of a theory may cause

incoherencies among those solutions (second flaw).

Service Level Management solutions are process-

driven and not service-driven. These solutions are

designed to work individually as processes but the

interactions between these processes (such as

Request Fulfilment, Service Level Management and

Incident Management) are usually unclear. For

instance, the connection between an incident and an

SLA is neither clearly explained in ITIL nor in

CMMI.

There are some solutions to specify the services

quality that originated in the web services

community. In (Sahai, Durante, & Machiraju, 2002)

the authors show how to use Web Service

Description Language (WSDL) and Web Service

Flow Language (WSFL) to specify SLAs. However,

this work suffers from the web vision tunnel as it is

focused on the web services and does not try to

specify business services. For instance, the

specifications do not include penalties or prices. The

researches in (Tosic, Patel, & Pagurek, 2002),

(Dobson, 2004) and (Frolund & Koistinen, 1998)

have the same bottleneck. Despite this trend in the

web service community, there are some recent

researches that try to overcome the mentioned web

service tunnel vision. In (Keller & Ludwig, 2003) a

novel framework for specifying and monitoring

SLAs for Web Services is introduced: the Web

Service Level Agreement (WSLA) framework. This

framework is applicable to any inter-domain

management scenario such as business process and

service management or the management of

networks, systems and applications in general. In

(Andrieux, et al., 2007) and (Liu, Ngu, & Zeng,

2004) business criteria is also included in SLAs.

These three solutions represent a new movement in

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

432

the web service community; however, none is based

on a strong conceptual foundation.

Another contribution to the gaps problem is the

Generic Service Specification Framework (GSFF)

(Terlouw & Albani, 2011), which is based on the

following generic service definition (Albani,

Terlouw, Hardjosumarto, & Dietz, 2009): a service

is a universal pattern of coordination and

production acts, performed by the executor of a

transaction for the benefit of its initiator, in the

order stated in the standard pattern of a transaction.

We adopted this service definition in our

research since this definition is the only one that, as

our research, uses DEMO as a conceptual

foundation.

The GSSF defines four main areas of concern for

each service: the service executor, the service

production, the service coordination and the service

contract option. The first one defines who the

provider of the service is. The second focuses on the

production act to be performed by the executor. The

third gives the consumer all the information required

for conducting a successful communication with the

provider. And finally, the service contract option

specifies one or several contract options from which

service consumers can choose.

Even though the quality aspects are very basic,

the Generic Service Specification Framework

represents a large contribution to the service

specification research area. However, the level of

service quality specification is not always sufficient,

because sometimes customers and providers have

different expectations due to a lack of specification

(Terlouw & Albani, 2011).

3 THEORETICAL FOUNDATION

This section briefly describes the Enterprise

Ontology theory (the theory that supports our

proposal).

Enterprise Ontology (Dietz, 2006) is based on

four axioms – operation, transaction, composition

and distinction – and the organization theorem. The

operation axiom states that the operation of an

enterprise is constituted by the activities of actor

roles that are elementary chunks of authority and

responsibility, fulfilled by subjects. In doing so,

these subjects perform two kinds of acts:

production acts and coordination acts. These acts

have definite results: production facts and

coordination facts, respectively. By performing

production acts (P-acts) the subjects contribute to

bringing about the goods and/or services that are

delivered to the environment of the enterprise. By

performing coordination acts (C-acts) subjects enter

into, and comply with, commitments towards each

other regarding the performance of production acts.

The transaction axiom states that coordination

acts are performed as steps in universal patterns.

These patterns, also called transactions, always

involve two actor roles (initiator and executer) and

are aimed at achieving a particular result. A

transaction develops in three phases: the order phase

(O-phase), the execution phase (E-phase), and the

result phase (R-phase). In the O-phase the two actors

agree on the expected result of the transaction; in the

E-phase the executer executes the production act

needed to create the expected result; and in the R-

phase the two actors discuss if the transaction result

is equal to the expected result.

The composition axiom establishes the

relationships between transactions. This axiom states

that every transaction is either a) enclosed in another

transaction, b) is a customer transaction of another

transaction, or c) is a self-activation transaction. The

latter case refers to transactions that give rise to

further transactions of the same type.

The distinction axiom states that there are three

distinct human abilities playing a role in the

operation of actors, called performa, informa, and

forma. An ontological act (performa) is an act in

which new original things are brought about.

Deciding and judging are typical ontological

production acts. Regarding the coordination between

people, typical ontological acts are requesting and

promising. An infological production act is an act in

which one is not concerned about the form but,

instead, about the content of the information.

Typical infological acts are inquiring, calculating,

and reasoning. Regarding the coordination between

people, formulating thoughts (in written or spoken

sentences) and interpreting perceived (through

listening or reading) sentences are typical infological

coordination acts. Acts like copying, storing, and

transmitting data are typical datalogical acts, while

speaking, listening, writing, and reading are typical

datalogical coordination acts.

4 PROPOSAL

This section corresponds to the design and

development step of DSRM. Our proposal is

composed by the following steps:

1. Identify the services;

2. Specify the executor, production and

coordination of the services;

UsingDEMO-basedSLAsforImprovingCityCouncilServices

433

3. Specify the SLAs for each identified

transaction/service.

The first step is to identify the services of the

provider, using for that purpose a process based on

the methodology proposed in (Dietz, 2006). This

process is composed by six steps: Enterprise

Description, Performa-Informa-Forma Analysis,

Coordination-Actors-Production Analysis,

Transaction Pattern Synthesis, Result Structure

Analysis and Actor Transaction Diagram/Service

Identification.

The second step of the proposal is to apply part

of the GSSF (Terlouw & Albani, 2011): service

executor, service production and service

coordination. The service executor area defines who

the provider of the service is; the service production

area focuses on the production act to be performed

by the executor; and the service coordination area

gives the consumer all the information required for

conducting a successful communication with the

provider.

The fourth area, the contract options, is replaced

by our definition of SLA presented in Figure . That

means for each identified service, one should specify

the list of associated SLAs using our SLA definition

(step three).

Figure 1: Structure and attributes of the DEMO-based

SLA.

Our SLA proposal considers four main areas of

concern with their respective attributes that we will

now explain. The first section is called Generic

SLA Information and it defines the name of the

SLA (SLA Name) and the SLA purpose (SLA

Description). Additionally, the Generic SLA

Information describes who owns this SLA (SLA

Owner), it also provides a contact of this person

(SLA Owner Contact Information) and, finally, this

section defines the name of the service that the SLA

applies to (SLA Service).

The second considered section contains

information concerning the dates of the SLA and is

called Temporal SLA Information. In this section

the date on which the SLA was established (SLA

Creation Date) is defined as well as the time interval

on which the SLA is valid (SLA Validity Period), the

information related to the SLA modification dates by

the customer (SLA Version Control Information) and

the information concerning the SLA review dates

performed by an entity related to the service

provider (SLA Review Period Information).

Next, we define a section called Responsibility

SLA Information that regards the information

about the responsibilities of each actor in the

execution of this SLA. In this section two attributes

are specified concerning the obligations and duties

of the customer (SLA Customer Responsibilities) and

the service provider (SLA Provider Responsibilities).

Finally, the last section is called Specific SLA

Information and for each type of SLA (SLA Type) it

specifies six different types of targets (SLA Targets),

which can give rise to actions if they are not fulfilled

(SLA Penalties), but if they are fulfilled, this should

be rewarded (SLA Bonuses). Each type of SLA is

also associated to a price (SLA Price).

Several of these attributes can be gained from

DEMO diagrams, such as, for example, the SLA

Owner that can be identified by the Actor

Transaction Diagram (ATD) or the SLA Penalties

and the SLA Bonuses that can be gained from the

Action Model (AM).

Thus, this research intends to reduce the gaps by

formally specifying the SLAs, using EO theory as a

foundation.

5 DEMONSTRATION

This section corresponds to the demonstration phase

of DSRM. We evaluated the proposal using an

experiment in order to validate its applicability. The

demonstration occurred in a Portuguese city council

named Pombal (CMP). Pombal is located in Leiria

District and is composed of 17 parishes. It has a total

area of 626.1 km² and a total population of 58,617

inhabitants. The population of the city of Pombal is

about 18,000 inhabitants. CMP employs a total of

389 people with 203 men and 189 women and has

five major departments divided into divisions, units

and sections. In 2010 CMP spent a total of 20 553

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

434

200 € from which 7 542 250 € in human resources

(Pombal City Council, 2011).

In order to identify the services (step one of the

proposal) we interviewed individually 17 employees

from CMP. With the purpose of having an overall

perspective of the entire CMP we have selected

employees from all the departments. During the

interviews participants were asked to describe the

activities performed by CMP. The interviews were

recorded and transcribed as well as checked and

discussed by two interviewers each ensuring

unbiased findings and avoiding misinterpretation as

specified in (Kvale, 2007).

The interviews allowed us to develop an

enterprise description of CMP that was used as input

for the service identification step (proposal first

step). We do not fully describe the six sub steps of

the service identification step due to space

limitation, nevertheless these sub steps are based on

DEMO (Dietz, 2006) and are described in previous

publications (Mendes, Ferreira, & Mira da Silva,

2011) (Mendes, Ferreira, & Mira da Silva, 2012).

The result of this first step is called the Actor

Transaction Diagram (ATD) and we decided to

show only the ATD with information regarding the

Human Resources and IT Service due to space

limitations (Figure ).

Figure 2: ATD with HR and IT.

In the ATD, a transaction/service is represented

using a diamond in a disk that contains the

respective combination of C-acts and a P-act. Each

transaction is connected to two boxes, representing

the initiator and executor actor roles. The initiator is

connected to the transaction symbol using a solid

line, while the executor is connected to the

transaction using a solid line ending in a black

square. The grey boxes refer to composite actor

roles, i.e. elements whose exact structure is not

known. All the environmental elements, i.e.

elements outside the organization that we are

studying, are represented with grey boxes for that

reason. This also means that we can represent the

studied organization with a grey box when referring

to the kernel of the organization, which can be

further specified by using elementary actor roles

represented by white boxes.

We identified 173 services of which 145 are

ontological, 17 are infological and 11 are datalogical

(see Section 3). These services correspond to all

services provided by the five major departments that

constitute the City Council. Figure 2 illustrates

seven major Composite Actor Roles (President, IT

Kernel, Employee, HR Kernel, Special Mobility,

Candidate and Doctor) on a total of nearly 50 across

the entire City Council.

Table 1: Service “Application Development”

Specification.

Service Specification - Application Development (T30)

Service Executor

Actor Role Developer (A02)

Contact

Information

General Email: suporte@cm-pombal.pt

Email: xxxxx

Phone: xxxxx

Service Production

Production Act

Application Development is the act of designing

and developing new applications or features

required by other units of the Municipal Council

or by the IT Division.

Production

Information

Used

The object classes used in this services are:

Application, Handbook, Employee, Course

Production Fact Application App has been developed

Production Kind

The production kind of this transaction is:

Ontological

Production

World

Semantics

Application, Handbook, Employee and Course

(See Figure )

Preconditions Hardware Installation (T28)

Postconditions ND

Service Coordination

Coordination

Acts

See

Figure 4: PSD for Service ‘Application

Development’

Figure

Coordination

Kind

This service is a Human Service

Protocol

1. Contact the IT Division;

2. Specifies the Application requirements;

3. Wait for the end of development;

4. Test Application;

5. Application available to the target

audience.

Location Email, Phone

UsingDEMO-basedSLAsforImprovingCityCouncilServices

435

The IT division operates and maintains the

computer equipment, develops new tools, supports

their applications, and conducts courses to enhance

learning of the new features. We identified 13

services provided by this division: Network

Configuration (T26), Hardware Installation (T28),

Hardware Uninstallation (T29), Hardware

Substitution (T39), Application Development (T30),

Incident Resolution (T31), Database Management

(T32), Software Installation (T35), Software

Uninstallation (T36), Backup Realization (T40),

Handbook Definition (T41), Training (T42) and

Business Intelligence Study Realization (T43).

In order to proceed to the second step of the

proposal (specify the executor, production and

coordination of the services), first we had to model

the Process Model, the Action Model, and the State

Model of Pombal City Council, since some aspects

of the GSSF (used in the second step of the

proposal) depend on these models. We do not

present all of these models due to space limitations.

We applied the Generic Service Specification

Framework (GSSF) to specify the services of Figure

(second step of the proposal). An example of this

specification is illustrated in Table . This table

describes the attributes of the service Application

Development that implements the T30 transaction

and is provided by the IT Division.

The service specified above is called Application

Development and is carried out by a Developer

whose contact is available in the Service Executor

section. In the Service Production area, it is

specified that a new application is produced, which

makes this an ontological service and, based on the

State Model (Figure ), we found the information

classes used in this service: Application, Handbook,

Employee and Course.

Figure 3: State Model for Service 'Application

Development’.

The ontological coexistence rules between these

classes are the following: an Application may have a

Handbook, several Courses and Employees using it.

Additionally, a Course concerns an Application and

it is taught in a certain period to a number of

Employees.

Before this service execution, a precondition

must be guaranteed, since the application

development must be preceded by the hardware

installation (T28). Regarding the post conditions

there are none associated with the Application

Development service. Concerning the Coordination

area, the coordination acts involved in this service

are illustrated by the Process Structure Diagram

(PSD) (Figure 4).

In this diagram we see that the employee makes

a request for a new application or feature and this

request is handled by the developer that, when the

development of the new application ends, starts two

new transactions: write the user manual (Handbook

Definition – T41) and schedule a training on this

new application (Training – T42). In addition, in the

Coordination area the procedure or protocol to

successfully contact the service provider is specified

as well as the location of the service that, in this

case, is by email or phone.

In the third step of the proposal we specified the

SLAs associated to the CMP services using our SLA

proposal (Figure 1). Table illustrates an example of

this specification for the service Application

Development.

Figure 4: PSD for Service ‘Application Development’.

This SLA concerns the development of an

application named WebDoc2.0 and the SLA Owner

is the IT Division Chief. Multiple contacts of the IT

Division Chief are specified in order to be contacted

by the service customer at any time. This SLA was

made on the first day of January of 2011, was valid

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

436

until 31 of November of 2011 and was changed in

July 15. To fulfil this SLA, the service provider

Table 2: SLA “WebDoc 2.0 Development” Specification.

needs to complete all the points specified in the SLA

Provider Responsibilities (completion of ERP

integration, improvement of communication

between citizens and putting into production all

features developed).

This SLA has three types that depend on the date

of completion of the development. Penalties and

bonuses are translated into career points that

influence the career development.

On the one hand, in case of SLA Type

“Overcome goal”, the SLA Owner wins five career

points and can evolve in his career.

On the other hand, if the service provider does

not meet the deadline the SLA Type “Non

Fulfilment Goal” applies and the SLA Owner only

wins one career point and can be fired with probable

cause.

Note that this SLA has no price defined because

this is an internal service to the CMP and no

chargeback is made among the CMP departments.

6 EVALUATION

This section corresponds to the evaluation phase of

DSRM and in order to explain the evaluation we use

the framework proposed in (Pries-Heje, Baskerville,

& Venable, 2004). This framework identifies what is

actually evaluated, how it is evaluated and when the

evaluation takes place.

Table illustrates the answers to the three main

questions that this framework proposes to answer:

 What is Actually evaluated? The artifact

evaluated is the proposed set of steps of

Section 4 (a design process) and the results

of applying these steps to the CMP

(Services and SLAs; a design product);

 How is it evaluated? We used CMP

employees’ feedback to evaluate the

DEMO-based SLA structure and the CMP

services and SLAs. This represents a

naturalistic evaluation since it was

conducted using a real artifact in a real

organization facing real problems;

 When was it evaluated? It was evaluated

ex post (after the design artifact was

developed).

P summarizes the essential characteristics of the

evaluation Process, while C indicates the evaluation

Criteria.

The evaluation was naturalistic since we applied

our proposal in a real organization with real data.

The evaluation was ex post since it occurred after

Service Level Agreement Specification

Generic SLA Information

SLA Name

WebDoc 2.0 Development

SLA Description

SLA concerning the development of a

tool for document management.

SLA Owner

Nome: Nuno Salvador

Category: IT Division Chief

Organic Unit: IT Division

SLA Owner

Contact

Information

Email: xxx

Phone: xxx

SLA Service Application Development (T30)

Temporal SLA Information

SLA Creation

Date

January 1, 2011

SLA Validity

Period

Until November 31, 2011

SLA Version

Control

Information

July 15, 2011

SLA Review

Period

Information

Responsibility SLA Information

SLA Customer

Responsibilities

SLA Provider

Responsibilities

1 - Finish ERP Integration;

2 - Improve communication between the

citizens;

3 - Place in operation all the features

developed.

Specific SLA Information

SLA Type

Overcome Goal

SLA Targets

Performance Until November 15, 2011

SLA Penalties NA

SLA Bonuses

Evaluation Score: 5

Career: Allows career evolution

SLA Price 0 €

SLA Type

Fulfillment Goal

SLA Targets

Performance Until November 31, 2011

SLA Penalties NA

SLA Bonuses Evaluation Score: 3

SLA Price 0 €

SLA Type

Non Fulfillment Goal

SLA Targets

Performance After November 31, 2011

SLA Penalties

Evaluation Score: 1

Career: Can be fired with probable cause

SLA Bonuses NA

SLA Price 0 €

UsingDEMO-basedSLAsforImprovingCityCouncilServices

437

Table 3: Evaluation strategy.

the demonstration in the CMP. We evaluated both

the proposed set of steps of Section 4 (including the

DEMO-based SLA structure) and the results of

applying these steps to the CMP (Services and SLAs

specification). In order to evaluate the DEMO-based

SLA structure (see Figure ) and the Services and

SLAs specification from CMP we collected

feedback from 23 CMP employees and 7 CMP

customers. They were arbitrarily chosen and were

asked to classify the attributes of the SLA proposal

from 1 to 10 according to the importance (being 1

irrelevant and 10 essential). Figure 5 illustrates the

average and standard deviation per attribute. As can

be seen there was little variation in the answers of

the interviewees in most attributes.

Figure 5: Rating of proposal attributes.

The first 14 attributes had high average scores

(from 7.60 to 8.40) and the remaining two (Bonuses

and Price) had lower classifications (5.90 and 6.85).

These results indicate that the majority of the

proposed attributes (14 in 16) were classified as

important since they scored a minimum of 7.60 in 10

possible points. The remaining two attributes

(Bonuses and Price) scored 5.90 and 6.85 revealing

that they were classified as less important when

comparing to the first 14. These results can be

explained by the fact that there are no chargeback

among the departments of the 30 inquired persons.

Therefore, they value more the attributes that

describe the service quality than the ones that

capture the costs.

Besides this validation, we also validated the

service catalogue from the IT Division with the IT

chief that agreed with all 13 identified services. The

same validation method is being used for other units

of the City Council hoping to have the same

acceptance that in the IT Division.

The definition of a city council service catalogue

has great potential, because the services that this

type of organizations provides to the citizens are

similar. For instance, in Portugal there are 308 city

councils and in theory they all have the same

purpose. Having identified the service catalogue of

one city council we can validate if it is applicable to

other city councils and eventually find some services

that could be provided in cooperation.

Knowing the services and the type of services

provided allows one to understand how the service

provider can improve his performance. For instance,

the service Collect Water Usage – T155 is a

datalogical transaction since it neither involves the

creation of new original facts (ontological) nor

information processing (infological). Hence, it has

large potential to be optimized, since technology can

be used to reduce the effort needed to execute

datalogical acts (Dietz, 2006).

CMP has four employees dedicated to collect

water usage. Assuming that CMP spends 77 555 € a

year with these four employees ((7 542 250 € / 389

employees) * 4 employees) and the other 307 city

councils in Portugal do not have this service

automated and use similar resources on it, then this

would represent an expense of 23 809 468 € a year

in a service that has great potential to be automated.

This value is estimated using only the costs of the

employees’ wages so if we add the supporting costs

(IT support, HR support, etc.) then the estimated

value would certainly be higher.

This kind of analysis has special value because

of the current situation Portugal is in since it points

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

438

some solutions to a number of current challenges

imposed by the Troika memorandum.

7 CONCLUSIONS

There are several solutions that contributed to

closing the gaps, but none solved the problem

completely. Some lacked detail in specifying the

services’ quality (like the Generic Service

Specification Framework), others were not based on

a strong conceptual foundation (such as ITIL,

CMMI or WSLA) and the majority of the web

services based solutions suffer from the web service

tunnel vision.

In this paper we apply our most recent proposal

to specific services in a Portuguese city council. This

proposal is based on the Enterprise Ontology theory

and identifies services that do not create new

original results, therefore services with great

potential to be automated. In this particular city

council we found services that if automated could

save the Portuguese state millions of euros.

Our new SLA proposal revealed to be more

flexible and mature when comparing to the previous

one. The employees’ and customers’ feedback

revealed that in the CMP context 14 in 16 of the

proposal attributes were considered important. This

indicates that this new version of DEMO-based

SLAs has more potential to capture the customers’

expectations than the older version of the proposal

that had only 5 attributes. By specifying these

attributes, customers can structurally define their

expectations which may help the alignment between

customers and service providers. The specification

of the customers’ expectations into SLAs helps the

service providers to understand those expectations

and consequently reduce the gaps among the two.

The last step of DSRM, communication, is being

achieved through scientific publications aimed at the

practitioners and researchers within the service

science area.

As future work, we intend to evaluate the

proposal using practitioners’ feedback. Presently, we

are interviewing recognized practitioners involved in

SLAs specification. Furthermore, we will apply our

proposal in an organization with a more complex

service exchange.

REFERENCES

Albani, A., Terlouw, L., Hardjosumarto, G., & Dietz, J.

(2009). Enterprise Ontology Based Service Deﬁnition.

Amsterdam, The Netherlands: 4th International

Workshop on Value Modeling and Business

Ontologies.

Andrieux, A., Czajkowski, K., Dan, A., Keahey, K.,

Ludwig, H., Nakata, T., et al. (2007). Web Services

Agreement Specification (WS-Agreement). Open Grid

Forum.

CMMI for Services, Version 1.3. (2010). Software

Engineering Institute - Carnegie Mellon University.

Dietz, J. (2006). Enterprise ontology - theory and

methodology. Springer.

Dobson, G. (2004). Quality of Service in Service-Oriented

Architectures. Retrieved June 2, 2011, from http://

digs.sourceforge.net/papers/qos.html

Frolund, S., & Koistinen, J. (1998). QML: A Language for

Quality of Service Specification. HP Software

Technology Laboratory.

Hevner, A., March, S. T., Park, J., & Ram, S. (2004).

Design Science in Information Systems Research. MIS

Quarterly.

Keller, A., & Ludwig, H. (2003). The WSLA Framework:

Specifying and Monitoring Service Level Agreements

for Web Services (Vols. 11, No. 1). Journal of

Network and Systems Management.

Kvale, S. (2007). Doing interviews. London: Sage

Publications.

Liu, Y., Ngu, A., & Zeng, L. (2004). QoS Computation

and Policing in Dynamic Web Service Selection. New

York, USA: 13th International World Wide Web

conference on Alternate track papers & posters,

ACM.

Mendes, C., & Mira da Silva, M. (2012). DEMO-based

Service Level Agreements. Geneva: 3rd International

Conference on Exploring Service Science, Lecture

Notes in Communications in Computer and

Information Science, Springer.

Mendes, C., Ferreira, J., & Mira da Silva, M. (2011).

Comparing Service Using DEMO. Paris: 3rd

International Joint Conference on Knowledge

Discovery, Knowledge Engineering and Knowledge

Management - Special Session on Enterprise

Ontology.

Mendes, C., Ferreira, J., & Mira da Silva, M. (2012).

Identifying Services from a Service Provider and

Customer Perspectives. Lecture Notes in

Communications in Computer and Information

Science, Springer.

Ministry of Finance and Public Administration. (2010,

October). 2011 State Budget. Retrieved April 09,

2012, from http://static.publico.clix.pt/docs/economia

/PropOERel2011.pdf

Office of Government Commerce. (2007). ITIL v3 –

Service Design. The Stationery Office.

Parasuraman, A., Zeithaml, V., & Berry, L. (1985). A

Conceptual Model of Service Quality and its

Implication for Future Research. Journal of Marketing.

Peffers, K. e. (2008). A Design Science Research

Methodology for Information Systems Research.

Journal of Management Information Systems, M. E.

Sharpe.

UsingDEMO-basedSLAsforImprovingCityCouncilServices

439

Pombal City Council. (2011). 2011 Pombal City Council

Budget. Pombal City Council, Pombal.

Pries-Heje, J., Baskerville, R., & Venable, J. (2004).

Strategies for Design Science Research Evaluation.

16th European Conference on Information Systems

(ECIS), pp. 255-266.

Sahai, A., Durante, A., & Machiraju, V. (2002). Towards

Automated SLA Management for Web Services.

Technical Report, Hewlett-Packard Company.

Terlouw, L., & Albani, A. (2011). An Enterprise

Ontology-Based Approach to Service Specification.

IEEE Transactions on Services Computing.

Tosic, V., Patel, K., & Pagurek, B. (2002). WSOL – Web

Service Offerings Language. (C. '. '02, Ed.) London,

UK: International Workshop on Web Services, E-

Business, and the Semantic Web, Springer.

KEOD2012-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

440