ProActive Service Ecosystem Framework

1

In: International Journal of Computer Integrated Manufacturing, Volume 26, Issue 11, 2013, pp 1021-

1041.

Pro-Active Services Ecosystem Framework

Tiago Cardoso, Luis M. Camarinha-Matos Faculty of Sciences and Technology - New University of Lisbon, Portugal [email protected], [email protected]

Service Orientation is a main trend in the development of support infrastructures for Collaborative Networks, namely in the manufacturing sector. Nevertheless, a literature review still shows the lack of a common understanding of the service concept, namely when we compare the software and business perspectives. As a contribution towards the integration of these two perspectives, this article introduces a conceptual framework based on the notion of representation of services in an ambassador like manner through the introduction of a pro-activeness mechanism. An example scenario and results with a prototype implementation are also discussed.

Keywords: Pro-Active Services; web-service; multi-agent system; collaborative networks

1 – Introduction

1.1 – Service Orientation

Although widely used nowadays, the notion of service orientation is understood

differently when considered from the business perspective or the Information and

Communication Technology (ICT) perspective in manufacturing, as well as other

economical sectors. There is a gap between these two worlds regarding the service

concept itself. At the business level, the focus is put on value creation towards client

satisfaction or resources management. At the ICT level, the focus is put on remote

procedure calling or interoperability issues. In fact, although considerable progress was

achieved in Service Oriented Architectures (SOA) in the last decade, the way a business

entity sees the services it is able to provide to customers is distinct from the used

software, in terms of systems or components created to represent such services and / or

automate (part of) their provision.

Naturally this gap can also be observed in the case of a Virtual Organization

Breeding Environment or a Professional Virtual Community, from now on referred to as

a Collaborative Business Ecosystem (CBE). The way a CBE member sees the services it

2

is willing to provide is distinct from the notion conveyed by the available software

approaches, namely the Web-Services case.

Nevertheless, the adoption of Service Orientation, particularly the case of Web-

Services, has got wide acceptance in systems development, namely in terms of

interoperability (Phaithoonbuathong, 2010), for example towards wrapping the

functionality of some machine in a shop-floor landscape. In fact, the usage of SOAP,

WSDL and UDDI–related concepts and mechanisms, opened the possibility for

providers to publish their services and for the clients to find and call them. In a typical

simple scenario, enterprises create their web-services and publish them in a catalogue.

Afterwards, clients query such catalogue and receive a list of matching results. Finally,

the client selects the one that best fits its needs and calls it.

In fact, the adoption of Web-Services improved the way Information Systems

can be integrated. The main reasons for this improvement are based on the adoption of a

number of standards that are independent of the provider's environment or development

platform. This fact facilitates the development of value-added services by composition

of simpler services provided by different members of a collaborative network.

Nevertheless, although this approach is suitable for many scenarios, a number of

limitations can also be identified (Cardoso and Camarinha-Matos, 2011):

• Web-Services are “static” entities - Providers may publish and register them,

making their functionality available for clients to discover and call, but these

constructs stay still, waiting for the clients' initiative. In other words, they do not

perform any action in order to attract clients or promote their functionality.

• Difficult selection process from clients’ perspective - If the list of services

provided by a catalogue in response to a query has a large number of items, a

problem arises on how to make a selection.

• Limitation of the selection process from the providers’ perspective - In the

same case of a large number of matching available Web-Services, the providers

face a problem on how to improve the chances that their Web-Services have to

be selected.

• Catalogues may become out-of-date - There are scenarios where Web-Services

frequently change their availability. In such scenarios, the information

catalogues have may easily become out of date.

• No aggregation - A workflow definition can support a composition of Web-

3

Services in order to achieve a higher level goal or a higher abstraction level

functionality, which may itself become a Web-Service, as well. However, from a

contractual perspective, having one of these higher level Web-Services

composed of several simpler services, each one from a distinct provider, is

typically much more “expensive” than having the same number of services

provided by fewer entities. In other words, under similar circumstances it could

be desirable to select two services from the same provider, instead of resorting to

two distinct providers, which could lead to higher costs related to the agreement

reaching processes or assignment of responsibilities if something goes wrong.

The aggregation of distinct services from the same provider within a single

construct would improve this situation.

1.2 – Proposal at a Glance

The proposal made in this article is inspired in the service orientation paradigm and

aims at creating a bridge between the business and ICT worlds, through the creation of

pro-active systems that act as ambassadors for the representation of business services,

which in turn are intended to be executed mostly by persons. This framework intends to

reduce the identified gap with particular focus on overcoming some current limitations

of the Web-Service approach. Two key actions are addressed in order to illustrate this

auto-initiative representation: 1) finding new collaboration opportunities; and 2)

improving the chances that the represented services have to be selected among

competitors. The application scenario selected for the proof of concept is the provision

of consultancy services by senior professionals in the manufacturing sector.

2 – Motivation

2.1 – Service Orientation in the Manufacturing Sector

A growing number of cases can be found concerning the application of service

orientation in the manufacturing sector. Some relevant examples include:

4

• Shop-floor Processes – At the shop-floor level, there is the possibility to wrap

each machine’s functionality using Web-Services and the manufacturing process

becomes an integration or composition of web-services. This is a research line

where a considerable effort has been put on, as shown in (Souza et al., 2008),

(Karnouskos et al., 2007) or (Jammes et al., 2005). Still under the process

perspective, there are also researchers working on issues like scheduling or

reconfiguration of the processes, as it can be found in (Shen et al., 2007) or

(Ribeiro et al., 2009). The main advantage of this approach is the loosely

coupling of the production resources, as well as the usage of standard

technologies, resulting in a smooth systems integration.

• Supply-Chain Management – at the next abstraction level the synchronization of

distinct parts of a supply chain can also be made using SOA. Here, the focus

leaves the shop floor level and goes up to the integration of distinct production

lines. This research-line has also gained considerable attention, as for example in

(Preist et al., 2005) or (Tarantilis et al., 2008). The advantage of using SOA at

this level is also a smooth integration, based on the standards that are

independent from the underlying resources, but this time typically in a

distributed environment rather than an integration made locally.

• Monitoring / Supervision / Maintenance – Finally, another example application

of SOA in the manufacturing sector can be found concerning supervision and

maintenance of machinery. One can consider, for example, a provider of

manufacturing machinery that is located in one city but has the ability to

remotely supervise the machines located in his client’s factories. This

perspective has also been the focus of various research works, as for example

(Yang et al., 2004) or (Karnouskos et al., 2010). The main advantage of SOA in

this context is more concerned with the geographical distances, namely

considering that the specialists from the enterprise that supplies the

manufacturing machinery have the ability to remotely monitor them and travel

only if something goes wrong. Yet another branch in this research line is the

remote maintenance, as mentioned in (Zeeb et al., 2008) or (Wollschlaeger et al.,

2003).

It is interesting to assess the application of some of the identified limitation aspects of

5

web-services to these cases. One might consider, for example, a shop-floor scenario

where the functionality provided by the machines is wrapped using web-services that in

turn are listed in a catalogue. In this scenario, all the machine selection might be made

dynamically based on such catalogue information.

As mentioned, the services are modelled as “independent constructs”, even if two

distinct services may be provided by the same machine / cell. If, for example, some

machine provides a raw material assembling service and also has the ability to dry up

materials before assembling, the existing web-service selection mechanisms do not

consider this fact for a case where the two services are in need. In fact the “traditional”

service selection within a process composed of several services is made independently.

In what concerns the problem of service catalogues becoming outdated, problems may

occur whenever one machine becomes unavailable due to some malfunctioning

situation. In this case the information stored in the catalogue is not updated, at least in

an automated manner, and this may result in the selection of this damaged machine for a

new process, which will lead to undesired delays.

2.2 – Services executed by persons

Although the main focus of the service orientation / web-service research community

has been put in the ICT systems integration, other services not executed by information

systems also have to be considered. Although these “other” services may also involve

the interaction with information systems, their execution has to be performed by persons

and that fact raises a new set of issues that need to be analysed, as summarized in table

1:

Table 1 – Some issues in services provided by persons

Determinism

When a person is executing a specific service, this task usually does

not occur virtually instantaneously, unlike the case of a traditional

web-service. As a result, the time spent in the execution of some

service can not be deterministically assessed.

Assignment

Although in a similar way as information systems, mechanisms have

to be created concerning the assignment of services to professionals,

eventually based on their abilities. The approach of the Web-Service

6

Human Task standard, proposed by OASIS, is the usage of the role

concept.

Provider

Representation

When a person has the ability / willingness to provide a specific

service, he or she would benefit from a representation mechanism

towards looking for collaboration opportunities and informing that

person whenever a new opportunity is found, or even performing

promotional tasks for the corresponding services.

2.3 – Manufacturing, Ageing Workforce and Know-How

The demographic evolution has risen new challenges in developed countries, namely in

Europe. According to (Fornasiero et al., 2009) , “it is becoming more and more obvious

that western societies should carefully consider how to deal with some ongoing

dramatic demographic changes, showing that senior citizens and workers are going to

play an increasingly relevant role”. In fact, “the increase in aged population is a critical

issue in sustainability. As the expectancy of a longer and healthy life increases, the issue

of extending the active professional life of senior people becomes an important topic”

(Camarinha-Matos and Afsarmanesh, 2010). Particularly with the retirement of the so-

called “baby boomer” generation, companies namely in the manufacturing sector are

losing a considerable part of their knowledge base.

In fact, the knowledge attained during a life-time experience is an asset that companies

should not underestimate. Although several mechanisms do exist in order to perform a

knowledge transfer between senior professionals to juniors within companies, there is a

considerable amount of knowledge that is tacit and thus difficult to transfer. According

to (Grant and Gregory, 1997) , “the acquisition of tacit knowledge has been attributed to

learning by doing and learning by using”. As a result, in a manufacturing landscape, the

employees that have a long experience, within a company or factory, possess a valuable

asset built of tacit knowledge. On the other hand, the same authors highlight that “it is

easy enough to transfer hardware-blueprints, specifications, price lists, product

samples; but much harder to ensure the transmission of the intangible 'know-how'

which is in the minds of those who use the hardware”.

With the increase in both the percentage of senior citizens and expectation of healthy

7

life, it is timely to reassess the understanding of ageing and retirement, and questioning

elderly’s work, happiness, leisure, community involvement etc. and their inter-

relationships with respect to old age. Many elderly citizens, following retirement,

quickly become marginalized, losing most of their social and professional links which

often leads to loneliness. Some consider the elderly as a cost burden rather than a

resource that is capable of “value creation” in the society. And yet retired senior

professionals represent an important source of experience, wisdom, and talent. As the

older population increases and the growth in the middle-aged population slows down,

older adults are becoming an increasingly important labour source. They typically bring

maturity, dependability, and years of relevant experience to the workplace. Nowadays

with more people remaining in good health at older ages and increasingly more jobs not

involving physical strength, more old adults are able to continue working than ever

before. Many seniors would be willing and even enthusiastic about continuing to

contribute to society and the economy. Active ageing, through a balanced combination

of leisure, social interaction, and continued work involvement, is central to meeting

citizen expectations and maintaining mental and physical health (Camarinha-matos and

Afsarmanesh, 2011). The critical challenge for the society in respect of the ageing

process is to identify new organizational structures – and to support these with relevant

technologies - to avoid the exclusion of senior citizens from the market and society, and

to promote the use of their knowledge.

In addition to the traditional initiatives focused only on socialization and entertainment

activities for elderly, a number of other organizational forms and mechanisms, already

existing or emerging, are focused on providing ways to help seniors remain

professionally active after their retirement. In fact, as a response to the active ageing

challenges, in recent years we can observe the emergence of a growing number of

associations of senior professionals. Although bringing some benefits to their members,

namely in the scope of socialization, the current model of these associations may lead to

some form of “organized ghettos” for seniors with little inter-generational interaction. In

order to guarantee a more holistic integration, the different socio-economic stakeholders

must develop a whole series of business and behavioural models of varied structure,

composition and characteristics to allow a variety of alternatives for the integration of

senior professionals.

Nevertheless, associations of retired professionals already represent an important

organizational structure supporting active ageing (Camarinha-Matos and Afsarmanesh,

8

2010). The aim of these associations is essentially to provide assistance to organizations

and people that cannot afford market prices for such services, or assistance to

organizations located in developing countries. Offered services are basically business

consultancy and mentoring. The majority of them work on a voluntary basis sustaining

themselves through membership fees and/or private or public funding and sponsorship.

In terms of ICT, the majority of these organizations only use basic tools; nevertheless

they are willing to use new and modern tools to help in their daily activities. The

development of new environments facilitating the operationalization of such services

would represent an important contribution to the sustainability of the manufacturing

knowledge base.

3 – Related Research

This research is based on two main inspiration lines: a conceptual baseline combining

the Collaborative Networks (CN) and the Services Science areas; and a technological

baseline gathering elements from the Web Services and the Multi-Agent Systems areas.

The following sub-sections highlight the main elements selected from these research

areas.

3.1 – Services Science

The services sector has experienced a growing importance path, especially in the last

two decades. In fact, the world labour has partially moved from agriculture and

manufacturing into the services sector (Maglio et al., 2006). It is interesting to notice

that the service designation appeared in opposition to agriculture and manufacturing, in

the 1930s (Chesbrough and Spohrer, 2006). At that time, the first two sectors of

economy were the major sectors in terms of employment and “services was a residual

category for other activities that did not fit into agriculture and manufacturing”.

Nowadays the services sector is the one with more jobs in developed countries.

Nevertheless, despite the growth of the tertiary sector in the world’s economy,

productivity in this sector is low, when compared with the manufacturing industry (Abe,

2005). One of the commonly accepted factors that dictate this low-productivity aspect is

that the service sector is heavily based in intuition and experience of employees, rather

than systematic processes. The magnitude of this phenomenon gains a particular

9

meaning when the dimension of the baseline grows from one service to the integration

of distinct services from independent entities forming service systems. In other words,

the “service system complexity is a function of the number and variety of people,

technologies and organizations linked in the value creation networks, ranging in scale

from professional reputation systems of a single kind of knowledge worker or

profession, to work systems composed of multiple types of knowledge workers, to

enterprise systems, to industrial systems, to national systems and ultimately to global

service systems” (Maglio et al., 2006).

As a result of this situation, and as foreseen in (Horn, 2005), the new concept of

“Services Science” has emerged with the goal of creating the basis to increase

productivity in the services industry, promote innovation and create greater viability and

transparency when assessing the value of investments in services. According to

(Spohrer et al., 2007), “services science aims to understand and catalogue service

systems and to apply that understanding to advancing our ability to design, improve and

scale service systems for practical business and societal purposes.”

Although a common definition has not yet been reached, a simplistic statement

from (Paulson, 2006) argues that “in essence, it represents a melding of technology with

an understanding of business processes and organization”. From another perspective,

“service science aims to categorize and explain the many types of service systems that

exist as well as how service systems interact and evolve to co-create value” (Maglio and

Spohrer, 2008). Yet another definition of service (Ferrario and Guarino, 2009; Ferrario

et al., 2011), states that “a service is present at a time t and location l iff, at time t an

agent is explicitly committed to guarantee the execution of some type of action at

location l, on the occurrence of a certain triggering event, in the interest of another agent

and upon prior agreement, in a certain way.” This definition is particularly applicable in

the context of Services Science, given the fact that it includes the commitment aspect,

coping with the business perspective, the temporal and logistics facet, and the

management perspective, and it can be materialized into a technological

implementation, for example using the multi-agents paradigm and / or service-oriented

architectures.

10

3.2 – Collaborative Networks

Collaboration among distinct independent entities is an old practice. The usage of

Information and Communication Technology to support such collaboration and break

out the barriers of geographical distribution, along with actual market turbulence and

the so-called globalization are some of the drivers of the new Collaborative Networks

scientific area.

This area has also gained a special attention from the research community in the

last two decades. According to (Camarinha-Matos and Afsarmanesh, 2008) there is

already a sound empirical knowledge, and a preliminary theoretical foundation for

collaborative networks. The “emergence of the virtual enterprise (VE) / virtual

organization (VO) paradigm falls within the natural sequence of the restructuring

processes in traditional industrial paradigms that is enabled by advances in information

and communication technology”. According to the same authors, “the idea of VE/VO

was not invented by a single researcher; rather it is a concept that has matured through a

long evolution process”. In fact, this scientific discipline can be included in the path of

systems integration, towards a global integration level.

Along its history, the CN area has been enriched with a set of concepts defining

base collaborative manifestations. In what concerns the collaboration between

enterprises or organizations, the concepts of Extended Enterprise, Virtual Enterprise or

Virtual Organization, are the collaboration opportunity (CO) oriented definitions, whilst

the Virtual Organization Breeding Environment targets long term collaborative

agreements. Another branch from this area developed an understanding of the concepts

concerning collaboration among human professionals. In this case the developed

concepts include Virtual Team and Professional Virtual Community. The former is

oriented to a response to a collaboration opportunity that may arise and the later targets

long term agreements between professionals towards being prepared to faster address

collaboration opportunities.

The formation phase form the CBE lifecycle is the main target of this work.

Within this period this scientific discipline addresses distinct perspectives, some of

which are closely related to this work: 1) Competences / Skills Profiling – needed in

order that the CBE members express what services they are able / willing to provide

(Chenghua and Weston, 2007); 2) Partners Search and Selection – needed after detailing

the services needed to accomplish a specific CO (Baldo et al., 2008); 3) Performance

11

Measurement – needed in order to assess the quality of services provided by CBE

members and support selection purposes (Alfaro et al.); 4) assessment of the most

suitable network configuration for a given CO, based on distinct value creation

scenarios (Elhabib et al., 2010).

3.3 – Technological Contributions

The Web-Services, as the main inspiration area for this work, have evolved through

three main phases that can be identified by the main keywords: Publish, Register / Find

and Compose. Publish, when the concept was presented and used by any entity wanting

to make some procedure available towards reaching potential worldwide clients. As

time shown, this expected large number of worldwide potential clients did not become

effective, since they were not notified and could not find the published services. At that

time, UDDI and WSDL were proposed and services could be registered in catalogues,

in order to be found. Afterwards, the creation of Value-Added-Services, made through

composition of other Web-Services was the advent of the third generation, resorting to

mechanisms like Workflow and technologies like BPEL4WS. Other initiatives like

OSGi provided the technological support for dynamic composition of web-services

(OSGi, 2007), (Jammes and Smit, 2005). The Web-Service Eventing introduced a

protocol that allows Web services to subscribe to or accept subscriptions for event

notification messages. Finally, the notion of Service Entities (Franco et al., 2009)

introduced a first aggregation approach for distinct Web-Services within a single

construct. Figure 1 briefly represents the main stages of this evolution.

12

Figure 1 – Web-Service Evolution

The other main technological contribution comes from the Multi-Agents research area,

as mentioned. “Agent technology has been recognized as a promising paradigm for next

generation manufacturing systems” (Shen et al., 2006). This approach has four

important characteristics (Wooldridge, 1998): autonomy, reactivity, social-ability, and

pro-activeness. First, agents can operate without a direct human intervention. Second,

agents perceive the environment surrounding them and can react to it. Third, they can

interact with peers / other agents, towards achieving common goals. Fourth, agents do

not simply respond to their environment "triggers" but they can take the action

initiative, towards the goal they pursue.

The active representation of the business services that CBE members are able /

willing to provide within a CBE, made through an ambassador-like manner, is presented

in section 4. This proposal gathers the pro-activeness notion that will be used to behave

in an auto-initiative basis towards pursuing the business interests of the represented

CBE member, namely looking for new Collaboration Opportunities, as mentioned

before.

3.4 – Other Research Initiatives

Other research initiatives tackle the usage of service orientation and also Multi-Agent

Systems in the Collaborative Networks area, namely ICT-I (Rabelo et al., 2006),

ManBree (Franco et al., 2009) and KIMM Framework (Kuk et al., 2008). These three

initiatives have particularly inspired this research work, as highlighted bellow.

ICT-I is the approach used in the ECOLEAD project to help members of CBEs

in making businesses and collaborations more efficiently. The approach followed in this

initiative relies on the service orientation and acts as a bus allowing different and

distributed organizations to interact. The main objectives of ICT-I are: 1) collaboration

and negotiation among people; 2) Interoperability between ICT systems, as well as their

ability to adapt to the surrounding environment; 3) Support for knowledge and resources

discovery and sharing; and 4) Synchronization and possible interconnection between

processes.

Although ICT-I supports a some degree of auto-initiative ability for the base

constructs to adapt themselves to the surrounding environment, this solution also shares

13

some of the identified limitations of web-services, namely their passiveness in what

concerns pursuing business interests.

ManBree is a VBE Management System based on Service Entities, that

addresses three main functionalities: VBE Management, VO Management, and Service

Entities Management. The most innovative aspect of ManBree is its base concept, the

Service Entity, which embraces a finite set of business services and a finite set of

attributes aiming the characterization of such entities towards distinguishing them.

Furthermore, Manbree supports the definition of Abstract Service Entities (ASE) in a

first place, where a skeleton of attributes is included, and the instantiation of such ASEs

into Concrete Service Entities (CSEs) made in a second place.

As mentioned before, the ManBree is one of the inspiration initiatives,

especially because of its first notion of aggregation, including distinct services from the

same entity within a single construct – the Service Entity. Nevertheless, this approach

also lacks the introduction of pro-activeness or auto-initiative towards actively

representing CBE member’s business interests instead of “delegating” all the initiative

to the client side.

KIMM is an engineering framework based on a service-oriented architecture and

agent’s technology aiming to provide an integrated environment to support

collaboration among the elements participating in a product development process. The

main features of this framework are the integration of distributed resources and the

orchestration of engineering activities. Although KIMM’s initiative addresses the

passiveness issue identified as a “bottleneck” of web-services, the pro-activeness of the

proposed base elements is restricted to a negotiation between agents mainly concerning

workflow issues. One limitation of this approach can be stated as a lack on the usage of

the auto-initiative aspect for other business-oriented goals, like finding business

opportunities or the ability to improve the chances a CN member has to see its services

selected among competitors, based on Quality of Service (QoS), for example.

4 – Suggested Approach

Starting from the limitations identified in the introduction section, as well as the vision

of a CBE member’s service active representation in an ambassador-like manner, the

Pro-Active Service Ecosystem Framework (PASEF) is now introduced. This framework

targets shortening the distance between the business world and the ICT counterpart

14

views of services and gathers the contribution of both Service Oriented Architectures

and Multi-Agent Systems. The base idea is to provide an environment fostering the

collaboration between CBE members through an active representation of the business

services they are able / willing to provide, as mentioned before. This representation is

made by constructs that the CBE members configure; both including the services to be

represented and selecting the behaviours such elements will perform in an auto-

initiative basis towards pursuing the desired goals or objectives. Two initial behaviours

were identified for this representation: 1 - Finding new business opportunities where the

services may be included as contributing components; 2 – Improving the chances that

such services have to be selected among competitors.

4.1 – Service stereotyping related concepts

The definitions used in the proposed framework start with the Service Stereotyping (SS)

related concepts group. This group of concepts is needed for distinct CBE members to

share a common understanding of the services they provide within the ecosystem

towards a smooth interaction between the services, namely for composition purposes.

The group is composed of three concepts: Meta-Service, Service Category, and Service

Taxonomy. These elements are instantiated by the ecosystem initiator.

Definition SS 1: Ecosystem’s Service Taxonomy An Ecosystem’s Services Taxonomy, Taxonomy for short, is the specification of a

common understanding about services to be provided within that Ecosystem. A

Taxonomy is composed of generalization / specialization relationships among a super-

type of services – the Service Categories - under which the Meta-Services are defined.

The main goal of this taxonomy is to foster better communication among users,

avoiding misunderstandings concerning the service definitions. A Taxonomy T may be

defined as a tuple:

T = < N, D, SCS > where:

• N – Ecosystem Taxonomy Name – the identification of the Taxonomy.

• D – Description of the Taxonomy – including the economic area to which it concerns.

• SCS – Service Category Set.

The following rule applies to every Taxonomy T, stating that none of the values N or D

may be null nor the set of service categories SCS may be empty:

15

∀T, ∃�N, D, SCS�|N ≠ null, D ≠ null, SCS ≠ ∅

Definition SS 2: Service Category A Service Category is a group of Meta-Services that share common characteristics,

within a given Taxonomy. A Service Category SC can be defined as a tuple:

SC = < N, D, T > where:

• N – Service Category Name – the identification of the category of services.

• D – Service Category Description – description information including common

characteristics of the services from SC.

• T – Service Category Taxonomy – the Ecosystem’s Services’ Taxonomy.

The following rule applies to a Service Category SC, stating that the elements of the

Service Category have to be defined, a priori:

∀SC, ∃�N, D, T�|N ≠ null, D ≠ null, T ≠ null

Definition SS 3: Meta-Service A Meta-Service is an abstract definition of a service to which concrete services

provided by CBE members have to comply. A Meta-Service MS can be expressed as a

tuple:

MS = < N, C, D > where:

• N – Meta-Service Name – the identification of the Meta-Service.

• C – Category – the category to which the Meta-Service belongs.

• D – Description – particular characteristics of MS, like the specification of needed input

information and output result. This description follows the BPEL4WS standard for

interoperability purposes.

4.2 – Membership Modelling related concepts

The second group of concepts included in PASEF is the Membership Modelling (MM)

related concepts, which are needed to model CBE members, what they are able / willing

to provide and the ecosystem itself.

Definition MM 1: Service

16

Service is the concept that models what a CN member is willing to provide to the

clients. In other words, a service is a concrete instance of a Meta-Service. This concept

also involves another element defined by the CN member: Provision Conditions –

specifying constraints for the service provision to be accepted.

A Service S can be expressed as a tuple:

S = < N, M, MS, SPC > where:

• N – Name – the identification of the Service.

• M – CN Member – the identification of the CN member that provides the specific

service.

• MS – Meta-Service – the Meta-Service, defined in the Ecosystem’s Service Taxonomy,

of which S is an instance.

• SPC – Service Provision Condition set – ����� �i ∈ ℕ}.

Definition MM 2: Service Entity Service Entity is a construct that models CN members from the service provision

perspective. It includes the Services that a particular CN member is willing to provide

and a set of attributes modelling relevant characteristics of that CN member. A Service

Entity SE can be represented as a tuple:

SE = < M, ATS, PSS >

where:

• M – CN Member – the identification of the CN member.

• ATS – Attribute Set – ������ |i ∈ ℕ} –set of relevant attributes of the corresponding CN

member

• PSS – Provided Service Set – ���� |i ∈ ℕ} – the set of services provided by the CN

member.

Definition MM 3: Behaviour Definition

A Behaviour Definition specifies the actions that a PSE will perform and the event that

triggers such behaviour. A Behaviour Definition BD can be expressed as a tuple:

BD = < ID, D, TM, BWD, PREC, POSC >

where:

• ID –Identifier – the Identifier of the behaviour.

• D – Description – a description of the behaviour.

• TM – Triggering Mechanism – timings, frequency and / or data-flow conditions

specifying when the execution will be launched.

17

• BWD – Behaviour Workflow Definition – specification of the base functions that are

used within the behaviour, their input information needs, output results and their

execution flow graph.

• PREC – Pre-Condition Set – ��� �� |i ∈ ℕ} – a set of conditions to be verified before

the behavior is launched.

• POSC – Post-Condition Set – ��!���� |i ∈ ℕ} – a set of conditions to be verified after

the behaviour finishes, assessing its success.

Definition MM 4: Pro-Active Service Entity (PSE)

The Pro-Active Service Entity is a concept that includes a CN members’ Service Entity

and a set of behaviours selected and configured by this CN member towards

representation purposes. A Pro-Active Service Entity PSE can be expressed as a tuple:

PSE = < N, ID, SE, BDS >

where:

• N – Name – the identifier of the PSE.

• SE – Service Entity – the Service Entity from the CN member including what this

member is willing to provide.

• BDS – Behaviour Definition set – �"#� |i ∈ ℕ} – a group of behaviour definitions,

selected and configured by the CN member, which specify the pro-activeness of the

construct.

Definition MM 5: Services Ecosystem The collaborative Services Ecosystem concept, Ecosystem for short, brings together

CBE members through their PSEs, final clients and intermediaries or brokers into a

single space fostering collaboration. This is a central “place” that manages information

both concerning service providers and clients, as well as the Collaboration

Opportunities launched by clients and performed by the providers – the CBE members.

The Ecosystem also stores information concerning the Brokers that make the bridge

between clients and providers, as well as Performance Measurement information used

for QoS purposes. Finally, the Ecosystem provides a set of functionality groups towards

fostering the intended collaboration among its members, including a certification

mechanism that attests PSEs performance on demand. A collaborative Services

Ecosystem SEcoSys can be expressed as a tuple:

SEcoSys = < N, ST, PS, CS, BS, CO, PM, CR, BF > where:

• N – Name – the identifier of the collaborative Services Ecosystem.

18

• ST – Service Taxonomy – the base Service Taxonomy used in all interactions by the

PSEs within the Ecosystem in order to guarantee a common understanding on service

definitions.

• PS – PSE Set – ��� �|i ∈ ℕ} – set of PSEs that represent CN members’ services within

the collaborative Services Ecosystem.

• CS – Client Set – ��$%�� |i ∈ ℕ}.

• BS – Broker Set – �"�&� |i ∈ ℕ}.

• CO – Collaboration Opportunity set – ��!� |i ∈ ℕ} – set of client needs / collaboration

opportunities specified by the clients in the first place and further detailed by the

intermediaries or brokers, within the Ecosystem.

• PM – set of Performance Measurement Information tracked within the ecosystem every

time a PSE participates in a CO. This set can be expressed as:

�pm),*|i, j ∈ ℕ, ∀pm),*∃pse) ∈ PS, ∃co* ∈ CO} meaning that every performance measurement information element pmi,j corresponds to

a specific psei performance within a collaboration opportunity coj.

• CR – set of Certification Information {cri | i Є ℕ } – the set containing all the

certification information provided by the Ecosystem concerning specific PSEs.

• BF – Built-in Functionality – {freg, fpost, fperf, fcertify, fother} – set of groups of functionality

built-in the Ecosystem, based on which the PSE Behaviour Definitions are composed

of. The 5 built in functionality groups identified are:

o freg – Registration – enabling the CN members to register themselves and launch

their PSEs.

o fpost – CO Posting – enabling clients to post their needs in a blackboard-like

infrastructure from the EcoSystem and PSEs to reply with their service

provision proposals.

o fperf – Performance Measurement – providing PSE performance measurement

mechanisms, as well as enabling brokers to grade performance, in order to

increase the information on every PSE that is registered in the Ecosystem.

o fcertify – Certification – based on PM information, the Ecosystem may have a

component responsible to certify some PSE’s QoS upon request.

o fother – other than the 4 identified functionality groups, each particular

Ecosystem may provide specific functionality under this group, that constitutes

a possible extension point for the ecosystem.

19

5 – Logical Architecture / PASEF design

5.1 – Multi-level Modelling

The usage of the Pro-Active Service Ecosystem Framework comprises three abstraction

layers, as represented in Figure 2: The Actors Space, the Service Market Space and the

Integrated Service Space. In the Actors Space, for the chosen application scenario, we

can find senior professionals from the manufacturing sector that possess a life-long

experience and may provide consultancy services – the CBE members. At the Service

Market Space, there are the representatives of the services such CBE members are able /

willing to provide – the PSE layer. At the top level, the Integrated Services Space, there

are the higher level services that result from the composition of simpler services made

by the brokers whenever a client specifies a high-level need, and implicitly correspond

to consortia created in response to Collaboration Opportunities.

This multi-layered modelling space has some advantages: 1 - Pro-Activeness –

PSEs may find interesting COs, prepare proposals and submit them in an auto-initiative

basis; 2 - Aggregation - distinct services from a CBE member are aggregated within a

single construct - this can be useful in a composition process, in order to decrease

consortia dimension, based on the inclusion of partners that can provide more than one

needed service. To some extent, the PSEs, at the middle layer, make the bridge between

the two other layers.

Figure 2 - Three Abstraction Layers

4.2 – Actors, Roles and Logical Architecture

The PASEF design starts with the identification of the system Actors, as well as their

20

main roles and mechanisms. Five Actors are identified:

• Clients - making high-level specifications of the needs,

• Brokers - responsible to prepare proposals in response to clients' needs and

select the Services that best fit these needs,

• Pro-Active Services Ecosystem Administrator.

• CBE members - the providers of the services.

• Pro-Active Service Entity - although not a human actor, PSEs are considered

like actors, given their pro-activeness.

The identified roles and the corresponding mechanisms are:

• PSE-Ecosystem-Portal Administration - providing Monitoring tasks and

Managing Performance Information,

• Performance Measurement - providing the needed mechanisms to assess the

performance of Services, consortia, as well as client’s satisfaction,

• Accounting - providing the billing mechanisms corresponding to the single and

consortium Services’ provision,

• Contract Management - providing the legal consortium contractual support,

• Service Integration - providing the Service composition mechanisms,

• Workflow Engine - providing the Service execution mechanisms,

• Assistants to both CBE member, Clients and Brokers,

• SE Representation – the PSE role.

The Logical Architecture of the Pro-Active Services Ecosystem Framework follows a

star-like structure, as shown in Figure 3, as the central system – the Services Ecosystem

Portal - is surrounded by the CBE member representative systems – the PSEs. In a first

place PSEs are configured by the corresponding CBE member. Next, the Clients, the

Brokers and lately the PSEs interact with the EcoSystem through the portal, as detailed

bellow in the service composition process.

21

Figure 3 - PASEF Star-like structure

5.3 – Service Composition Process

The second step for the development of PASEF is the specification of the Service

Composition Process, leading to the establishment of consortia among the CBE

members. The framework usage process is divided into 5 temporal phases, Figure 4.

Figure 4 - Framework usage phases

• I - Configuration Phase

(1) PSE Configuration - first a CBE member downloads a PSE template and

configures it. This includes setting up the Services to be represented by such

PSE, filling in information concerning the provider – the CBE member and

22

selecting / configuring a set of behaviours for representation purposes.

Afterwards, the PSE has to be launched so it starts looking for COs.

• II – Collaboration Opportunity Specification Phase

(2) Clients Needs Specification - Through a Client Assistant, a high level

specification can be made starting a new Collaboration Opportunity.

(3) Broker details Business Needs - Based on the high level CO specification the

Broker creates a Business Process Model (BPM), through a Service Integrator

Module, detailing the required Services, taken from the Ecosystem Service

Taxonomy, to accomplish the specified needs.

(4) Client BPM Commit - When the BPM specification is ready, the Client is

requested to commit to that BPM.

(5) Broker writes needs on Blackboard - After the Client Commits to the BPM, the

Broker posts the specific service needs in a blackboard from the Ecosystem.

• III - PSE Proposals Phase

(6) PSE checks for Business Needs - The PSE looks for Collaboration Opportunities

on the Ecosystem’s blackboard towards matching one of the Services that it

represents;

(7) PSE prepares a Proposal to submit in the case of a positive match.

(8) Provider Commits to the Proposal - PSE can be configured to send the proposal

automatically or to ask the provider to complete / review it before submission.

• IV - Service Selection / Negotiation Phase

(9) Broker checks Proposals – the Broker checks the received proposals and selects

the ones that best fit the needs. A negotiation process can also take place a this

stage.

(10) Contractual Commitment - Both the Client and all the selected Providers have to

commit with each other through a contract generated for that purpose.

• V - Execution Phase

(11) BPM Launch - After the BPM has all the needed providers selected, it is

possible to launch it, through a trigger made by the Client. After that, a

23

Workflow Engine is responsible for asking the provision of each Service at the

right moment.

• Edition at Runtime - at anytime during the execution phase, the Broker can

complete or change the BPM, rolling back:

o to phase II - when the BPM did not get complete before starting

execution. In this case, the Broker has to complete it, and eventually,

change some Services;

o to phase III / IV - when some providers have not been selected yet or

there is some ongoing negotiation process.

5.4 – Requirements Engineering

After the identification of the actors, their roles and the mechanisms that should be

supported by the framework, the next step is to go down into a more detailed descriptive

level, through a systematic requirements engineering specification. The selected

approach for this phase is the usage of the i* (i-star) framework based on two main

factors: 1) i* is goal-oriented, resulting in more intuitive diagrams, which help

connecting this modelling framework with the concepts defined; 2) the bridge between

i* models and the tools used in later stages of software analysis and design phases is

also extensively addressed in the literature, namely for the case of UML.

The two i* model types presenting this specification are: one “Strategic

Dependency” Model (SD) that describes an overview of the goals / interactions between

the identified actors, and two “Strategic Rational” Models (SR) that “zooms in” the ICT

systems, identifying how the main goals are performed. Figure 5 represents the

described overview of PASEF. The main “Soft-Goal” of the framework is the “service

provision” from the CBE member (at the right-bottom “corner” of the diagram) to the

Client (at the left-bottom “corner” of the diagram). It is reasonable to say that all goals

(hard and soft) and Tasks in the diagram, between these two actors, “positively

contribute” to this higher level soft-goal.

24

Figure 5 – Strategic Dependency Model

It is also interesting to notice that this diagram corresponds to the elements from the

BPMN-like diagram defined in the previous stage of the software lifecycle in Figure 3.

As mentioned before, a zoom in is now made in order to specify how the PSE and the

Service Ecosystem, as the main ICT systems, are intended to perform the goals

identified in this diagram. Figure 6 and Figure 7 show these two “Strategic Rational”

Models.

Figure 6 – Services Ecosystem – Strategic Rational Model

25

The main internal goals of the Services Ecosystem include:

• Service Taxonomy Management – this goal provides a common understanding

within the Ecosystem. All providers (PSEs and the corresponding CBE

members) and all the consumers (Brokers and the corresponding Clients) have to

comply with the definitions made by this goal.

• Collaboration Opportunity Management – this goal is responsible to manage all

the COs and the corresponding PSE proposals.

• Workflow Edition – this goal is the one upon which the creation of the detailed

workflow plans depend. This goal is decomposed in the task “addition of

services” that itself is decomposed into two tasks: the insertion of activities and

transitions. This group of tasks create the skeleton of a workflow plan.

Afterwards, there is the need to post Calls for Proposals in the Services

Ecosystem blackboard, so that PSEs may become aware of the needed services.

Finally, the selection of the best proposals concludes the creation of executable

workflow models.

• Workflow Execution – this goal is responsible for using the executable

Workflow Models created and launch each service at the right moment.

Figure 7 - PSE – Strategic Rational Model

The PSE SR follows the same approach as the Services Ecosystem SR. The main goals

are mapped and decomposed in their main tasks. As identified in the global SD model,

the PSE interacts with the Services Ecosystem software system and the CBE member.

As the main aim of this software system is to represent the services from the CBE

26

member in the Ecosystem following an auto-initiative approach, the CBE member needs

to configure it, not only concerning some profile data and the services he or she is

willing to provide to the Ecosystem, but also how autonomous the PSE should be, as

well as how it should behave.

On the other side of the diagram, the main dependencies between the PSE and

the Services Ecosystem are mapped as the goal of collaboration opportunity

management. This goal is then divided in two main tasks:

• Check existing collaboration opportunities – this task is performed by the PSE in

a pre-defined frequency rate, allowing the PSE to become aware of new COs in

a reasonable time-frame, after such COs have been posted by a broker.

• Prepare / Submit Proposals – in a later stage, whenever a CO is found, matching

the represented services, the PSE is responsible to prepare a proposal and ask the

represented CBE member to edit and commit to such proposal, so it can be

submitted.

After this Requirements Engineering phase, the UML models are used towards further

detailing the systems composing PASEF. In a first stage, all the goals identified in this

stage are considered as candidates to become elements of Use Case Diagrams.

Furthermore, the SR models give an important contribution for the definition of UML

collaboration diagrams that may be created towards specifying how the systems should

behave whenever each use case is triggered. Although this phase is not included in this

article, for this later stage, Sequence Diagrams or State Transition Diagrams were used.

6 – Implementation Approach and Validation

6.1 – Developed Prototype

Three possible approaches were considered for the implementation of the PASEF

prototype, namely:

(1) Develop the whole system from scratch, including the multi-threaded

mechanisms in order to create independent and autonomous PSEs, as well as the

message exchange mechanisms;

(2) Build the framework on top of an existing MAS middleware solution;

(3) Build the framework on top of an existing Web-Services middleware solution.

27

The adopted solution is based on the integration of options 2 and 3. The specific

platform on which the prototype system was developed is JADE which is supported by

an active research community, offers an easy development approach and provides three

agent characteristics that are aligned with the Pro-Active Service Ecosystem Framework

concepts (Bellifemine et al., 2005): 1 - an agent is autonomous and pro-active; 2 -

agents are loosely coupled, meaning that the communication is asynchronous and no

temporal dependency exists between message senders and receivers; 3 - the system is

peer-to-peer, meaning that each peer or agent is equally privileged. The integration of

JADE and Web-Services was made through the usage of the Web Service Integration

Gateway (WSIG), defined by FIPA, since it is a user-friendly bridge already existing

between the two “worlds” – MAS and SOA. Therefore the development targets the

proof of the defined concepts through a Web-based prototype infrastructure built on top

of the JADE platform.

In terms of practical validation, the proposed framework is applied to a

collaborative network (PVC-like) of senior professionals that want to remain

professionally active after retirement, as mentioned before. The purpose of this

application is to “support active ageing and facilitate better use of the talents and

potential of retired or retiring senior professionals”, as proposed in (Camarinha-Matos

and Afsarmanesh, 2010). In fact, three main perspectives can be identified concerning

the current early retirement of people in many countries:

(1) Retirement age typically happens long before the age when elderly people’s

working capabilities strongly decrease.

(2) Many senior professionals prefer to continue working, although under a more

flexible schema, instead of starting a process that many times leads to loneliness

based on the lack of activities to fulfil the time retired people have, although

they are not used to.

(3) The knowledge attained during a life-long experience is an asset that the socio-

economic system could benefit from and many elderly persons feel glad to

share.

The implemented prototype is composed of 6 modules:

(1) PASEF Toolbox – Designed to enable a fast scenario definition, as well as to

launch, test, and monitor all the other modules that form the Pro-Active Service

28

Ecosystem Framework prototype (Figure 8).

Figure 8 – PASEF Toolbox

(2) Service Taxonomy Management – a module was developed for the specification

of a service Taxonomy made by the ecosystem initiator (Figure 9 – lower side).

(3) Seniors' Community Management – in this case, PASEF is intended to support a

Professionals Virtual Community and, for that reason, a community

management module is needed in order to enrol the Senior Professionals that are

willing to provide services where their life-time expertise can be used. This

module provides the functionality of service specification. After the initial

configuration phase a Pro-Active Service Entity is launched in order to represent

the corresponding professional, i.e. the PSE acts like an “ambassador” in the

service ecosystem representing the senior.

(4) Senior Pro-Active Services Ecosystem Portal – the module that monitors the

activity of the ecosystem, showing active PSEs, open Business Opportunities

and the messages exchanged among distinct actors of PASEF (Figure 10).

29

Figure 9 – Community / Taxonomy

Management

Figure 10 – Pro-Active Services

Ecosystem

(5) Workflow Editor – In order to make the specification of BPMs, to satisfy the

corresponding COs that are posted by clients or brokers, a Workflow Editor was

created. Figure 11 shows a simple example BPM (still without PSEs assigned to

activities). In a second stage, after the PSEs find the newly posted CO they

submit proposals in case there is a matching between the needs and the

represented services. The final step for brokers is to select the proposals that best

fit the needs.

Figure 11 – Sample BPM

(6) Workflow Engine – Finally, a workflow engine is needed in order to launch the

execution of the BPMs. Figure 12 illustrates a BPM that is being executed,

through messages exchange (engine, PSE) and (PSE, SP).

30

Figure 12 – Sample BPM in Execution

In the current implementation, the Workflow Editor and the Workflow Engine are built

as extensions of the open source solutions from Together Corporation.

In the considered application context, each senior professional starts by

registering his / her abilities or services he or she is willing to provide to the

community; next the PSE configuration and launch steps are performed. After this first

set-up process, when clients post their needs in the PASEF’s blackboard, all PSEs

become aware of the newly added CO within a limited time-frame. After that, the

matching mechanism takes place, leading to a group of potential providers for that CO.

At that point, each involved PSE notifies the corresponding senior professional in order

to get the “ok” for the proposals. And the process goes through.

The overall framework worked well under this experiment, although considering

it is a prototype environment. As a major feature of PASEF in comparison to current

solutions, the “inversion of the process”, putting the initiative on the PSEs side turned

out to be an interesting improvement for two reasons: 1 – from the client or broker

perspective, the number of potential providers increases proportionally to the scale of

the CBE and does not get restricted to a limited set of providers, as in the case when the

initiative of asking for proposals goes directly from a broker to some providers; 2 –

from the senior professionals’ perspective, whenever a CO matching their abilities

appears, they are sure they will be represented (by their PSEs) and have the chance to

submit a bid.

31

On the other hand, from this experiment it also became clear that PASEF should

work with other complementary systems that could manage contracts and accounting,

for example, in order to cover the real case needs.

6.2 – Benchmark

Other than the prototype development, two benchmark exercises were carried out

towards further validating both the proposed approach and the achieved solution. In the

first case, the approach was compared “against” the three systems mentioned in the

Related Research Section. The prototype itself was compared against two existing

consultancy service sites: lifeperson.com and freelancer.com. The TOPSIS method

(Hwang et al., 1993) was used for these benchmarking exercises. At a glance, this

method follows a sequence of 7 steps: 1) identifying the parameters that are relevant to

compare between the elements being considered; 2) classification of each element under

the selected parameters; 3) normalization of the classifications; 4) decide the relative

weight of the classification parameters; 5) identify the ideal and anti-ideal values by

selecting the best and worst classifications; 6) calculate the distance of each

classification to the ideal and anti-ideal values, D+ and D-, respectively; and 7) based

on D+ and D- calculate a relative proximity value C, that is the result of the method.

The two exercises are summarized below, showing only the selected parameters, the

classifications and the final result.

6.2.1 – Approach Benchmark

Table 2 shows the comparison parameters under which each approach was classified,

including PASEF.

Table 2 - Approach Benchmark - Classification Parameters Name Description

Pro-Active Constructs

How pro-actively do the constructs representing Business Services behave towards business success?

Aggregation Does the approach aggregate distinct services provided by the same entity, taking benefit from that?

Adaptability Does the approach benefits from large scale scenarios, meaning that it may evolve to adapt to the environment?

Table 3 shows the classification of each approach and the reason for the each

classification value:

32

Table 3 – Approach Benchmark – Classification of Approaches Pro-Active Constructs Aggregation adaptability

ICT-I * Although behavioral aspects are out of the scope at base web-services constructs, they are carried out at higher levels.

3 * Services in the base constructs.

3 * Key design goal 5

ManBree * Although behavioral aspects are out of the scope at base service entity constructs, they are carried out at higher levels.

3 * Services + Attributes in the base construct * Aggregation is a keyword

4 * Service entities wrap interoperability constrains

2

KIMM’s framework

* Negotiation behaviour considered

4 * Not considered 1 * Engineering Service Server hides interoperability constraints

5

PASEF * One of the key factors for PSEs

5 * Services + Attributes in the base construct

4 * PSEs wrap interoperability constrains

3

The final results of this benchmark exercise, after the normalization and weight of the

parameters, are presented in Table 4.

Table 4 – Approach Benchmark – distance to ideal values and relative proximity Approach D+ D- C

PASEF 0,0106 0,0269 0,7175

ICT-I 0,0138 0,0224 0,6191

ManBree 0,0195 0,0238 0,5500

KIMM’s framework 0,0245 0,0168 0,4078

The final results of the TOPSIS method applied to this benchmark exercise put PASEF

in the first place. This fact was expected, to some extent, because the selection of the

classification parameters was centred on the main limitations of current approaches,

which were indeed taken as the base for PASEF specification. The classification of the

other approaches was also expected. On one hand, KIMM’s framework is close to

PASEF in what concerns the multi-agent systems characteristics. On the other hand, the

ManBree approach is also close to PASEF, this time in what concerns the aggregation

factor. Nevertheless, despite these two intermediate classifications, the “second place”

goes to the ICT-I approach because, although it does not gather the “closest”

classification in any specific factor, it has a relatively strong classification in several

classification parameters. This fact has a particular magnitude in the adaptability

perspective, which is a classification factor where ICT-I had the best classification.

33

After ICT-I, Manbree benefits from the aggregation classification parameter where it

gets the higher classification. Finally, the KIMM’s framework is not far from the

ManBree approach because it gains in the pro-activeness factor along with PASEF.

6.2.2 – Solution Benchmark

Table 5 show the comparison parameters under which each solution was classified,

including PASEF.

Table 5 - Solution Benchmark - Comparison Parameters Name Description

Specificity How specific is the solution? Can it be applied to distinct economical consultancy areas?

Effectiveness How often do the services from Consultancy Service Providers become requested?

Collaboration Does the solution consider the collaboration between distinct consultants?

Table 6 shows the classification of each solution and the reason for each classification

value:

Table 6 – Solution Benchmark – Classification of Solutions Specificity Effectiveness Collaboration

lifeperson.com * Applied to many activity areas * 13 main activity categories

5 * Provides distinct communication channels * 1641 available expertise

4 * No collaboration between providers is considered

1

freelancer.com * Applied to many activity areas * 10 main activity categories

4 * Provides private message board communication channel * 425 average open projects

4 * No collaboration between providers is considered

1

PASEF * General, but especially used for senior professionals (in current implementation)

3 * Support bidding * Direct / Indirect Matching between projects and provided services.

5 * BPM support. * High-level client needs detailed in workflow BPMs by brokers

5

The final results of this benchmark are presented in Table 7:

Table 7 – Solution Benchmark result

34

Solution D+ D- C

PASEF 0,0095 0,0638 0,8700

lifeperson.com 0,0638 0,0095 0,1300

freelancer.com 0,0639 0,0048 0,0693

In this Benchmark exercise the final results put PASEF in the first position, far from the

other two solutions. Actually these results were also already expected because of two

reasons: 1 – the selected comparison parameters were the base problems that originated

PASEF; 2 – PASEF is the only solution considering a collaboration factor among the

members of the community using the solution.

Although this benchmark exercises suffers from the limitation concerning own

judgement, as “one cannot do a thing that he is a proper judge of it” (Oscar Wild), an

effort was made to overcome this limitation, namely through the usage of a

straightforward and transparent benchmarking method.

7 – Conclusions and Future Work

The adoption of Service Oriented Architectures is the most commonly used approach in

the development of platforms for collaborative networks, nowadays. Nevertheless, a

number of limitations exist, namely in what concerns the distinct understandings that

business and ICT perspectives have regarding the notion of service. The proposed Pro-

Active Service Ecosystem Framework provides a contribution to the reduction of the

aspects that separate these two perspectives. Furthermore, the current “static” approach

of Web Services, on one hand, and the inexistence of aggregation of distinct Web

Services from the same provider on the other hand, have also been two driving forces

for the creation of the PASEF framework. With the Proactive Service Entity concept,

distinct services that a CBE member can provide are represented within a single

“ambassador” construct which actively pursues business objectives, instead of waiting

for a client initiative.

On the other hand, the auto-initiative basis of PASEF tackles the problem of

catalogues becoming outdated in certain scenarios. In other words, the proposals / bids a

broker receives concerning a given business need have the guarantee of being up-to-date

because they are made directly by the providers, through their PSEs. In the case of

traditional service catalogues, it may happen that the lists of available services become

outdated, whenever a given provider becomes unavailable.

35

As future work, towards making the concepts included in PASEF better support

real-world scenarios, the development of complementary systems is needed, namely in

what concerns the management of accounting and contracts. Another future extension,

this time improving PASEF, concerns the creation of a repository of “most commonly

used Business Process Model Templates”.

Acknowledgments. This work was supported in part by FCT (CTS multiannual

funding) through the PIDDAC Program funds, funded by the European Commission. It

is also supported in part by the European funded research project “BRAID – Bridging

Research in Ageing and ICT Development”.

References

Abe, T., 2005. What is Service Science. Tokyo, Japan: Fujitsu Research Institute, Economic Research Center.

Alfaro, J. J., Rodriguez Rodriguez, R., Verdecho, M. J., & Ortiz, A., Business Process Interoperability and Collaborative Performance Measurement. International

Journal of Computer Integrated Manufacturing, 22(09), 877-889.

Baldo, F., Rabelo, R., & Vallejos, R., 2008. Modeling Performance Indicators' Selection Process For Vo Partners' Suggestions. In: A. Azevedo (Ed.), Innovation in

Manufacturing Networks. Springer Boston, 67-76.

Bellifemine, F., Bergenti, F., Caire, G., & Poggi, A., 2005. Jade — A Java Agent Development Framework. In: R. Bordini, M. Dastani, J. Dix & A. Fallah Seghrouchni (Eds.), Multi-Agent Programming. Springer US, 125-147.

Camarinha-Matos, L. M., & Afsarmanesh, H., 2008. Classes of Collaborative Networks. In: G. D. Putnik & M. M. Cruz-Cunha (Eds.), Encyclopedia of Networked and

Virtual Organizations. 193-198.

Camarinha-Matos, L., & Afsarmanesh, H., 2010. Active Ageing Roadmap – A Collaborative Networks Contribution to Demographic Sustainability. In: L. Camarinha-Matos, X. Boucher & H. Afsarmanesh (Eds.), Collaborative

Networks for a Sustainable World. Springer Boston, 46-59.

Camarinha-matos, L. M., & Afsarmanesh, H., 2011. Active Aging with Collaborative Networks. Technology and Society Magazine, IEEE, 30(4), 12-25.

Cardoso, T., & Camarinha-Matos, L. (2011). Proactivity in collaborative Services

Ecosystems. Paper presented at the PRO-VE - Working Conference on Virtual Enterprises, São Paulo, Brazil.

36

Chenghua, D., & Weston, R. (2007, 23-25 May 2007). Role Based Modelling Approach

to Designing Multi-product Systems. Paper presented at the Industrial Electronics and Applications, 2007. ICIEA 2007. 2nd IEEE Conference on Industrial Electronics and Applications.

Chesbrough, H., & Spohrer, J., 2006. A research manifesto for services science. Communications of the ACM - Services science, 49(7), 35-40.

Elhabib, N., Boucher, X., & Peillon, S., 2010. Engineering of Service Oriented Collaborative Network. In: L. Camarinha-Matos, X. Boucher & H. Afsarmanesh (Eds.), Collaborative Networks for a Sustainable World. Springer Boston, 461-468-468.

Ferrario, R., & Guarino, N., 2009. Towards an Ontological Foundation for Services Science. In: J. Domingue, D. Fensel & P. Traverso (Eds.), Future Internet – FIS

2008. Springer Berlin / Heidelberg, 152-169.

Ferrario, R., Guarino, N., & Fernández-Barrera, M., 2011. Towards an Ontological Foundation for Services Science: The Legal Perspective. In: G. Sartor, P. Casanovas, M. A. Biasiotti & M. Fernández-Barrera (Eds.), Approaches to

Legal Ontologies. Springer Netherlands, 235-258.

Fornasiero, R., Berdicchia, D., Zambelli, M., & Masino, G., 2009. Methodologies for Active Aging in the Manufacturing Sector. In: L. Camarinha-Matos, I. Paraskakis & H. Afsarmanesh (Eds.), Leveraging Knowledge for Innovation in

Collaborative Networks. Springer Boston, 733-741.

Franco, R., Ortiz Bas, Á., & Lario Esteban, F., 2009. Modeling extended manufacturing processes with service-oriented entities. Service Business, 3(1), 31-50.

Grant, E. B., & Gregory, M. J., 1997. Tacit knowledge, the life cycle and international manufacturing transfer. Technology Analysis & Strategic Management, 9(2), 149-162.

Horn, P., 2005. The New Discipline of Services Science. Business Week, January 21, 2005.

Hwang, C.-L., Lai, Y.-J., & Liu, T.-Y., 1993. A new approach for multiple objective decision making. Computers & Operations Research, 20(8), 889-899.

Jammes, F., & Smit, H., 2005. Service-oriented paradigms in industrial automation. Industrial Informatics, IEEE Transactions on, 1(1), 62-70.

Jammes, F., Smit, H., Lastra, J. L. M., & Delamer, I. M. (2005, 19-22 Sept. 2005). Orchestration of service-oriented manufacturing processes. Paper presented at the Emerging Technologies and Factory Automation, 2005. ETFA 2005. 10th IEEE Conference on.

Karnouskos, S., Baecker, O., de Souza, L. M. S., & Spiess, P. (2007, 25-28 Sept. 2007). Integration of SOA-ready networked embedded devices in enterprise systems via

a cross-layered web service infrastructure. Paper presented at the Emerging Technologies and Factory Automation, 2007. ETFA. IEEE Conference on.

37

Karnouskos, S., Savio, D., Spiess, P., Guinard, D., Trifa, V., & Baecker, O., 2010. Real-world Service Interaction with Enterprise Systems in Dynamic Manufacturing Environments

Kuk, S. H., Kim, H. S., Lee, J.-K., Han, S., & Park, S.-W., 2008. An e-Engineering framework based on service-oriented architecture and agent technologies. Computers in Industry, 59(9), 923-935.

Maglio, P., Srinivasan, S., Kreulen, J., & Spohrer, J., 2006. Service systems, service scientists, SSME, and innovation. Commun. ACM, 49(7), 81-85.

Maglio, P., & Spohrer, J., 2008. Fundamentals of service science. Journal of the

Academy of Marketing Science, 36(1), 18-20.

OSGi. (2007). OSGi Service Platform Core Specification, Version 4.1: OSGi Alliance.

Paulson, L. D., 2006. Services Science: A New Field for Today's Economy. Computer,

39(8), 18 - 21.

Phaithoonbuathong, P., 2010. Web Services- based Automation for the Control and Monitoring of Production Systems. International Journal of Computer

Integrated Manufacturing, 23(02), 126-145.

Preist, C., Esplugas-Cuadrado, J., Battle, S., Grimm, S., & Williams, S., 2005. Automated Business-to-Business Integration of a Logistics Supply Chain Using Semantic Web Services Technology

Rabelo, R., Gusmeroli, S., Arana, C., & Nagellen, T., 2006. The Ecolead ICT Infrastructure For Collaborative Networked Organizations. In: Network-Centric

Collaboration and Supporting Frameworks. Springer Boston, 451-460.

Ribeiro, L., Barata, J., & Colombo, A., 2009. Supporting agile supply chains using a service-oriented shop floor. Engineering Applications of Artificial Intelligence,

22(6), 950-960.

Shen, W., Hao, Q., Yoon, H. J., & Norrie, D. H., 2006. Applications of agent-based systems in intelligent manufacturing: An updated review. Advanced Engineering

Informatics, 20(4), 415-431.

Shen, W., Hao, Q., Wang, S., Li, Y., & Ghenniwa, H., 2007. An agent-based service-oriented integration architecture for collaborative intelligent manufacturing. Robotics and Computer-Integrated Manufacturing, 23(3), 315-325.

Souza, L., Spiess, P., Guinard, D., K, M., Karnouskos, S., & Savio, D. (2008). SOCRADES: a web service based shop floor integration infrastructure. Paper presented at the Proceedings of the 1st international conference on The internet of things.

Spohrer, J., Maglio, P. P., Bailey, J., & Gruhl, D., 2007. Steps Toward a Science of Service Systems. Computer, 40(1), 71-77.

Tarantilis, C. D., Kiranoudis, C. T., & Theodorakopoulos, N. D., 2008. A Web-based

38

ERP system for business services and supply chain management: Application to real-world process scheduling. European Journal of Operational Research,

187(3), 1310-1326.

Wollschlaeger, M., Geyer, F., Krumsiek, D., & Wilzeck, R. (2003, 16-19 Sept. 2003). XML-based description model of a Web portal for maintenance of machines and

systems. Paper presented at the Emerging Technologies and Factory Automation, 2003. Proceedings. ETFA '03. IEEE Conference.

Wooldridge, M., 1998. Agent-based Computing. Interoperable Communication

Networks, 1(1), 71-97.

Yang, B.-S., Kwon Jeong, S., Oh, Y.-M., & Tan, A. C. C., 2004. Case-based reasoning system with Petri nets for induction motor fault diagnosis. Expert Systems with

Applications, 27(2), 301-311.

Zeeb, E., Priiter, S., Golatowski, F., & Berger, F. (2008, 25-28 March 2008). A Context

Aware Service-Oriented Maintenance System for the B2B Sector. Paper presented at the Advanced Information Networking and Applications - Workshops, 2008. AINAW 2008. 22nd International Conference on.