CONTRACT SERVICES FOR POST-DISCOVERY GUARANTEE

MANAGEMENT

Josef Spillner, Bastian Buder, Torsten Schiefer and Alexander Schill

TU Dresden, Nthnitzer Str. 46, 01062 Dresden, Germany

Keywords:

SOA, SLA management, SLA management, Contracts.

Abstract:

Contract services let service providers and consumers manage service level agreements (SLAs) and other

contractual documents. These services let users negotiate SLAs, select tariffs, submit selected service offerings

to auctions, view and manage running contracts and ﬁnally submit feedback about the satisfaction with the

usage of contract-bound services. We deﬁne a number of contract services speciﬁcally for consumers which

run as platform-level services with a common associated user interface called Contract Wizard. The increased

interactivity over purely autonomous approaches is crucial for a wider acceptance of contract-bound service

execution.

1 INTRODUCTION

Service discovery is the process of matching func-

tional and non-functional consumer requirements

against provider offers. Many interesting algorithms

for the matchmaking exist, but most descriptions stop

at the interesting question: Once a service is found,

what can a consumer do with it? Technically, service

discovery returns a service description and a commu-

nication endpoint, both of which can be used to in-

voke a service from another service or, if the search

was initiated interactively by the consumer, integrate

associated user interfaces into the client system or

render a form dynamically. Sometimes, however, ad-

hoc invocation is not desirable. Instead, guarantee

terms shall be negotiated so that the service consumer

can rely on the offered properties.

Contracts can be established between service

providers and service users to obtain veriﬁable and

legally valid records of speciﬁc agreements regarding

service quality or other potentially conﬂicting goals

on the technical, business and legal level. Accord-

ing to (Truong et al., 2008), contract management

involves several phases, including speciﬁcation, ne-

gotiation and monitoring. Other approaches add re-

negotiation and feedback as additional aspects. Usu-

ally, research on this management is restricted to au-

tonomous agents. However, in order to let the con-

cept of contracts gain acceptance in service-oriented

computing, concepts need to be developed to let users

(both providers and consumers of services) control

each phase.

In our approach, the web application Contract

Wizard handles all contract-related matters from a

consumer’s point of view, including the initial nego-

tiation of non-functional properties (NFPs), selection

of tariff options and handover to auctioning platforms

in the negotiation phase. The monitoring and feed-

back phases are supported with a contract visualisa-

tion and management tool and the collection of feed-

back regarding the satisfaction with the contract, re-

spectively. The contract services implementing this

functionality are seen by us as a set of useful, stan-

dardisable parts of service brokering architectures. In

this paper, we ﬁrst describe the usefulness of contract

services and report on existing approaches of contract

management. We present a reasonable set of them

on a conceptual and logically linked level. To illus-

trate the concept, we then show to which extent our

current prototype already improves the user experi-

ence on service marketplaces. Finally, we report on

open issues which need to be solved before contract

services can eventually replace inﬂexible manual con-

tract management approaches.

2 PROBLEM STATEMENT

In next-generation networks, service usage contracts

– also known under the acronym SLA, for Service

Level Agreements – are said to play an important role

369

Spillner J., Buder B., Schiefer T. and Schill A. (2009).

CONTRACT SERVICES FOR POST-DISCOVERY GUARANTEE MANAGEMENT.

In Proceedings of the 4th International Conference on Software and Data Technologies, pages 369-375

DOI: 10.5220/0002285903690375

 SciTePress

to ensure a high quality as well as technical, ﬁnan-

cial and juridical reliability (Braun et al., 2008). Such

contracts usually contain clauses about minimum or

maximum average values of technical characteristics,

tolerance ranges, usage constraints and other restric-

tions. In addition, they may contain compensation

rules. Finally, any contract will have a set of metadata

such as the unambiguous identiﬁcation of the partici-

pating partners – or parties, in legalese –, the duration

of validity, associated tariff models and disclaimers.

Without the metadata, the contract would just be an

agreement without legal relevance. Often, the term

SLA is used to refer to purely technical contracting

approaches and Quality of Service (QoS) guarantees.

On the other hand, formal service interface deﬁni-

tions as described by the Web Service Description

Language (WSDL) or the older Interface Description

Language (IDL) are also named contracts in the con-

text of software engineering. Our understanding of

contracts requires the mutual acceptance of an SLA

by at least two involved parties.

Several declarative languages for expressing agree-

ments and contracts have been created over the last

decade. Among them, WS-Agreement, WSLA and

SLA-ng have been subject to a considerable amount

of analysis, critique and proposed improvements

(Tian et al., 2004; Frankova et al., 2006; Truong et al.,

2008). A number of service execution platforms have

been speciﬁcally developed or extended to exploit the

contents of such contracts and act accordingly in de-

cisive situations where the contracts conﬂict with the

state of the platform. Depending on the nature of

the situation, either the contracts can be modiﬁed by

renegotiating or cancelling them, the service execu-

tion can be modiﬁed by terminating, reconﬁguring,

rebinding or otherwise adapting services, and the plat-

form can be modiﬁed by acquiring more resources,

migrating workload to other servers or equivalent ac-

tions which help solve the conﬂict. These technical,

corrective actions belong into the functionality of all

adaptive platforms for contract-bound service execu-

tion. The term Service-Level Management (SLM) en-

compasses adaptivity but especially includes changes

of a contract within its lifecycle as well as admission

control functionality to reduce the need for adaptivity.

An approach to SLM is proposed by (Wang et al.,

2007) and refers to technical QoS management and

admission control using a custom QoS speciﬁcation

language. Some of its goals like end-to-end QoS guar-

antees are feasible in closed environments but unsuit-

able for Internet-scale service delivery. Furthermore,

no contract lifecycle guidance for users is offered by

the system. In (Ludwig and Franczyk, 2006), current

SLA languages are compared and evaluated for suit-

ability for automated contract management in service

grids. The authors conclude that several ambiguities

are present in all of them which makes automated ap-

proaches hard or impossible for more complex condi-

tions. When elastic infrastructures are used beneath a

user-oriented participative Internet of Services (pIoS),

user guidance through contract services becomes even

more important.

While the autonomous handling of contracts is suf-

ﬁciently covered by existing approaches, additional

SLM aspects are needed for conveying the conﬂicts

and decisions to the user in an intuitive and compre-

hensible way. No previous concepts are known at this

point which could ﬁll this gap. Hence, we introduce

dedicated, reusable contract services as a novel con-

cept. Furthermore, we propose Contract Wizard as

an intuitive user interface to contract services in the

context of pIoS.

3 CONTRACT SERVICES

Given that the processes of creating, storing, man-

aging and terminating contracts encompass multiple

steps, we propose a set of individual contract services

which are loosely coupled so they can be used both

as standalone platform services and as a composite

service which manages all of the contractual aspects.

The research is focused on initial, user-driven ne-

gotiation of SLA parameters (technical perspective),

tariff selection and auctioning (business perspective),

contract monitoring and status visualisation, interac-

tive feedback collection and the adherence to contract

law (legal perspective) in any of these steps. We do

not consider contract services aimed at providers who

place service offerings at marketplaces in this paper,

although it can be assumed that the basic functional-

ity would be comparable. Figure 1 shows the overall

picture.

3.1 Initial SLA Negotiation

The scope of the initial negotiation is the user-driven

transformation of a service description or an SLA

template into an SLA. A set of templates is attached to

a service description within the service registry, from

which the user may choose one to initiate the negotia-

tion process. If no templates are present, an uncon-

strained negotiation can be initiated. The resulting

SLA will be linked to the user to form the base for

a valid contract between two parties. The term ini-

tial serves to differentiate this step from autonomous

renegotiations which may happen once the SLAs have

become binding. It is assumed that only NFPs are

ACT4SOC-EHST 2009 - 4th International Conference on Software and Data Technologies

370

Figure 1: Contract relationships between pIoS hosters, ser-

vice producers and consumers.

subject to negotiation, while functional properties are

ﬁxed from the moment of service selection. The ex-

tension to allow the contract peer to ﬁnd function-

ally equivalent or similar services is not considered

by us. The negotiation service runs as an agent act-

ing on the user’s behalf and representing the user’s

will. The negotiation peer would usually be pro-

vided by the service execution environment. In sev-

eral projects, an SLA manager agent assumes this

role. The transformation uses form generation tech-

niques to create a user interface representing the pos-

sible options and NFPs for negotiation. It contains

ﬁxed values as well as variable properties which may

be altered by the user. Alteration can happen either

on a per-contract basis using roundtrips to the negoti-

ation peer, or for more sophisticated needs and more

intuitive understanding of inter-property dependen-

cies on a per-property basis. Property ranges may be

initially restricted and preset by importing the user’s

preferences from the service discovery process. At

the moment, we are not aware of a standard ﬁle for-

mat for expressing and transmitting user goals. We

propose to leverage the prioritised requirement/offer

match from the discovery, which may be expressed as

a semantic query in the Web Service Modelling Lan-

guage (WSML). The preferences are expressed as ei-

ther a quadruple consisting of property name, value,

unit and comparison operator. Each preference item is

thus transformed into a service level objective (SLO)

in the contract as shown in ﬁgure 2.

3.2 Tariff Selection

Tariffs regulate how much a user has to pay to ac-

cess a service. Every tariff is based on an economi-

cally viable pricing model which aims at maximising

Figure 2: Preference-inﬂuenced negotiation example for re-

sponse time (rt).

the income for the service provider while keeping the

price competitive enough to not drive potential and

current users to competitors. In addition, there is a

mutual relationship between the applicability of tar-

iffs and the set of outcomes of negotiation processes.

For example, selecting a low-budget tariff restricts the

guarantees on response time and encryption strength,

whereas negotiating an agreement of near-100% up-

time will lead to the expulsion of low-budget tariffs

in subsequent tariff selections. Due to the loose cou-

pling of contract services, either order of initial nego-

tiation and tariff selection will be possible. Techni-

cally, either a reference to a tariff or the tariff terms

themselves are added to the SLA.

3.3 Contract Monitoring and

Management

While the creation of a contract only takes a lim-

ited amount of time in the range of minutes up to a

few days, the contract validity period can often reach

months or years. During this time, the user needs

to stay informed about the set of running contracts

(passive control). There should also be facilities for

modifying, terminating or otherwise managing con-

tracts (active control). It is generally expected that

the user-centric Internet and user-driven actions like

service composition or content creation will increase

rapidly in the pIoS. Giving information and control

back to the user is mandatory for the broad acceptance

of these concepts.

The information gathered by the various moni-

toring sources in service execution environments can

be used to present contract status information to the

user. In particular, generated graphs related to ne-

gotiated properties as well as to the overall con-

tract compliance of the service execution are of great

help to decide whether a renegotiation or cancellation

will be beneﬁcial. Further information such as de-

tailed lists of incidences can be retrieved through a

Monitoring-as-a-Service (MaaS) interface to the mar-

ketplace monitoring database. Of course, the trans-

parency and level of detail for such views can be re-

CONTRACT SERVICES FOR POST-DISCOVERY GUARANTEE MANAGEMENT

371

stricted at the discretion of the hosting operator. There

are various reasons for such restrictions, including the

fear for providing hints on the used infrastructure to

competitors and the risk of leaking data due to tech-

nical glitches. For the purpose of our research, we

assume full access of users to data directly related

to their contracts. Some of the features will hence

not apply if restrictions are in place. Key relevant

information displayed to the user ideally includes a

timeline of values per monitored property, a timeline

of incidents linked to explanations by the provider

and statements about whether the user’s speciﬁc con-

tract has been affected, payment status information

and overall average and current values. In practice,

it again depends on provider policies of how much of

this information will be available to the user.

3.4 Feedback Collection

While monitoring data gathered from service execu-

tion will be objectively accurate, there might always

be situations in which the user’s perceived quality of

a service deviate from the measured value. A classic

example is a fully-operational host with a broken con-

nection to the user’s computer. In such cases, addi-

tional feedback from the user regarding the fulﬁlment

of the contract by the provider can be very helpful and

might indeed protect the provider from greater dam-

age of having to compensate a number of consumers

when problems propagate. Such ratings provide in-

novative value over traditional, contract-agnostic ap-

proaches to automatic and user feedback on service

execution (Maximilien and Singh, 2002).

3.5 Adherence to Law

In order to turn an agreement (or SLA) into a widely-

recognised contract, the peculiarities of contract law

need to be considered for the whole process of con-

tract creation, modiﬁcation and termination. Even in

legally harmonised systems like EU member states,

the acceptance and legal relevance of qualiﬁed digital

signatures, electronic contracts and terms of service

varies greatly. Therefore, we only consider the legal

issues for a limited number of countries in our work

on contract services. The issue is complicated further

by the nature of consumers as either business users or

private consumers. Already identiﬁed requirements

encompass the prominent display of the court loca-

tion of the provider company and double conﬁrma-

tion according to the Rome Convention of 1980 and

its successor e-commerce laws, the unambiguous act-

ing of Contract Wizard on behalf of and in full coop-

eration with the user, and the possibility to print con-

tracts for postal delivery in case electronic contracts

are not recognised by law (Parrilli, 2008). Users must

be able to modify all parts of SLA offers before ﬁ-

nalising them. Contracts must explicitly inform about

potential renegotiation attempts in the future. Detail-

ing these requirements and adding further legal safe-

guards is an ongoing process in the deﬁnition of use-

ful and legally viable contract services. We plan to

report on the outcome of these activities based on a

cooperation and continuous review from researchers

in the area of contract and civil law.

4 PROTOTYPICAL EVALUATION

We have implemented all mentioned contract services

and consolidated them into a single application called

Contract Wizard. The name reﬂects the user-centric

step-by-step guidance nature of the application. In ad-

dition to the application itself, this section is going to

present major integration points with additional bro-

kering components, namely the service discovery for

the initiation of contract creation, as well as the back-

end services for all of the steps in Contract Wizard:

The pricing models and corresponding pricing model

repository, auctioning and brokering platforms, the

usage feedback service, the monitoring history ser-

vice, and ﬁnally the SLA manager. The architecture

and workﬂow is shown in ﬁgure 3.

4.1 Contract Wizard

Contract Wizard is typically invoked in the match-

making or brokering phase of any service usage life-

cycle, directly after the service discovery step and

before the service usage phase. It also extends into

the usage and feedback phases. It is located on user-

focused portals. The use of this application is aimed

at service consumers. The application itself is imple-

mented with eRuby and Java/JSP. It acts as a shell

containing page modules for all the possible steps dur-

ing the creation and lifetime of a contract. Each page

module interacts with a platform service, which is in

turn connected to a repository or another web service.

Page modules can be skipped if no choice is avail-

able. In particular, service providers may opt to not

offer conﬁgurable SLA levels or tariffs, or only in-

clude ﬁxed SLA templates without negotiable parts.

In parallel to the active page module, Contract Wiz-

ard permanently displays a list of running contracts

for easy access. The ﬁgures 4 and 5 demonstrate typi-

cal negotiation controls as seen by a user and a result-

ing SLA section.

ACT4SOC-EHST 2009 - 4th International Conference on Software and Data Technologies

372

Figure 3: Contract document workﬂow within Contract Wizard.

Figure 4: Screenshot of initial negotiation in Contract Wiz-

ard.

ws ag : G uar ant e eT e rm N am e = " A vai lab ili ty " {

ws ag : S erv ice Sco pe S e rv i ce Nam e = " Ex Ser vic e " ;

ws ag : S e rvi ceL e ve l Obj e ct i ve {

ws ag : K PIT ar g et {

ws ag : K PI Nam e " A vai lab ili ty " ;

ws ag : C u st o mSe rvi c eL e vel " 90 " ;

}} }

ws ag : G uar ant e eT e rm N am e = " R eli ab i li t y " {

ws ag : S erv ice Sco pe S e rv i ce Nam e = " Ex Ser vic e " ;

ws ag : S e rvi ceL e ve l Obj e ct i ve {

ws ag : K PIT ar g et {

ws ag : K PI Nam e " A vai lab ili ty " ;

ws ag : C u st o mSe rvi c eL e vel " 84 " ;

}} }

Figure 5: Resulting WS-Agreement SLO deﬁnition in com-

pact notation.

Both the initial negotiation and the feedback mod-

ules are driven by the NFPs offered by the service.

Our implementation supports conﬁgurable category

systems so that most users can opt to only conﬁgure

the most interesting properties, e.g., performance and

price, while advanced users can negotiate all proper-

ties separately.

4.2 Service Discovery

Contract Wizard provides a direct initiation mode

which requires the user to give it the URL to a service

description and contract template. Usually though,

service discoveries like ConQoMon link several ac-

tions to found services, including the option to invoke

the service directly and to negotiate a contract with

it (Stoyanova et al., 2008). Either way, the descrip-

tion and associated contract templates are fetched by

Contract Wizard through HTTP GET or passed to it

through HTTP POST and used to create the contract

creation session.

4.3 Pricing Models Repository

In order to increase the acceptance of autonomous

service hosting, rather complete pricing models and

taxonomies have been developed on an abstract level,

such as (Lehmann and Buxmann, 2008). This model

can be expressed in either XML or in ontology lan-

guages. Eventually, Contract Wizard’s tariff selec-

tion will be based on this model. Currently, an older,

less complex model derived from an implementation

called WS-Accounting (Br, 2008) is used for this pur-

pose. The tariffs are expressed in XML and stored

in a ﬁle system repository. We have implemented a

helper application called tariffguide which scans this

repository and is able to select the best tariff based on

criteria like the number of expected invocations. In

addition, a web page lists the tariffs and offers them

for selection and inspection of details. Both parts of

the implementation are considered to be reusable once

the more complex model will be used.

4.4 Auctioning Platforms

As soon as service users have several contracts with a

provider, the optimisation potential moves from the

modiﬁcation of all single contracts to the improve-

ment of the set of contracts. Multiple services might

be offered for less cost than the sum of all of them

when contracted individually, for example. In these

situations, Contract Wizard allows the user to sub-

mit a negotiated offer together with the preferred tar-

iff to an auctioning platform. Two such platforms

for multi-attributive and combinatorial auctions (van

Dinther et al., 2008) have been integrated on a con-

ceptual level. The integration efforts are currently still

CONTRACT SERVICES FOR POST-DISCOVERY GUARANTEE MANAGEMENT

373

ongoing and are based on HTTP references to con-

tract offers and tariffs. During the auction, the Con-

tract Manager does not proceed. Once an auction is

ﬁnished, the respective auctioning platform invokes a

re-entry page of Contract Wizard, therefore signalling

it the results of the auction and triggering it to pro-

ceed to the contract ﬁnalisation step. Due to the ex-

pected delay until a contract becomes ﬁnal and valid,

the auction will only be applied to contract creation

processes without realtime requirements. The combi-

natorial auction needs information about a set of ser-

vices. This is realised by having the discovery expose

the BPEL structure of composite service.

4.5 Service Usage Feedback

Explicit feedback and rating on the usage of contract-

bound services is collected from the user through an

automatically generated form. The contents of the

form are submitted to the usage feedback aggregator.

The form is shown in ﬁgure 6, with ﬁgure 7 explain-

ing the process.

Figure 6: Screenshot of completed rating form in Contract

Wizard.

Figure 7: SLA negotiation and feedback loop.

4.6 Monitoring History

For the purpose of integrating the monitoring results

into Contract Wizard, several possibilities exist. In

order to display lists and tables, our prototype can

connect to the Monitoring-as-a-Service interface and

query for relevant data such as daily average response

times. Displaying graphics, on the other hand, inte-

grates the results of visualisation components which

themselves use the MaaS interface. In most cases,

Contract Wizard embeds references to already gener-

ated graphs.

4.7 SLA Management

While the SLA negotiation part of Contract Wizard

turns an SLA template into an SLA offer, the task of

the SLA manager component is to evaluate the offer

and accept or reject it. The two services communicate

with each other using a negotiation protocol which is

the on-wire representation of a negotiation algorithm.

There are a number of algorithms around, but only

a subset can be used for this communication: Some

algorithms only exist on paper without or with only

insufﬁciently performing implementations. Likewise,

some algorithms are bound to the SLA format in use.

Avoiding these two subsets, the remainder is the junc-

tion of all SLA format-agnostic algorithms and WS-

Agreement-based algorithms. The default negotiation

mandated by WS-Agreement is known to be insufﬁ-

cient, though. In our work we closely watch the de-

velopment of a successor with more modular split be-

tween ﬁle format and negotiation protocol, while at

the same time looking into creating a generic negotia-

tion toolbox with pluggable algorithms. That way, we

can cover multi-phase negotiation, incremental con-

straints and other advanced algorithms.

5 OUTLOOK AND NEXT STEPS

In the future, we want to discuss the general idea

of contract services and the speciﬁc implementa-

tion Contract Wizard within the wider research com-

munity. There is generally a high interest to treat

SLAs as contractually valid contracts to enable the

semi-automatic creation and management of contracts

in real-world applications. Contract Wizard is al-

ready offered publically as part of our SOA plat-

form TECLA

, in which it cooperates with other

components, especially the service discovery. The

TECLA and Contract Wizard website:

http://texo.inf.tu-dresden.de

ACT4SOC-EHST 2009 - 4th International Conference on Software and Data Technologies

374

focus of TECLA is to provide a well-integrated,

open source platform which covers all important as-

pects regarding contractually guaranteed properties

and quality. TECLA includes additional compo-

nents, such as a dynamic invocation tool (Dynvoker),

a context- and quality-aware semantic service discov-

ery (ConQo) and another step-by-step web applica-

tion aimed at service providers who want to publish

services (Provider Wizard). We believe that both the

theoretical foundations for a real-world adoption as

well as a ready-to-use and freely available implemen-

tation are mandatory to raise awareness and accep-

tance for the idea of contract services.

6 CONCLUSIONS

Our research on establishing and managing con-

tracts between service users and service providers has

turned up a number of concepts for individual, yet

loosely-coupled platform services which we call con-

tract services. Aligned with the vision of a participa-

tive Internet of Services, we have identiﬁed the need

to incorporate these concepts into the marketplace and

link them to related research activities by various re-

search groups concerned with contract law, pricing

models, auctioning and brokering as well as SLA

management. Therefore, we have introduced Con-

tract Wizard, an essential user-visible service market-

place component for contract management. Its inte-

gration with semantic service discovery helps users to

ﬁnd suitable contracts based on initial preferences and

custom negotiation.

ACKNOWLEDGEMENTS

The project was funded by means of the German Fed-

eral Ministry of Economy and Technology under the

promotional reference “01MQ07012”. The authors

take the responsibility for the contents.

REFERENCES

Br, M. (2008). Steuerung der Nutzung von Web Services

durch Vertrge und Bezahloptionen. Thesis, Technis-

che Universitt Dresden.

Braun, I., Reichert, S., Spillner, J., Strunk, A., and Schill, A.

(2008). Zusicherung nichtfunktionaler Eigenschaften

und Dienstgte im Future Internet of Services. PIK

- Praxis der Informationsverarbeitung und Kommu-

nikation, 31(04/08):225–231.

Frankova, G., Malfatti, D., , and Aiello, M. (2006). Se-

mantics and Extensions of WS-Agreement. Journal

of Software (JSW), 1(01):23–31.

Lehmann, S. and Buxmann, P. (2008). Preisstrategien fr

Software- und Serviceanbieter. In Schriften zur Quan-

titativen Betriebswirtschaftslehre, number 4.

Ludwig, A. and Franczyk, B. (2006). SLA Lifecycle Man-

agement in Services Grid - Requirements and Current

Efforts Analysis. In Proceedings of the 4th Interna-

tional Conference on Grid Services Engineering and

Management (GSEM), pages 219–246. Erfurt, Ger-

many.

Maximilien, E. M. and Singh, M. P. (2002). Reputation

and Endorsement for Web Services. ACM SIGecom

Exchanges, 3(1):24–31.

Parrilli, D. (2008). SLAs and Legal Issues. In Workshop on

Infrastructure Services from a Business Perspective.

ServiceWave, Madrid, Spain.

Stoyanova, G., Buder, B., Strunk, A., and Braun, I. (2008).

ConQo A Context- and QoS-Aware Service Dis-

covery. In Proceedings of IADIS Intl. Conference

WWW/Internet. Freiburg, Germany.

Tian, M., Gramm, A., Ritter, H., Schiller, J. H., and Winter,

R. (2004). A Survey of current Approaches towards

Speciﬁcation and Management of Quality of Service

for Web Services. PIK - Praxis der Informationsver-

arbeitung und Kommunikation, 27(03/04):132–139.

Truong, H.-L., Gangadharan, G., Treiber, M., Dustdar,

S., and D’Andrea, V. (2008). On Reconciliation of

Contractual Concerns of Web Services. In 2nd Non

Functional Properties and Service Level Agreements

in Service Oriented Computing Workshop (NFPSLA-

SOC). Dublin, Ireland.

van Dinther, C., Holtmann, C., Setzer, T., Stage, A., and

Stathel, S. (2008). Auctions for Service Brokerage

in Business Value Networks. In Group Decision and

Negotiation. Coimbra, Portugal.

Wang, C., Wang, G., Wang, H., Chen, A., and Santiago

(2007). Quality of Service Contract Speciﬁcation, Es-

tablishment, and Monitoring for Service Level Man-

agement. Journal of Object Technology, Special Issue,

6(11):25–44.

CONTRACT SERVICES FOR POST-DISCOVERY GUARANTEE MANAGEMENT

375