|
Organizations are increasingly using the Internet and Grid to offer their
own services and to utilise the services of others. Presence of a wide
variety of services and resources creates new opportunities for providing
value added, inter-organizational services by composing multiple existing
services into new Composite Services (CSs). This naturally leads to resource
sharing across organizational boundaries. An inter-organizational business
relationship is commonly referred to as a virtual organization (VO). Whether
in the context of large-scale scientific experiments or eCommerce,
eGovernment or any other collaborative effort, organizations need
cost-effective ways of finding, purchasing and managing services performed
by other organizations. It is therefore necessary for organizations to be
able to set-up and manage business links with other organizations in a
rapid, dynamic and flexible manner. However Virtual Organisations blur the
distinction between outsiders and insiders, yet organizations forming a VO
wish to preserve their individual autonomy and privacy. A central problem in
VO management therefore concerns how organizations can regulate access to
their resources by other organizations in a way that ensures their
individual policies for information sharing are honoured. Regulating access
to resources by other organizations is difficult because the organizations
might not trust each other. Organizations will therefore require that their
interactions with other organizations be strictly controlled and policed.
In the paper-based world, businesses have been conducted using contracts.
The concept and the use of contracts are not new to today’s society. Legal
contracts can be traced back to ancient times. To form and manage VOs, we
need to emulate electronic equivalents of contract-based business management
practices; this means that relationships between organizations for
information access and sharing will need to be regulated by electronic
contracts, and then enforced and monitored.
In the Gridmist project, we have begun by investigating the types of
middleware services required. We will investigate whether it is possible to
determine a small set of generic services that can be used to support
arbitrarily complex interactions between organizations. Figure below shows
two basic interaction styles, where a cloud represents middleware services
shared between the organizations: (a) organizations trust each other
sufficiently to interact directly with each other; and (b) no direct trust
exists between the organizations, so interactions take place through trusted
third parties (TTPs) acting as intermediaries. A given contractual
relationship could be implemented by either interaction style. In a dynamic
setting, it may be that interactions initially take place through TTPs and
once sufficient trust has been established, organizations agree to interact
directly.
We have investigated how a contract can be represented electronically (an
executable version). Each entry in a contract is called a term or a clause.
The clauses of a contract stipulate how the signing parties are expected to
behave. In other words, they list the rights and obligations of each signing
party.
A right is an action that a signing entity can perform if it wishes to.
For example, a contract might stipulate that Alice, as a manager of
enterprise E1, has the right to send an offer to Bob, the manager of
enterprise E2.
Because this is a right, it is up to Alice to send or not to send the
offer to Bob; Bob need not be disappointed if he does not receive the offer.
Similarly, an obligation can be defined as a duty that an entity is expected
to perform.
A failure to perform such a duty means a breach of the contract. For
example, a contract might stipulate that upon receiving an offer to sell
from Alice, Bob has the obligation to reply to her with an OfferAccepted or
OfferRejected message.
Of course, there are no easy ways of automatically transforming a
contract written in a natural languge into an executrable version, but
several systematic, semi-automatic approaches have been proposed and are
under investigation. We have developed a systematic way of representing
contracts as finite state machines (FSMs) and described how rights and
obligations can be monitored and enforced.
Secondly, we have developed a composite service (CS) coordination system
that can be used for enactment of business processes of a VO represented as
workflows. The system provides the option of both centralised coordination
as shown in figure 2 and decentralised coordination, as shown in figure 3.
Should the coordination of the service be spread across multiple servers as
in figure 3, a higher level of fault tolerance is provided. In such cases,
each server makes the invocations of the constituent Web/Grid services for
its part of the CS, and communicates via a coordination protocol with its
peers to orchestrate the overall execution. Should a coordinating server
fail or leave, it is possible to move the CSs that server was coordinating
to another server. The cost of doing so is proportionate to the number of
CSs being coordinated and the complexity of those CSs.
Further work is required in integrating contract management. This work
will be undertaken with the help of future e-science projects.
top |
