2.3 Grid Computing and Virtual Organisations
2.3.5 Trust Issues in VO Functioning
During the functioning phase of the life cycle, the agents work collaboratively under a predefined agreement. They work towards a common goal, such as providing a composite service (that is an amalgamation of their individual services) to an end user.
When a VO is formed, the agents are bound by a legal contract, which typically requires them to conform to a particular behaviour or requires them to meet a set of definite expectations. This is reason enough to assume that under such restrictions it is not necessary to consider trust – everyone is likely to achieve what is expected. Additionally, calculating trust may be computationally expensive for an agent, and therefore it may decide that due to the presence of a contract binding the actions of others, there is no need to evaluate their trustworthiness. In this case, all agents may adopt a trusting disposition towards their fellow VO members.
However, if complete trust is assumed and trust values are not used in the VO, the question of whether an agent still needs to update the information about others, which it uses to calculate trust levels, needs to be addressed. It is necessary to evaluate the performance of other agents, with respect to trust, even in a closed system such as an individual VO. This is because bad partners may be selected in the formation phase, due to lack of information available when cal- culating the trustworthiness of agents in an open system (prior to VO formation). The continual assessment of the VO partners with respect to trust is essential so that an agent may make a more accurate partner selection decision next time a VO is formed.
Whilst a VO is functioning, its internal organisation may be analysed using social relations. For example, Figure 2.3 shows a VO that may be classified as having a hierarchical structure (with one agent at the top of the social structure) and one with a peer-based structure (where all agents are socially equal). Social relationships are not a prerequisite for agents to form VOs with each other, however certain relations may emerge between agents that regularly interact together. These relationships may lead to agents behaving in a particular manner towards each other, for example an agent may behave biased towards another with which it regularly forms a VO. Therefore, observing and learning the social relations that emerge from regular interaction is crucial in assessing the expected behaviour of an agent (both with respect to its behaviour inside and outside a VO). Social structure plays an important part in trust and trusting, we can see this from human societies. For example, an organisation is unlikely to take advice from a competing entity without considering the fact that the information it receives may be misleading. During the functioning of the VO there may be occasions when the VO fails to meet a require- ment that was agreed upon in the contact. From a trust perspective, it is important to isolate the individual or group of individuals within the VO that were the cause of this failure, due to the
A
B C D
E F
A B C
HIERARCHICAL STRUCTURE A is more important than B, C and D (which is more impor- tant than E and F).
PEER-BASED STRUCTURE A, B and C are all socially equi- valent. No one agent is more important than another.
FIGURE2.3: Examples of social structures in virtual organisations.
fact that agents in the system are likely to lose trust in the parties upon whom the blame falls. The identification of who to blame is outside the scope of this study2, however it is important to consider the way in which trust changes when blame is assigned (to an individual or group) and when it is unattributed. For example, a consumer of the service offered by a VO may reduce their trust in all the members of a VO if they cannot assign a blame for a fault in the service to a single or group of agents. Likewise, they may lose trust in only those that are assigned the blame for the fault.
From the analysis of the trust issues in the VO formation stage of the VO life cycle, we obtain the following requirements:
General Requirement 9
Maintaining interaction history — Regardless of an agent operating in a VO or
outside a VO (in the mix of free agents in the multi-agent system) it is necessary for the trust model to store the experiences of the agent for future trust calculations.
General Requirement 10
Use social information — Since social structures are present in VO environments,
and are important in calculating trust, it is important for the model to include social factors and information in the calculation of a trust value.
General Requirement 11
Dynamic trust value — The model should provide a means of changing a trust
2
value (that an individual agent calculates for another) based on the experience of an individual agent. More specifically it should dynamically change trust values based on the individual’s observations over time. Every observation should have a defined impact on trust.