A cooperative distributed systems approach is an appropriate design paradig
m for many applications, such as manufacturing control and network manageme
nt. A fundamental problem in this paradigm is how to manage the interdepend
encies among the entities. The solution to this interdependency problem is
called coordination. In this paper, coordination has been described in term
s of structure and mechanism. This paper focuses on the mechanism aspect of
the coordination. It is viewed as a composition of decision points and int
eraction devices directed to deal with different aspects of the interdepend
ency problem. This paper also provides a coordinated, intelligent rational
agent (CIR-Agent) model for cooperative distributed systems. The structure
and the architecture of the agent are based on its mental state regarding a
chieving a goal. The agent's structure consists of four main components: pr
oblem-solver, pre-interaction, interaction, and execution. There is no glob
al control of coordination allowed in the proposed model, for which the age
nts utilize different interaction devices to resolve their interdependencie
s.