In this work we examine some cooperative features of hymenoptera and apply them to the management of physical components in a network. Cooperation within an ant colony ensures the population functions harmoniously, survives and evolves - desirable characteristics for the management of future networks. Our initial approach is to model cooperating (ant-like) code processes within a live communications network environment. The primary task being to gather and filter network management information at a local level, whilst distributing knowledge throughout the network entirety via inter-agent communication.
The work attempts to define a set of local strategies that allow a collection of autonomous distributed entities to impose global order within a managed network. The architecture consists of multiple autonomous mobile agents, each capable of moving within a restricted domain of the managed network. Through interaction, co-operation and information interchange the distributed entities organise to form a proficient and robust management architecture. Each domain is centred on a host-node and includes all other nodes directly connected to that host. Within the host node a Queen agent rules over the localised region and creates offspring in response to stimuli from her daughter agents who report on events within the domain. Daughters are spawned to carry out specific tasks, each with the capacity to communicate, to some degree, with their sister agents. This is an iterative and dynamic process that results in the creation of a global network management framework through an interacting local domain architecture where each Queen is aware of events occurring only in her domain.