In the middle 80s David Gelernter from the Yale University developed a programming language called LINDA. It is applied in the field of distributed environment development. Moreover, LINDA also describes a coordination model which is fully distributed in space and time. Based on its characteristic, although being used in parallel computing, it is restricted to a certain amount of servers while it cannot cope with adaptiveness and scalability in an open environment. As a solution SWARMLINDA is proposed based on a decentralized multi-agent system which got its inspiration by nature in the field of swarm intelligence. The ability of the architecture is characterized by a very scalable behavior. The system can grow to enormous size while still be very effective since the principle is based on only local interaction with the surrounding neighborhood. The behavior patterns are observed from natural individuals (aka swarms) like ants, birds, termites and bees. Each agent is characterized by simplicity, dynamism and locality. The research for the report has been performed in evaluating as well as examining properties, problems, behavior pattern and advantages of swarm intelligence used in a SWARMLINDA system. In particular, the work discusses involved mechanisms and describes the development of a SWARMLINDA. Further on, it defines different metrics being used in the system. A main part is based on the distribution of information objects in a 2D environment by forming clusters which hold similar objects. The cluster itself is defined by a spatial region containing several associative memories. Finally, the report closes by presenting plenty of test runs that have been executed on the SWARMLINDA system. In order to rate the performance an evaluation metric has been developed defining the spatial network entropy.