Despite enormous technical developments on computer hardware and the resulting abilities in the last decades, many applications are still not and will never be computable by one machine alone. For collaboration, various machines are thus assembled to distributed systems; however, as these machines may be heterogeneous in hardware and operating systems, this integration shows to be a challenging task. The contribution of this diploma thesis is an enterprise Java implementation of the Peer Model, a data-driven model for collaboration of heterogeneous systems based on the concepts of Timed and Colored Petri Nets, introduced and specified by the Space Based Computing Group of the Institute for Computer Languages, TU Wien. One of the primary requirements of collaboration between machines is to enable communication between them. To that end, the thesis evaluates several serialization and communication formats and defines a platform-independent mechanism of instance discovery and data exchange between Peer Model instances. Thus, it allows to build up scalable and distributed solutions while not requiring the Peer Model instances executed on the collaborating machines being implemented in the same programming language. Furthermore, the implementation allows to add and remove entities during the runtime of the Peer Model instance and thus enables dynamic adaptions while being executed. The implementation is integrated into a developing toolchain composed of various tools and systems in the Peer Model context. Furthermore, a few diploma theses and one PhD thesis are already using or extending this implementation for their use cases.