With the rising amount of networked devices in industrial, edge, fog and IoT applications, gathering, managing and presenting data from a multitude of distributed sources is an increasingly vital task. Computing systems that locate and centralize data from a subordinate network domain - complex gateways and aggregating servers - are frequently employed and described in this capacity. However, when such "aggregating systems"are encountered in research or industrial applications, one quickly realizes that functions and capabilities vary tremendously. The term appears to encompass anything from simple fieldbus gateways to intelligent information processing nodes for harmonizing information models. This article presents a classification scheme for aggregating systems, that is systems with the ability to integrate information from other, distributed devices into their own datastore. A particular focus lies on intelligent systems supporting OPC UA, a standard that is rapidly establishing itself as a key protocol for semantic communications. This article first introduces a general architecture model for aggregating systems and proceeds to detail functionality and cross-cutting concerns that effect these systems. Finally, the applicability of these models is tested on two case studies.