AbstractOperating telecommunication networks and services economically, reliably, and efficiently has been the major goal for every Administration and Operating Company since the very beginning of telecommunication systems. As soon as computers became available, applications aimed at increasing levels of efficiency and reducing the cost of operations were developed by mechanizing a number of commercial and administrative procedures as well as operations, such as billing, service order management, archive conversion to data files, etc.More recently, the introduction of more advanced network equipment (with increasing intelligence capacity), together with the availability of computers and peripherals providing attractive cost‐to‐performance ratios, has led to the development of systems designed to centralize and mechanize a variety of network operations and management activities. To date, the growth in the number of these two classes of support systems has been,instance‐driven, as they have been thevertical and dedicatedresponses to specific needs generated in individual sectors of the company organizations.Today, a number of new factors are strongly and urgently leading towards the evolution of the Operations Support Systems (OSS). In fact, while telecommunications networks are dramatically evolving in terms of technological levels or complexity, there is an even more rapidly increasing demand for high quality custom‐tailored services.The integration of support systems to meet the progressive integration of networks and services has therefore become essential in order to meet customer needs quickly and flexibly, adjust to the company's organization without major overhauls, reduce operation costs and ensure the desired quality levels so as to maintain or stay on the leading edge of technology.The evolution towarddata‐drivenarchitectures for the operations support systems is certainly not an easy task, made still more difficult by the requirement of preserving the existing systems, as much as possible, for both operational and economic reasons. Therefore, having defined the integrated OSS target by a top down approach, the practical way adopted to accomplish the integration process must follow a bottom‐up approach where the present reality is taken into account with the possible and predictable evolution. Naturally, one must also consider the trade‐offs between investments and derived benefits.This approach, aimed at progressively interconnected OSSs, will minimize the investments, by taking the maximum advantage out of them, and should allow a fine, progressive tuning of the Operating Company needs. The interconnectivity issues may be satisfied by the overlay of a new network, the Telecommunication Management Network ‐ TMN ‐, onto the Telecommunication Network. In this subject, the efforts of the International Standard Organizations are concentrated on the definition of the interfaces and communication protocols between the Network ElementS (NE) in the Telecommunication Network and the Operations Systems (Os) in the TMN. The implementation of an overlayed TMN is the only viable approach for the short ‐ medium term, but new architectures may become possible as switching systems evolve. In the long term, we may come to a point where these two networks, the support TMN and the telecom network, collapse into a single Information Network. However, the medium term evolution of the TMN networks, as it is suggested by the worldevolution, could be nothing else but a completion of the ideas of integration for the current systems. In the 92‐96 we will witness the application of the standards being approved or studied today.In a more distant future, new architectures will become available for telecommunication systems providing higher levels of flexibility, high speed links and high capacity distributed data bases.The ability of operations support functions to keep pace with the high speed, software based, environments of the future in an automated fashion, may be enhanced by new system technologies both in software concepts and in hardware constructs. For example, the structured, process oriented, datadriven, open system architecture should provide the flexibility to adjust to the dynamics of high speed networks and to the rapid introduction of software services developed by the Operating Companies. This structured approach should minimize the distruptions to the overall, complex, integrated, automated operations support functions while maintaining, without human interventions, maximum automated flow‐through of the functions. It should also enhance the ability of evolving to new software design concepts such as knowledge engineering.A look at the picture of future computing systems indicates that the movement of parallel processing techniques from academy to business applications, the increasing capabilities of database machines, and the advancing neural‐network horizons for expert systems engines, will all be active pursuits into the twenty‐first century. Flexible and Open Systems Architectures will enhance the Operating Company capabilities to economically adopt these emerging software and hardware developments.This paper addesses first some general concepts on OSS, the features of the present situation, with particular reference to Italy, and the rationale behind the present pressure for interconnecting and integrating the OSSs. How to identify the evolutive steps towards this integration and how to support the progressive implementation of the TMN is then discussed, together with the status and the perspectives of the standardization studies in progress in many European and International Bodies. Even if the attention is focused on the predictable evolution within the next decade, a quick look at the following years is also given, trying to discuss the possibility that the two networks, the support TMN and the telecom network, collapse into a