Try to imagine a railway network that did not check its rolling stock, track, and signals whenever a failure occurred, or only discovered the whereabouts of its loΒ comotives and carriages during annual stock taking. Just imagine a railway that kept its trains waiting because there were no available locomotives. Similar thoughts could apply to any transport network of trucks, buses, or taxis covering a wide geographical area. It is quite clear that the quality of service and safety, and the cost efficiency of any network depend on the effective and timely manΒ agement of network resources. The same is true of telecommunications networks. For a long time now, telecomΒ munications networks have been designed and built with remote monitoring and control devices. The very dimensions of networks and the need to intervene rapΒ idly in the event of an incident made these devices imperative. On the other hand, the means of measuring quality of service did not appear till much later. Data transmission networks developed during the 1970s were often designed without any network management features. The technical traditions of the world of information technology had been forged on small systems with just a few dozen terminals spread over a small area. At the time, the extension of networks to cover wide geographical areas was not perceived as a change of dimension.