As networks grow in size, speed and flexibility, the role of network management becomes increasingly important. Recent developments in Internet technologies might provide the capabilities that are well suited to the solutions of some very challenging, outstanding problems in network management. Advances in network structures and technologies are increasingly dependant on the way in which these networks will be managed. The paper seeks to pierce the mystique which surrounds the recent developments in network management techniques and standards and to indicate their significance for new network technologies.
Networks today are managed typically through the use of powerful and general purpose monolithic management platforms. These to some degree of success provide integration of management tools, but pose a number of disadvantages:
1. Management platforms are expensive, both in terms of software as well as the cost of the hardware required.
2. They are typically complex do install, run and maintain.
3. Management platforms are based on the centralized paradigm of management: a small number of sites (typically a single one) collect data from the network and analyze them. This can create bottlenecks and thus delays in reacting to network problems; in addition, ability to scale becomes an issue.
4. In general, it is difficult to remotely access data and tools on the management platforms. The means available are typically primitive such as telnet, and do not match the capabilities offered on the consoles.
Recent emergence of Internet technologies such as the World Wide Web (Berners-Lee, 1992) and the Java language (Arnold, 1996) offers new means to overcome some if not all disadvantages of today’s network management platforms.
The roots of the network management protocols that are in common use these days are traced back to developments within the late Eighties. Before that point, network management had usually been performed using low-level signaling techniques to send special management info. Receipt of this info would cause receiving hardware to stop traditional operation and enter a special diagnostic mode during which it more responsible commands contained within the info. This approach worked well in homogeneous networks that used identical interface technology throughout. However, with the arrival of protocol stacks and abstraction of the lower level network characteristics, networks began to support multiple differing types of interface technology that meant that a unique approach was necessary so as to support network-wide management. At now, each the TCP/IP and Open System Interconnect (OSI) protocol stacks began to define network management protocols that operated at the appliance layer. This alteration in approach had each benefits and drawbacks. The foremost advantage was that management may currently be performed using identical tools at any purpose within the network; the foremost...