Integrated management solutions based on standardized management architectures aim to support the network and service providers' efforts in maintaining a high degree of quality and availability of their services while decreasing the cost of running the information technology infrastructure. Although providers can choose their management solution from a large number of implementations, the increasing complexity and heterogeneity of distributed systems still represents a major challenge for the Operation, Administration, Maintenance and Provisioning (OAM&P) of large-scale public and corporate networks.
From a management point of view, the situation has become more complicated due to the implications of telecom deregulation: Nowadays, an enterprise is not only free to choose between numerous IT service providers and network carriers but also needs to verify the QoS of the subscribed services. Furthermore, the role of the Internet Protocol (IP) as ``lingua franca'' for worldwide data communication and the layering of its standardized services and protocols (e.g., EMail (SMTP), Domain Name Service (DNS), World Wide Web (HTTP)) present opportunities of outsourcing IT activities for cost savings. This leads to layered service provider hierarchies as depicted in the left part of Figure 1 where a service provider is at the same time the customer of another provider: The provisioning of end-user services like electronic mail, WWW and DNS therefore depends on the availability of the IP service which, in turn, depends on, e.g., an ATM service. The problem domain dealing with the fulfillment of service-level agreements (SLA) between service providers and their users is usually known as Customer Service Management and is currently a subject of ongoing research (see e.g. [14]).
In this paper, we will concentrate on the right part of Figure 1, i.e., the question how the Management Systems (MSs) of the service providers and their customers can cooperate effectively. Today, it is very likely that these systems cannot interoperate seamlessly because different service providers usually deploy MSs that differ not only by their vendor (IBM/Tivoli, HP, CA), but also by the underlying management framework (OSI/TMN, Internet, CORBA, Java/WWW, proprietary). Furthermore, when these systems were initially purchased, there was often no need to exchange management information with peer MSs operated by another authority. This situation has changed: End users now check the quality of their subscribed services by retrieving data with WWW browsers from SNMP-based MSs and SNMP-based MSs that initially controlled the local area networks of an enterprise need to exchange management information with TMN-compliant MSs surveying long-haul telephony links etc. As MSs are the point of control for services and networks, management policies have to be enforced and surveyed by them. All this makes MSs crucial for successful enterprise management. It is therefore not only necessary to (re-)configure MSs at runtime but also to control whether they are working properly. We refer to these duties as ``Managing the Management''. This paper presents a CORBA-based approach to tackle this problem.
The above discussion shows that two questions must be answered to ensure the successful deployment of MSs in an service markets:
Questions relating to the design of a Management Information Base (MIB) for MSs are: How is the management model structured? Can we reuse object classes specified by standardized reference models as base Managed Object Classes (MOCs) for our model? Section 4 describes the principles that guided us in designing our management model and how we made use of RM-ODP concepts. We also describe the impact of CASE technology for the design of the MIB. Section 5 deals with our implementation and analyses how the prototype uses general-purpose distributed object technologies like CORBA and Java. Section 6 concludes the paper and presents issues for further research.
