Scalability

  JDMK does neither provide standardized directory nor naming services. Features for achieving location transparency (like the CORBA Interoperable Object References) are also not available: An M-Bean is identified by a protocol identificator, the host address and port number, and its object name. These parameters and some knowledge about the registered beans in a given CMF must be present in order to make use of the M-Beans. The scope of the Metadata Service that can be used to retrieve information on registered M-Beans is limited to a single CMF, i.e., there is no global Metadata Service. Consequently, the only way of finding the currently active CMFs is the JDMK Discovery Service which sends a broadcast message that is answered by all running agents. The establishment of domains and the structuring of the agents in functional groups is not supported by the development environment and has to be done by the developer.

The conceptual weaknesses mentioned above impede the development of management applications for large IT infrastructures where a high number of different JDMK-based agents is needed. The JDMK services are useful for small, local environments where the amount and the degree of diversity of the agents are restricted. Due to the absence of focus on large systems, the scalability of JDMK-based solutions may be critical at the current stage of the toolkit.