next up previous
Next: Agents Architecture Up: Architecture for Automated Model Previous: Architecture for Automated Model

Overall Generation Process

  The following explanation of the process assumes that the implementation is already installed in the managed environment. Figure [*] depicts all steps to be carried out by the IT-managers and the involved algorithms during the most important parts of the model's `lifecycle'.


Steps of the Modeling Process  [r]








In the first step the administrator has to select all services, components, departments etc. that should be part of the model. These will later appear as the objects in the model. Of course, this selection highly depends on the management tasks the model will be used for. There are two variants, on which level the selection process can take place: Either (i') on the abstract (class) level, e.g., ``web servers'' and ``web clients'' followed by (possibly platform supported, automated) instantiation, to select the real objects belonging to the chosen classes. Or (i'') directly on the level of objects in the real world, e.g., web server on host a, web client on host b, etc. Note, that domains are represented as objects, too.

In step ii the matching components, applications etc. in the real world must be chosen for each object. This requires no special actions for simple objects, but only for those where the realization is dispersed over distinct real objects, e.g., distributed applications and also the previously mentioned domains. In our example, two main routers could be selected to represent the object communication infrastructure.

Step iii, where the appropriate probes to meter the objects' activity (as explained in section [*]) must be installed, is the last step of the installation phase. As for all measurements in distributed environments special care has to be taken on where to place the means of collection and the model generating algorithms; esp. in the TCP/IP world, where the management data is transferred `inband' through same channels as the user data. Therefore, the architecture presented in the following section supports different kinds of collectors, assuring its suitability for various environments.

During normal operation of the systems and networks, activity data is collected (step iv). After a certain period of time the information is transferred to places where a model can be generated (step v) using the method described in section [*]. From hereon management tools can start to use the generated model (vii,a)). Special management functions (vii,b))--as described at the end of section [*]--may be applied on sequentially generated models (step vi). For example, pointing out the differences to the administrator or generating alarms if heavy (or certain) changes occur.


next up previous
Next: Agents Architecture Up: Architecture for Automated Model Previous: Architecture for Automated Model
Copyright Munich Network Management Team