After you determine the manageable aspects of your component and define the groups you will use to organize the attributes, you are ready to define the attributes according to the MIF. Understanding the MIF provides a detailed description of the syntax and conventions used by the MIF, including the different definitions used to define your MIF file.
The attribute is the smallest piece of your component that can be defined and identifies a discrete manageable aspect of the component. The way you define an attribute affects not only how the attribute value is stored and accessed but also how the component is managed. To ensure that management of your component is as efficient as possible, consider these points when defining the attributes in your component:
For those attributes that do change, determine whether the value should be available to a management application. If so, read-write access is appropriate for the attribute, but if not, specify write-only access for the attribute.
When defining key attributes to be used as indexes into a table, use as many keys as necessary to provide management applications efficient and flexible access to the data in the table. The attributes you specify as keys should be appropriate to the component and not simply arbitrary indexing values.
If you are developing a component that does not have instrumentation associated with it, ensure that the order of the keys in the MIF file is sequential. This maximizes performance of the service layer, particularly across SNMP.
Note: The values for each row in a table, including the key value, must be supplied either from the MIF database or by component instrumentation. You cannot supply values from the database and from instrumentation for the same row.
For example, ProductID and SerialNumber are likely candidates for database attributes because their values are not likely to change. However, Available_Bytes_on_Hard_Drive is an attribute that is appropriate for component instrumentation.