Cisco Network Mgmt Protocol FAQ: Management Information
Q1. What does the acronym MIB stand for?
Answer: Management Information Base.
Figure: MIB and MOs
Q2. Name four categories of management information and tell what distinguishes them.
Answer: State information reflects the current state of physical and logical resources. It is used mainly for monitoring. It tends to be dynamic in nature and subject to constant change, and it cannot be modified by management applications. Physical configuration information is static in nature, changing rarely, if at all; it cannot be modified by management applications.
Logical configuration information concerns parameter settings that are subject to modification by network administrators and management applications. It provides the management “knobs” for the managed device. Historical information contains periodic past snapshots of state information, as well as logs of events that have occurred in the past. It is less common than other categories of management information and is often retrieved in bulk from the device.
Q3. In what ways does a MIB differ from a database management system?
Answer: It is an abstracted view of an “active” real-world system, not a set of information that is actually stored somewhere in a file system. It is optimized for management requirements—for example, omitting general-purpose database capabilities such as joins— and has a smaller footprint. Also, the managed objects contained in a MIB tend to be much more heterogeneous in nature than data maintained by database management systems.
Q4. Name two of the different paradigms that can underlie a MIB definition language.
Answer: Table orientation and object orientation.
Q5. Can you think of a MIB object for which it would make sense to define a maximum access of write only?
Answer: An example is security-sensitive information such as a password.
Q6. What is the name of the language for the definition of management information used with SNMP?
Answer: Structure of Management Information (SMI)—specifically, SMIv2.
Q7. In SMI, what is an important difference between an OID designating an object type and an OID designating an object instance?
Answer: An OID that refers to an object type definition is globally unique. An OID that refers to an object instance is unique only within the MIB that it is contained in.
Q8. Why are SNMP MIB objects not considered objects in an object-oriented sense?
Answer: They lack features that are commonly associated with object orientation. One such feature is inheritance (the capability to derive specializations from existing object class definitions). Other object-oriented features that SNMP MIB objects lack but that were not mentioned in the chapter include polymorphism (the capability for instances of a subclass to be treated as if they are instances of a superclass) and the inclusion of methods as part of the object class definition, commonly associated with the property of encapsulation. SNMP MIB objects are essentially simply MIB variables.
Q9. SNMP MIBs use a hierarchical naming structure very similar to the structure many operating systems use to name files and folders. In which way is the object identifier tree of SNMP MIBs different from a naming tree for a file system?
Answer: The naming tree in a file system represents a containment hierarchy between the objects. If you delete an object that contains other objects—that is, that are in a subtree underneath the object—those other objects will be deleted as well. The object identifier tree of SNMP MIBs, on the other hand, does not reflect a hierarchy between objects. Instead, it reflects the structure of the underlying MIB definition, or the way in which the definitions of the object types are grouped that are instantiated by objects in the MIB. The objects in the MIB themselves are flat; every one of them is a leaf node in the MIB object identifier tree.
Q10. What does the granularity of a model refer to?
Answer: It refers to the degree to which management information is aggregated and lumped together (coarse granularity), or to which every individual real resource is separately accounted for (fine granularity). Generally, a coarse-grained model is more efficient but offers less detailed control capabilities than a fine-grained model.