This article is the sequel to Understanding WBEM, page 14, PCQuest, September. In this part, we move toward understanding WBEM driven CIM Standard and model, and then implement and test a CIM provider; CIM stands for Common Information Model.
As discussed in Part I, WBEM is an initiative to develop a core set of standard technologies for managing enterprise computing environments. CIM is a data model driven by WBEM that attempts to unify and extend the existing management standards (SNMP, DMI and CMIP) using an object oriented format.
After discussing about the WBEM architecture, let us now map the same to CIM. For this, let us first describe the CIM structure.
In order to manage a resource which can be any object, say, a setting in a configuration file, a device and a process running on a system, WBEM client requests the CIMOM for the same. This is shown in the figure that shows the CIM structure. These requests can be to enumerate, add, modify and delete the managed objects.
The CIMOM and the provider reside on the managed system. The CIMOM is responsible for calling the correct provider(s) in response to the client's requests.
The WBEM/CIM provider then processes the incoming request and returns the response to the CIMOM, which then passes on the response to the WBEM client.
Fundamentals of CIM modeling
The goal of CIM is to model various aspects of the enterprise management space. In order to achieve this, a CIM schema framework is provided, which covers the classes (along with attributes, methods and associations) describing the various managed resources. CIM schema is structured into three layers, namely:
Core Model: Describes concepts that are applicable to managing anything. For example, managing a small device to a very big network.
Common Model: Captures notions that are common to particular management field, and is independent of particular technology or its implementation.
Extension Schemas: Represents the technology specific extensions of the common models.
Any CIM model can be expressed visually as well as textually. A visual model is designed using UML (Unified Modeling Language) while MOF (Managed Object Format) is used to write the description of the common elements such as classes, objects, methods, attributes, qualifiers (meta-data) and associations that are used to describe a CIM Model.
Managed object format
The MOF presents concepts such as Inheritance (shown by placing colon and Base class name after the derived class name), Properties, methods and qualifiers. Qualifiers are the meta data, providing information a class, property, method or a reference in CIM Schema.
Visual CIM model using UML
Class: A class is a basic management unit which is described as collection of instances which support a common type, that is, a set of properties and methods. In UML the class can be represented as shown in the figure.
Property: It has a data type and a value used to denote a characteristic of a class. Some properties are designated as Keys, which means that they are required to be unique across all the instances of the class.
Method: It is an operation that denotes the behavior of the class.
Association: An association is a mechanism for providing a relationship between classes. ASSOCIATION qualifier (Meta data) is used in order to differentiate an association class from a normal class. An association is represented by a line between the classes it associates. The star at the ends of the association indicates the cardinality of the class.
In order to instrument a managed resource, a CIM provider has to be: modeled (that is, the extension schema), developed (that is, provider code) and tested.
To get an in-depth understanding of how this modeling is done you can refer to http://forums.pcquest.com and go the technology and tech trends section there.
Ramandeep Kaur, Sr Software Engineer, Novell Software, Bangalore