Researching Integration Requirements

The prerequisite of this stage is a blueprint of the CIs and relationships needed to be discovered by DFM, which should include the attributes that are to be discovered. For details, see Adapter Development and Writing Overview.

Modifying an Existing Adapter

You modify an existing adapter when an out-of-the-box or field adapter exists, but:

• It does not discover specific attributes that are needed

• A specific type of target (OS) is not being discovered or is being incorrectly discovered

• A specific relationship is not being discovered or created

If an existing adapter does some, but not all, of the job, your first approach should be to evaluate the existing adapters and verify if one of them almost does what is needed; if it does, you can modify the existing adapter.

You should also evaluate if an existing field adapter is available. Field adapters are discovery adapters that are available but are not out-of-the-box. Contact Micro Focus Software Support to receive the current list of field adapters.

Writing a New Adapter

A new adapter needs to be developed:

• When it is faster to write an adapter than to insert the information manually into the CMDB (generally, from about 50 to 100 CIs and relationships) or it is not a one-time effort.

• When the need justifies the effort.

• If out-of-the-box or field adapters are not available.

• If the results can be reused.

• When the target environment or its data is available (you cannot discover what you cannot see).

Model Research

• Browse the UCMDB class model (CI Type Manager) and match the entities and relations from your blueprint to existing CITs. It is highly recommended to adhere to the current model to avoid possible complications during version upgrade. If you need to extend the model, you should create new CITs since an upgrade may overwrite out-of-the-box CITs.

• If some entities, relations, or attributes are lacking from the current model, you should create them. It is preferable to create a package with these CITs (which will also later hold all the discovery, views, and other artifacts relating to this package) since you need to be able to deploy these CITs on each installation of Universal CMDB.

Technology Research

Once you have verified that the CMDB holds the relevant CIs, the next stage is to decide how to retrieve this data from the relevant systems.

Retrieving data usually involves using a protocol to access a management part of the application, actual data of the application, or configuration files or databases that are related to the application. Any data source that can provide information on a system is valuable. Technology research requires both extensive knowledge of the system in question and sometimes creativity.

For home-grown applications, it may be helpful to provide a questionnaire form to the application owner. In this form the owner should list all the areas in the application that can provide information needed for the blueprint and business values. This information should include (but does not have to be limited to) management databases, configuration files, log files, management interfaces, administration programs, Web services, messages or events sent, and so on.

For off-the-shelf products, you should focus on documentation, forums, or support of the product. Look for administration guides, plug-ins and integrations guides, management guides, and so on. If data is still missing from the management interfaces, read about the configuration files of the application, registry entries, log files, NT event logs, and any artifacts of the application that control its correct operation.

Guidelines for Choosing Ways to Access Data

Relevance: Select sources or a combination of sources that provide the most data. If a single source supplies most information whereas the rest of the information is scattered or hard to access, try to assess the value of the remaining information by comparison with the effort or risk of getting it. Sometimes you may decide to reduce the blueprint if the value or cost does not warrant the invested effort.

Reuse: If Universal CMDB already includes a specific connection protocol support it is a good reason to use it. It means the DFM Framework is able to supply a ready made client and configuration for the connection. Otherwise, you may need to invest in infrastructure development. You can view the currently supported Universal CMDB connection protocols in the Data Flow Management > Data Flow Probe Setup > Domains and Probes pane. For details about each protocol, see the section describing the supported protocols in the Universal CMDB Discovery and Integrations Content Help.

You can add new protocols by adding new CIs to the model. For details, contact Micro Focus Software Support.

Security: Access to information usually requires credentials (user name, password), which are entered in the CMDB and are kept secure throughout the product. If possible, and if adding security does not conflict with other principles you have set, choose the least sensitive credential or protocol that still answers access needs. For example, if information is available both through JMX (standard administration interface, limited) and Telnet, it is preferable to use JMX since it inherently provides limited access and (usually) no access to the underlying platform.

Comfort: Some management interfaces may include more advanced features. For example, it might be easier to issue queries (SQL, WMI) than to navigate information trees or build regular expressions for parsing.

Developer Audience: The people who will eventually develop adapters may have an inclination towards a certain technology. This can also be considered if two technologies provide almost the same information at an equal cost in other factors.

Summary

The outcome of this stage is a document describing the access methods and the relevant information that can be extracted from each method. The document should also contain a mapping from each source to each relevant blueprint data.

Each access method should be marked according to the above instructions. Finally you should now have a plan of which sources to discover and what information to extract from each source into the blueprint model (which should by now have been mapped to the corresponding UCMDB model).