Common Information Model
An organization is looking to improve the methods it uses to develop and integrate IT systems into its operations.
An organization is struggling to integrate its IT systems and business operations. This integration may be required to increase its efficiency, embrace new technology, expand its business, and/or improve its customer service.
Organizations today have many IT systems that have been developed over time. They are struggling to innovate while trying to maintain and manage existing IT systems. Market forces are driving organizations to become more integrated, to support new channels (such as mobile), exploit new platforms (such as cloud and big data), and support a more social way of working both inside and with external parties such as customers and business partners. What can they do to provide a stable foundation for maintaining a state-of-the-art IT capability?
A fictitious travel agent, GKDMR Travel, has many systems that have been acquired over the years as the business has grown. Its systems are getting old and it wants to make better use of modern approaches such as mobile, social media, cloud-based deployments, open source software, and analytics. The company’s budget is not large, but it anticipates having to change some of its systems to accommodate its vision. However, many of the existing systems will remain and there is a need to interface them to the new systems.
- Existing systems provide key capability to the organization. They are expensive to replace and must continue to operate while any changes are made to them.
- An organization’s information is distributed among its existing IT systems, in people’s heads, in unmanaged files on employee laptops, in paper documents, and externally on the Internet, or in business partner systems.
- The information within an existing system is rarely as good a quality as the organization that owns it believes.
- An organization will use inconsistent terminology across its many departments, professional disciplines, and internal fiefdoms. Sometimes the same term is reused for different purposes, or even when the meaning is consistent, assumptions about its timeliness, precision, and accuracy will differ among the different groups inside the organization.
- The database schema of an application does not document all the information available through the application interface. Some key values are derived. For example, GKDMR Travel’s booking system records all the trips a customer has booked with the company to the database. This information is visible in the database schema. However, GKDMR Travel operates a loyalty scheme, which provides discounts to customers who book trips regularly with it. The loyalty scheme was added to the booking system after it went into production. As such, to save time, the loyalty status was not added to the database schema. Instead it is calculated in the application logic whenever a new holiday is booked. As a result, an important piece of information about a customer that could be useful for analytics or another customer service application is locked in a single application’s logic.
The organization should develop a shared understanding of its information’s terminology, meaning, and structure in order to facilitate agile and effective integrated operations.
Information is at the heart of an organization’s ability to service its customers, deliver on its promises, and collect the expected rewards for its services. The types of information that an organization holds are stable, although the scope of information available has been growing recently with the digitization of many aspects of our lives.
The organization can develop a set of common information definitions that captures:
- The meanings of key concepts, facts, events, and activities used by the organization
- The preferred structures that should be used to store and exchange this information
- The types of values that must be captured to describe them
These common information definitions are implemented as a collection of models and definitions. This collection covers the portion of the organization’s information that needs to be shared and synchronized. It must represent many perspectives on this information and be consumable in different programming environments. Collectively, these models and related artifacts are called the common information model.
An organization can choose which types of models to create, and to which level of detail. The focus will depend on where change happens most often and where the cost of change is high or notoriously error prone.
The different types of models may not be entirely consistent. However, the closer they are, the less transformation is required as information flows around the enterprise.
- If an organization can understand the information it uses, then it is better able to assess which systems are important, how best to manage the information it has, what types of information need to be made available to new applications, and where special care must be taken to protect valuable or sensitive information.
- A well-formed common information model creates an adaptable definition of how information should be represented and shared at key places where the organization needs to synchronize or control its operations. Without it, information sharing can be ad hoc and developed as a number of inflexible point-to-point solutions.
- When a new project is started, the common information model is an invaluable planning tool for identifying what type of information is needed, where it is located, and how it should be structured in the new capability.
- When an existing application built to the guidelines of the common information model must be maintained, the development team has the common information to guide its understanding of the existing code.
- For a common information model to deliver value it must be treated as an asset, with an owner who is responsible for its ongoing maintenance and executive support to ensure it is properly governed and to encourage use of the common information model content.
- The contents of a common information model must be easily consumable by the teams that are building and maintaining the IT systems. Ideally, physical artifacts such as interfaces and schema would be generated from the common information model and included in the developers’ working toolset. Thus it becomes easier for the development team to use the common model rather than create its own. The common information model will be welcomed and adopted when it saves the development team time and effort.
- In some organizations, where data modeling is not widely understood, the generated physical artifacts are the only part of the common information model that many developers see. In this case, these physical artifacts should include comments and annotations covering the semantic definitions of the model.
GKDMR Travel has three major projects that will be assisted through the creation of a common information model:
- It is building smartphone and tablet applications for its customers and customer-facing staff. These applications will be specialized to the needs of the users being served and will integrate with the existing systems that GKDMR Travel has running. The people using these applications will also receive printed documentation and may use the company website, too. GKDMR Travel needs to develop these applications quickly, but at the same time believes the information these applications display should be consistent with the other information these people see.
- It wants to offer real-time alerts and actions to support its traveling customers with relevant information, rebooking when weather and other factors disrupt travel and offering other relevant offers. This capability is going to need information from existing systems that is integrated with external information.
- It wants to improve the management reporting on the state of the business and the trends they see. The aim is to optimize the operations of the business.
The plan is to create a common information model that includes the following:
- A glossary of terms2 that describes the meaning of the different types of information that will be used by customers and staff who are using the new capability. This includes the information displayed by the new applications along with the existing systems.
- A model of the objects that will be used in the mobile applications, ensuring they display consistent structures and each field in the structure has a consistent name so it can be matched with printed information and the information on the company website.
- A model of the service interfaces that define how information will be exchanged between the existing systems and the new ones.
- A persistence model that describes how data should be consolidated and linked together.
The consistency that occurs between these models will speed up the collaboration between the different teams by implementing new capability and extending and enhancing the existing systems, which will then enhance the quality of the resulting applications.
Some of the key uses of a common information model within an organization are
- Creating a consolidated repository of information from different sources, such as a data warehouse, operational data store, and master data management hub.
- Defining canonical information services for exchanging information in a service-oriented architecture, or for consumption by a business process management (workflow) engine.
- Identifying the information that needs to be supplied to the corporate reporting platform.
The following patterns are specializations of the common information model pattern. Each supports a different purpose and style of usage for the common information model:
- Concept Beads pattern: Supports simple, fine-grained models of core concepts. Suitable for an organization that needs to quickly assemble simple applications that would benefit from consistency in implementation since they use many common concepts.
- Continuous Fabric pattern: Supports an organization where there is high value in understanding the relationships between people, assets, events, and activities; this provides a blueprint for how the landscape of information links together.
- Encapsulated Views pattern: Supports an organization that wants to develop common interfaces to information irrespective of how or where the information is stored. This style is called information virtualization and is a common approach in service-oriented architectures.
- Unifying Context pattern: For an organization that has historically operated as multiple independent units and wants to become more consistent and integrated in its use of information.
The sections that follow describe these patterns in more detail.