Thursday, February 12, 2009

SOA CIM: Common Information Model

As I've written about before, there is a lot of confusion and different interpretations of what a "Common Information Model" (CIM) is. The most common reification of CIM is to imply that it is the same concept as in chapter 4 "XML: The Foundation for Business Data Integration" in David Chappell's seminal book Enterprise Service Bus from 2004. Dave describes the need for having a common XML data format for "expressing data in messages as it flows through an enterprise across the ESB", but never uses the term CIM or describe any modeling approach. This common format will naturally comprise any kind of XML data used in the service bus, including messages, business entities and event data.

The CIM concept in relation to SOA was described by Mike Rosen and
Eric Roch in 2006: it defines a common representation of business entity objects that provides a canonical format and unified semantics for a business domain.
The above figure is from Mike Rosen's article Business Architecture and SOA. As can be seen in the figure, CIM comprises a shared information model that is the basis for the documents that flows through the business processes built using the capabilities provided by the services.

Mike explains the role of CIM like this:

[The SOA Business Model] has to manage the sharing of services and information across processes. In other words, it needs to eliminate redundancy, overlap, and gaps between services so that each business capability is implemented once, by the organizational unit that is responsible for that capability. And that those services are used by all the different processes needing those capabilities. In addition, all of the information shared between services must be identified in the common information model. In other words, all services that are related to the same business concepts must use the same information to describe those concepts. Finally, the SOA Business Model must ensure that all of the information passed into and out of the business services (mostly in the form of documents) is defined in the [shared semantic] information model.

I hope this clarifies the definition of what a CIM is:
it is a model that comprises shared information entities such as "Customer", but not actual messages like "CustomerHasMoved" defined in the business process information model (BPIM). The message types such as "AddressChange" (document) are part of this logical model as projections of the shared entities.

The "Canonical Schemas" used in messages are part of the service model's standardized service contracts. Note that these schemas are generated from, but are not part of CIM. Generating the canonical schemas from CIM makes enforcing "Schema Centralization" unnecessary, as they already share a common basis. This helps alleviate the negative effects of a one true schema approach; e.g. the versioning ripple effect through the canonical data model that affects all related service contracts - even if not intended.

Do not confuse CIM with the centralized SOA logical data model (LDM) approach. The services speak "CIM" natively. Service messages are not transformed to/from CIM by a message broker as in the LDM schema bus approach. Still, using the message broker approach is handy to include non-conformant services in your SOA solutions. Enforcing a LDM for your information will cause the same problems as the traditional EAI hub-and-spoke canonical data model (CDM).

I recommend reading chapter 5 "Service Context and Common Semantics" in the "Applied SOA: Service-Oriented Architecture and Design Strategies" book by Michael Rosen et al. Figure 5-12, 5-14 and 5-15 in the book shows how a grapical projection DSL for documents (message types) might look.


Unknown said...

I am a soa consultant, and i am agree with you in this article, but i have a question. Technically, how you implement a approach where you maintain an MDM as a common logical data dictionary and map every physical object model to it"? If for example I have defined the common logical data dictionary in a generic xml schema, how I can to create a different customization of it, if I need only a few information of this big generic xsd? Is it possible to create a customized instance of a generic xsd in other schema for use it in my service interfaces? Thanks in advance

Kjell-Sverre Jerijærvi said...

It's been a long time since I've done anything in this area, but I belive the referenced book "Applied SOA: Service-Oriented Architecture and Design Strategies" book by Michael Rosen et al show how to do projections of the model.

Also, even if the CIM concept is similar to MDM, they are not the same. CIM is for the message contents that drive your business processes ("customer has moved"), while MDM is for the data ("customer") in the systems-of-record.