It's always good to get an idea of the big picture before diving into the details of any technology-centric topic. For this reason, we'll take the time to briefly mention the overarching goals and benefits associated with service-oriented computing as they relate to Web Service contract design.

Because these goals are strategic in nature, they are focused on long-term benefit - a consideration that ties into both the design and governance of services and their contracts. An understanding of these long-term benefits helps provide a strategic context for many of the suggested techniques and practices in this guide.

Here's the basic list of the goals and benefits of service-oriented computing:

• Increased Intrinsic Interoperability
• Increased Federation
• Increased Vendor Diversification Options
• Increased Business and Technology Domain Alignment
• Increased ROI
• Increased Organizational Agility
• Reduced IT Burden

Although it might not be evident, service contract design touches each of these goals to some extent.

Let's explore how.

Increased Intrinsic Interoperability
For services to attain a meaningful level of intrinsic interoperability, their technical contracts must be highly standardized and designed consistently to share common expressions and data models. This fundamental requirement is why project teams often must take control of their Web Service contracts instead of allowing them to be auto-generated and derived from different sources.

Increased Federation
Service-oriented computing aims to achieve a federated service endpoint layer. It is the service contracts that are the endpoints in this layer, and it is only through their consistent and standardized design that federation can be achieved. This, again, is a goal that is supported by the ability of a project team to customize and refine Web Service contracts so that they establish consistent endpoints within a given service inventory boundary.

Increased Vendor Diversification Options
For a service-oriented architecture to allow on-going vendor diversification, individual services must effectively abstract proprietary characteristics of their underlying vendor technology. The contract remains the only part of a service that is published and available to consumers. It must therefore be deliberately designed to express service capabilities without any vendor-specific details. This extent of abstraction allows service owners to extend or replace vendor technology. Vendor diversification is especially attainable through the use of Web Services, due to the fact that they are supported by all primary vendors while providing a non-proprietary communications framework.

Increased Business and Technology Domain Alignment
The service layers that tend to yield the greatest gains for service-oriented environments are those comprised of business-centric services (such as task and entity services). These types of services introduce an opportunity to effectively express various forms of business logic in close alignment with how this logic is modeled and maintained by business analysts.

This expression is accomplished through service contracts and it is considered so important that entire modeling processes and approaches exist to first produce a conceptual version of the service contract prior to its physical design.

Strategic Benefits
The latter three goals listed in the previous bullet list represent strategic benefits that are achieved when attaining the first four goals. We therefore don't need to map the relevance of service contracts to each of them individually.

If we take the time to understand how central service contract design is to the ultimate target state we hope to achieve with service-oriented computing in general, it's clear to see why this book was written.

Service-Orientation and Web Service Contracts
To understand SOA is to understand service-orientation, the design paradigm that establishes what is required to create software programs that are truly service-oriented.

Service-orientation represents a design approach comprised of eight specific design principles. Service contracts tie into most but not all of these principles. Let's first introduce their official definitions:

• Standardized Service Contract: "Services within the same service inventory are in compliance with the same contract design standards."
• Service Loose Coupling: "Service contracts impose low consumer coupling requirements and are themselves decoupled from their surrounding environment."
• Service Abstraction: "Service contracts only contain essential information and information about services is limited to what is published in service contracts."
• Service Reusability: "Services contain and express agnostic logic and can be positioned as reusable enterprise resources."
• Service Autonomy: "Services exercise a high level of control over their underlying runtime execution environment."
• Service Statelessness: "Services minimize resource consumption by deferring the management of state information when necessary."
• Service Discoverability: "Services are supplemented with communicative meta data by which they can be effectively discovered and interpreted."
• Service Composability: "Services are effective composition participants, regardless of the size and complexity of the composition."

Each of these design principles can, to some extent, influence how we decide to build a Web Service contract. With regards to the topics covered in this book, the following principles have a direct impact.

Standardized Service Contract
Given its name, it's quite evident that this design principle is only about service contracts and the requirement for them to be consistently standardized within the boundary of a service inventory. This design principle essentially advocates "contract first" design for services.

Service Loose Coupling
This principle also relates to the service contract. Its design and how it is architecturally positioned within the service architecture are regulated with a strong emphasis on ensuring that only the right type of content makes its way into the contract in order to avoid the negative coupling types.

The following sections briefly describe common types of coupling. All are considered negative coupling types, except for the last.

Contract-to-Functional Coupling
Service contracts can become dependent on outside business processes, especially when they are coupled to logic that was designed directly in support of these processes. This can result in contract-to-functional coupling whereby the contract expresses characteristics that are specifically related to the parent process logic.

Contract-to-Implementation Coupling
When details about a service's underlying implementation are embedded within a service contract, an extent of contract-to-implementation coupling is formed. This negative coupling type commonly results when service contracts are a native part of the service implementation (as with component APIs) or when they are auto-generated and derived from implementation resources, such as legacy APIs, components, and databases.

Contract-to-Logic Coupling
The extent to which a service contract is bound to the underlying service programming logic is referred to as contract-to-logic coupling. This is considered a negative type of service coupling because service consumer programs that bind to the service contract end up also inadvertently forming dependencies on the underlying service logic.

A Web Service contract can be negatively coupled to various parts of the underlying service implementation.

Contract-to-Technology Coupling
When the contract exposed by a service is bound to non-industry-standard communications technology, it forms an extent of contract-to-technology coupling. Although this coupling type could be applied to the dependencies associated with any proprietary technology, it is used exclusively for communications technology because that is what service contracts are generally concerned with.

An example of contract-to-technology coupling is when the service exists as a distributed component that requires the use of a proprietary RPC technology. Because this book is focused solely on Web service contract technology, this coupling type does not pose a design concern.

Logic-to-Contract Coupling
Each of the previously described forms of coupling are considered negative because they can shorten the lifespan of a Web service contract, thereby leading to increased governance burden as a result of having to manage service contract versions.

This book is focused on providing the skills necessary to achieve high levels of logic-to-contract coupling by ensuring that the Web service contract can be designed with complete independence from the underlying Web service implementation.

The most desirable design is for the Web Service contract to remain an independent and fully decoupled part of the service architecture, thereby requiring the underlying logic to be coupled to it.

Service Abstraction
By turning services into black boxes, the contracts are all that is officially made available to consumer designers who want to use the services. While much of this principle is about the controlled hiding of information by service owners, it also advocates the streamlining of contract content to ensure that only essential content is made available. The related use of the validation abstraction pattern further can affect aspects of contract design, especially related to the constraint granularity of service capabilities.

Service Reusability
While this design principle is certainly focused on ensuring that service logic is designed to be robust and generic and much like a commercial product, these qualities also carry over into contract design. When viewing the service as a product and its contract as a generic API to which potentially many consumer programs will need to interface, the requirement emerges to ensure that the service's functional context, the definition of its capabilities, and the level at which each of its design granularities are set are appropriate for it to be positioned as a reusable enterprise resource.

Service Discoverability
Because the service contracts usually represent all that is made available about a service, they are what this principle is primarily focused on when attempting to make each service as discoverable and interpretable as possible by a range of project team members.

Note that although Web Service contracts need to be designed to be discoverable, we do not discuss discovery processes or registry-based architectures.

Service Composability
This regulatory design principle is very concerned with ensuring that service contracts are designed to represent and enable services to be effective composition participants. The contracts must therefore adhere to the requirements of the design principles and also take multiple and complex service composition requirements into account.

•   •   •

About the Authors
Anish Karmarkar, Priscilla Walmsley, Hugo Haas, L. Umit Yalcinalp, Kevin Liu, David Orchard, Andre Tost and James Pasley are co-authors of this article. To learn more about this book and the authors, visit www.soabooks.com/wsc/.