A Service-Oriented Architecture Approach to Enterprise Capture

By Clay Mayers

What is a Service-Oriented Architecture?

Historically, integrating applications from multiple vendors into a cohesive solution has been a major challenge for all types of organizations. Integration efforts often require both in-depth application programming interface (API) knowledge of each system and numerous one-to-one integrations between each system. This creates far too many integration points for organizations to effectively develop and maintain.

In response, enterprise application integration (EAI) efforts were developed in an attempt to simplify the integration of different systems by creating a common layer to manage these integrations. While this simplified the integration of each component system, it still required complex integration code to combine multiple systems into an effective enterprise solution. Though EAI efforts simplified the architecture, they remained complex to create, difficult to both maintain and upgrade, and in effect, just concentrated “the mess in a closet.”

Today, the technology industry is buzzing about a new approach to integrating business systems: Service-oriented architectures (SOA). SOA is not a product or technology; it’s a technique for designing large, enterprise applications. Instead of requiring an in-depth knowledge of each underlying system’s API, SOA focuses on exposing systems as services. Developers responsible for a new business system can focus on the task they need the system to perform and the data required to perform the task. There is no need to be concerned with the programming implications of this task. By using standard data formats and communication protocols, many of the obstacles to integrating disparate systems are greatly simplified.

An increasing number of organizations are either planning or actively developing a SOA strategy. According to a 2006 Forrester report[1]:

• Of the respondents, 38 percent of firms are already using SOA.
• The use of SOA will grow in 2006. Seventeen percent of respondents have an enterprise-level strategy for SOA and an additional 21 percent use the technology selectively. More than half (69 percent) of these firms say they will increase their SOA usage in the next two years. None of the respondents said they would decrease usage.
• SOA adoption is greatest in the Global 2000, where firms are most likely to have an enterprise-level strategy. Of the respondents with more than 20,000 employees in their firms, 26 percent report having enterprise-wide initiatives around SOA.
• The drivers that are causing firms to implement or plan SOA implementations include using more standards-based technology, reducing the amount of code that must be developed and maintained, reducing ongoing maintenance of their enterprise solution – especially the integration layer(s), and enabling more independence with respect to vendor selection.

What’s the Difference Between SOA and Web Services?

Service-oriented architectures are being enabled by a number of open standards, including Web Services, SOAP, XML and HTTP. These standards provide a common method for software developers to expose their applications as services and to call other systems.

• Web Services, by themselves, are not SOAs; however, Web services provide the standards and tools to build an SOA. A Web service is a self-describing, self-contained, modular software application that provides one or more business functions to other systems through a standard Internet connection.
• Simple Object Access Protocol (SOAP) is the protocol that allows Web services to communicate using XML messages. SOAP messages are independent of operating systems and platforms, and they are transported using standard Internet protocols, most commonly HTTP. The use of SOAP and Internet protocols provides access to systems without regard to their location, operating system, or platform.
• Extensible Markup Language (XML) is a cross-platform, extensible, text-based standard used for communications and data exchange across the Internet. XML is a formal specification of the World Wide Web consortium (W3C) and has become the universal format for structured documents and data on the Web.
• Web Services Description Language (WSDL) allows Web services to be self-describing to potential users of the service, defining what operations the service will provide and how to access it. A WSDL is an XML file that describes the public interface to the Web service and allows the application to communicate with the service.

Leveraging the Benefits of SOA

Rather than depending upon detailed understanding of each system’s APIs, as well as those in an EAI layer, SOAs rely on lightweight interoperability, where the focus is on the result of the service rather than on the implementation. A simple example of a service is a centralized payroll system. Rather than requiring each business unit to understand complex rules for multi-jurisdictional payrolls, the service simply requires basic employee information and hours worked, and it generates payments and tax documents. The business requires no additional knowledge of the platform or programming implementations.

Another key benefit to service-oriented architectures is that they can be fully distributed by leveraging common Internet standards. By utilizing Internet protocols such as HTTP, Web services enables integration outside of a single enterprise. Integrations may be created between distant organizations or between corporations and business process outsourcing (BPO) firms. Using the previous payroll example, payroll processing may be easily outsourced and only requires the company to supply basic information through the Internet using Web services.

SOAs also help to reduce the implementation and, more significantly, maintenance costs associated with integrating multiple systems. By focusing the integration on the system’s runtime results rather than its APIs, the integration is not constrained by system implementation details. This lightweight integration enables each organization to upgrade its systems independently, as long as they continue to agree on how they will exchange information. Services also allow for code reuse; multiple applications can be built using many of the same services.

Applying an SOA Approach to Enterprise Capture

As document capture solutions are increasingly deployed across the enterprise, businesses are demanding the ability to leverage investments across multiple departments and multiple business lines. SOAs enable businesses to leverage common functions or abstract details, such as the document type or index fields, so that common processes may be used in different applications.

Some companies are already employing methods for building enterprise architectures that leverage SOA to include capture functionality. For example:

A capture process flow can be defined as a service, enabling enterprise architects to mix and match various services together to solve unique business problems. The process flow may include multiple capture tasks, such as image classification, document classification and manual classification validation. When called, the service accepts a batch of images, runs image enhancement filters to improve the quality of the images and classifies the document, including both automatic classification and manual classification validation before returning the batch for routing and additional processing.

A common service can also be used in different ways by each application in the enterprise. One example that allows service reuse involves an indexing service where document indexing is performed. In this case, the document types and index fields are unique to each application that calls the service. By designing the service to accept parameters that further define the process, a generalized service may be developed and reused across several different capture applications.

By leveraging Web services capability to provide remote integration between enterprise systems, input management services may be exposed to systems in other locations – or even other companies. A corporation may deliver scanned images to a service provider that performs document identification, indexing and validation before returning the required data and images to the corporation’s content management or other business system.

Selecting an Enterprise Capture Solution for SOA

Businesses should look for document capture solutions that provide flexible web services capabilities, which allow customers to create rich enterprise business systems and leverage their investment in document capture throughout the enterprise. This approach enables enterprise architectures to define services and integrations to match their unique architectures.

The goal is to select Web services that:

<h3 style="MARGIN: 6pt 0in">

• Enable capture solutions to be more tightly integrated into enterprise business systems including ECM, ERP, etc., rather than simply exporting data to the next system in the business process. As content management systems are more tightly integrated with other business systems, input management solutions must support this level of tight integration as well.
• Leverage capture services throughout the information lifecycle, not just as an input function. This allows capture services, such as image enhancement, rescanning, classification, or indexing steps to be called at any point during a business process, not only when specified by the image capture process.
• Use an open, flexible approach allowing businesses to focus on their enterprise architecture, rather than requiring that their architecture change to match proprietary interfaces.
• Provide access to the system at the level that the architecture requires, integrating both complex capture logic and discreet capture functions that serve a larger business process, as required.

Summary

SOA is an approach to integrating business systems that allows organizations to reuse code so that multiple enterprise applications can be built by using the same services. A growing number of companies are implementing SOA architectures for their enterprise content management and data processing systems in order to reduce operational costs and ease the maintenance of their enterprise solutions. To best leverage their ECM investments, these companies should implement an enterprise-level capture platform that provides a flexible architecture and Web services that are not fixed (defined by the vendor) so that the company itself can define the services to fit its unique architecture.

Clay Mayers is Chief Technologist of Captiva software for EMC Corporation. For more information, visit www.emc.com/captiva or email info@captivasoftware.com

[1] Herbert, Liz. Software and Services in Large Enterprises. Forrester, March 8, 2006