How Web Services Can Be Used to Help Create New Organizational Efficiencies

Storebrand (pronounced "Story-brand") is Norway's largest financial and insurance company a diversified financial services provider that provides health and life insurance as well as banking and asset management services to more than 280,000 clients throughout Norway.

In this highly competitive market Storebrand is well positioned for continued growth. For instance, by using innovative techniques and packaging, Storebrand Life (the company's life insurance unit) allows customers to "link" products together (such as defined contribution occupational pension products with individual investment choices) to determine how aggressively their pension funds are invested. Linking other products and services together and finding new distributors for these products is expected to be key to the company's continued growth.

Also important to continued growth will be the company's ability to control costs by introducing internal efficiencies as well as efficiencies in dealing with its business partners. The company knows that if it can automate certain business processes, it can lower operating costs, and those cost savings pass directly to the company's bottom line.

Sidebar 6-4
Storebrand: Using Business Process Management and Web Services to Reduce Operating Costs

Imagine serving 6,500 companies with approximately 390,000 employees and having to synchronize the records of those employees such that individual coverage could be accurately calculated for each employee under a variety of insurance schemes. For Storebrand, this monumental synchronization task has largely been a manual process that involves the verification, updating, and rechecking of data before it is entered into Storebrand's centralized database. And this process has been begging for an automated solution a solution that heretofore has not been forthcoming using existing Web standards.

Storebrand has attacked this problem in two ways: (1) by modifying its existing manual practices (in other words, changing its internal business processes such that people do not have to get as heavily involved in handling pension calculation and input into Storebrand's mainframe database); and (2) by enabling employee data to be captured from client payroll systems using electronic Web services, automatically performing calculations on that data, and then directly transferring the resulting information directly to its mainframe database.

As a result of automating this manual process, Storebrand has been able to improve operational efficiency by reassigning a number of employees who are currently involved in the manual pension calculation/input task to other tasks (thereby making better use of people and potentially saving money by not having to hire new employees to do other manual tasks). Storebrand's automation of this pension calculation/input task is expected to save the company thousands of hours in manual labor costs annually.

To assist in this effort Storebrand employed the assistance of IBM's jStart team (www.ibm.com/software/jstart) to help with the implementation of Web services. The project team identified the following objectives:

• Create Web services for Storebrand (as Servlets);

• Adapt vendor payroll systems to do Web services;

• Implement necessary infrastructure;

• Connect existing Storebrand business partners; and

• Provide a proof-of-concept to Storebrand that Web services can reduce operating costs.

Storebrand, with IBM's assistance, accomplished these goals. The product design involved the following:

To get the divergent environments [Windows-based and mainframe databases, as well as customer/business partner environments] to share data, jStart suggested using a SOAP (Simple Object Access Protocol) connection to bridge the gap by connecting the Windows environments of most payroll applications with the WebSphere Java environment at Storebrand Life.

The team determined that three options were available:

SOAP alone SOAP would provide the basic program-to-program "glue" that enables applications to bind together and enter into peer-to-peer communications.

SOAP + WSDL definitions WSDL would provide a standardized description of the service, but would not offer full repository lookup. Once the application service has been described, it would be published with a service broker that helps make the application/service known to Web "requestors" (in this case, new Storebrand distributors).

SOAP + WSDL + UDDI (including service description in registry) UDDI is also a framework for Web services integration. It contains standards-based specifications to describe the Storebrand products and services and allow discovery in a global registry architecture.

Storebrand initially opted to utilize a combination of all three technologies. To do this, the team created a COM object that takes the XML document as input and sends it to the Storebrand Web service, by transmitting it over the Internet via the HTTP protocol. The customer sites, in this case, primarily used Microsoft Windows NT or 2000 systems to perform their data entry for the benefits processing. Thus, the team decided on building COM interfaces in VB for the client-side and used Microsoft's SOAP toolkit for the interface. On Storebrand's server-side, the team converted an existing Java application to a service by simply using the Wizard provided in the IBM XML & Web Services Development Environment. This Web services enabled application receives the XML document and transforms the data for the verification and update process required to feed the DB2 central database of Storebrand. Data transport and manipulation at the backend is managed with MQSeries Integrator.

Source: http://www-106.ibm.com/developerworks/library/ws-asa/. Used by Permission.

Notice that Storebrand was able to use all of the most important Web services protocols in unison to create a sweeping, cross-platform program-to-program solution that enabled the company to reduce human overhead costs. Also notice how IBM's WebSphere platform made it possible to augment Web services protocols with other protocols and messaging services (such as its MQSeries Integrator) in order to enable cross-platform interoperability. Of all of the examples in this book, Storebrand made the most comprehensive use of Web services in creating its solution.

By making use of XML to capture and present data, by modifying business processes between Storebrand and its customers, and by using the SOAP Web standard for program-to-program communications, the company will now be able to save thousands of hours in information capture and input time thus reducing costs related to manual information processing.