Executives Guide to Web Services (SOA, Service-Oriented Architecture)

The emergence of a new technology in the software industry has a tendency to generate media attention and hype. Web services has perhaps been the most heralded and hyped new technology to date. The Gartner Group refers to this phenomenon as the “Hype Cycle.” ^[1] Figure 6.1 illustrates Gartner’s Web services hype cycle from late 1999 through 2005. Gartner believes that Web services will reach the peak of its hype cycle by early 2003 and then descend into the “Trough of Disillusionment,” where Web services are unable to live up to the mystical image created by the software industry.

Figure 6.1: The Web services hype cycle. Source: Gartner, 2002

Considering how over-hyped Web services are today, one could envision reality versus hype as two ends of a spectrum. On one end of the spectrum, there are the very real and tangible business benefits that can be attained today. On the other end, there is the speculation of what might be attained in the future. This concept is illustrated in Figure 6.2, the Web services reality versus hype spectrum. Although the hype of today is not what we will see and implement tomorrow, it is important to remember that over the coming years it is possible—even likely—that Web services will bridge the gap, making today’s hype tomorrow’s reality.

Figure 6.2: The spectrum of Web services reality versus hype.

The Reality

At the reality end of the spectrum, early adopters will gain significant benefits from Web services through internal integration and closer collaboration with partners, suppliers, and customers. The capabilities and standards required for the first steps of integration and collaboration pilots are already in place. Earlier pilots will primarily leverage XML messaging from the enabling standards tier and SOAP binding from the evolving standards tier. Early pilot projects will focus on cost reductions and efficiency gains through the following initiatives:

• Reduced System Integration Costs —Web services, using XML and SOAP, will become the de facto mechanism for systems integration. As an organization’s application portfolio migrates to a service-oriented paradigm, all systems will use Web services as a primary interface mechanism, removing the need for custom integration projects. This standards-based approach to application integration will significantly reduce the time, effort, and cost of future integration projects.

• Faster Time to Market —By exposing current application functionality as services, the promise of software reuse can at last become a reality. Reuse of services, rather than the more traditional “let’s reinvent the wheel” approach, will enable greater flexibility and faster time-to-market for implementation of capabilities or services.

• Real-time Operational Visibility —Without exception, every organization seeks to improve the availability, accuracy, and visibility of key operational metrics such as sales forecasts, operating expenses, cash flow, inventory levels, and so on. The implementation of real-time executive dashboards tailored to the specific needs of each executive will provide real-time mission-critical information, enabling better management of internal business execution.

• Closer Collaboration —Sharing much more information with partners, suppliers, and customers on a system-to-system basis, making it possible to take slack out of the value chain for a firm and its partners. For example, much like Dell, a firm might be able to improve order visibility, enabling reduced levels of on-hand stock—thus creating additional space to add new production lines. This situation in turn could increase production capabilities and improve purchasing economies of scale.

• Redeployment of Valuable Resources —By enabling both greater reuse and improving collaboration, it will be possible to greatly improve operational efficiencies.

Dell Computer is a great model of how Web services capabilities can be applied to achieve the advantages above. In their article “Break on through to the Other Side: A Missing Link in Redefining the Enterprise,” ^[2] John Hagel III and John Seely Brown discuss how Dell has continued to change the rules in the PC market:

Direct material purchases represent as much as 70% of Dell’s revenue, so even modest savings in the area can deliver significant bottom line impact. Equally important to Dell is the inventory asset exposure—in an environment where product prices recently have been declining at 0.6% per week, any excess inventory can become very costly. It is not surprising that Dell sees significant benefits to more effective supply chain management.

Dell began by focusing on its network of third-party logistics providers who operate distribution centers (“vendor managed hubs”) where raw material inventories are maintained for Dell’s assembly operations. Dell’s “ship to” target from the time of receipt of a customer’s order is 5 days, yet the average fulfillment time of their suppliers is 45 days. To meet their “ship to” targets despite the 45-day delivery time from suppliers, Dell used three business approaches at the outset: maintain a 26–30 hour inventory buffer at the Dell assembly plant, ensure vendors maintain a 10-day inventory buffer at the vendor managed hubs and distribute on a weekly basis a 52-week demand forecast to suppliers. The vendor-managed inventory at the hub helps to deal with unanticipated supply disruptions and inaccuracies in the near-term forecast.

In the first stage of its initiative, Dell focused on reducing the 26–30 hours of inventory held at their assembly plants. Using i2 supply chain software, Dell now generates a new manufacturing schedule for each of its plants every two hours to reflect customer orders received over the previous two hours. Publishing those manufacturing schedules as a web service via Dell’s extranet, Dell alerts the vendor-managed hubs regarding what specific materials are needed and directs them to deliver those materials to a specific building and dock door so that the materials can be fed into a specific assembly line. These alerts are sent in an XML format so that they can be integrated directly into the disparate inventory management systems maintained by its vendor-managed hubs. The distribution centers are implementing the capability to respond to the material requests automatically through the extranet and then have 90 minutes to pick, pack, and ship the required parts to the factory. With this automated, web services-based approach, Dell has been able to reduce its own holding of raw material inventories from 26–30 hours of production to 3–5 hours—a reduction of more than 80%. Eric Michlowitz, Director of Supply Chain E-Business Solutions for Dell explains, “what we’ve been able to do is remove the stock rooms from the assembly plant, because we are only pulling in materials that we need specifically tied to customer orders. This enabled us to add in additional production lines, increasing our factory utilization by one-third.”

Of course, lean manufacturing approaches often push back inventory exposure from the manufacturer to the supplier. Dell’s goal is to eliminate excess inventory, not simply push it back up the supply chain. This drove its next wave of initiatives designed to eliminate the need to hold buffer inventories in the vendor-managed hubs because of errors in the supply chain. To do this, they focused on checking the reliability of supplier delivery schedules early enough in the process to allow effective contingency planning to mitigate against unanticipated disruptions in supply (e.g., by temporarily withdrawing certain models on Dell’s web site). Michlowitz’s team is deploying a web services-based event management system to do this through its extranet. This approach avoids the need to directly access the application systems of its suppliers and instead relies on an approach consisting of automatic exchanges of inquiries and confirmations with its suppliers (e.g., inquiring whether the supplier shipped as promised). Because the system is automated, the supply chain team is able to focus its energy on handling the exceptions. Dell expects to be able to reduce inventory exposure in the vendor-managed hubs by 10–40%, while at the same time significantly improving gross margin performance through more effective matching of demand and supply.

In this example, Dell demonstrates superior performance using only basic Web services capabilities. As Web services standards evolve, Dell may be able to identify further opportunities to improve efficiencies, reduce costs, and increase profitability.

The Hype

At the hype end of the spectrum, a disproportionate amount of early Web services discussion has focused on the conceptual ability to automate application assembly on a Just-In-Time (JIT) basis. In this scenario, it is proposed that applications will be automatically and dynamically assembled as needed by using services published in public registries on the Internet, typically using the UDDI standard. This capability, although technically feasible, faces significant obstacles including the following:

• Trust —The fundamental ability to put trust in the source of a public service is an obstacle. Unless one knows that a public service is trustworthy, as with the Dunn and Bradstreet example in Chapter 2, “Standards, Concepts, and Terminology,” would a firm really be willing to use it? How does one know whether the service does what it is supposed to do? What if the service is malicious? The ability to trust the provider of a service will initially limit the use of Web services to collaboration with partners, suppliers, and customers where a pre-existing trusted relationship exists. It is unlikely that significant markets will emerge for publicly available services unless those services are provided by organizations that are deemed trustworthy by a large target audience.

• Security —It is possible to secure a private network with trusted partners, suppliers, and customers by using Virtual Private Network (VPN) capabilities today. But as the desire to use services that exist outside of a private network increases, the emerging Web services security standards will need to mature significantly. In the medium-to-long term we do not anticipate that security will remain an inhibitor to Web services adoption. A number of new security standards have recently been proposed by the WS-I organization to close the security gap (for example, WS-Security and WS-Policy).

• Management —A range of support tools are emerging that will be applied to the broad category of Web services management. This category includes tools for performance management, quality of service, logging and auditing, and so on. All of these capabilities must be evaluated as the number of Web services being used by an organization grows. To date, management requirements have received less visibility than Web services security, and there has been less attention focused on creating management standards. As a result, proprietary solutions have already begun to appear, and these may be followed by the eventual emergence and adoption of Web services management standards.

^[1]Gartner Group, September 2002, “Navigate the Web Services Hype Cycle” by Mark Driver.

^[2]“Break on Through to the Other Side: A Missing Link in Redefining the Enterprise” by John Hagel III and John Seely Brown, pp. 12–13.