A new computing paradigm known as web services is quickly gathering momentum in enterprise IT. Web services represent an emerging model for developing and deploying enterprise software applications that promises to fundamentally change not only the way companies build and deploy software, but also how they communicate with their partners and customers.
Imagine taking the functionality of an existing mainframe-based application, such as inventory management, and incorporating it into a new application to provide real-time inventory information to trading partners. Or perhaps your business would benefit from an employee portal that seamlessly connects third-party services such as travel agents, payroll outsourcers and health insurance providers with data such as current sales forecasts, marketing activity and customer account information.
This increased flexibility emerges because web services will likely change the fundamental cost structure of enterprise application integration (EAI). Observers have likened the effect web services will have on enterprise computing to the dramatic increases in flexibility and productivity that client-server architectures introduced in the late 1990s. Market research firm IDC forecasts that spending on enterprise integration will reach US$50 billion in 2003. More explicitly, Gartner suggests that the web services software market will reach US$1.7 billion in 2003. Still, notwithstanding the enthusiasm and the major investments in product development made by software vendors, the deployment of web services is very much in its infancy, and many questions remain: What are the practical implications for early adopters? What problems can web services help solve today? What are the first steps?
What is clear, however, is the tremendous potential for value creation. Software developers and forward-looking IT organizations have already begun to incorporate web services into their overall technology strategies. Ultimately, the bottom line is that web services offer the promise of doing more, doing it better, and doing it for less cost than previously possible. Most exciting, the point of entry for web services is substantially lower than previous software environments.
The experiences of A.T. Kearney and The Stencil Group in advising innovative software companies and leading-edge corporate users of IT suggest that web services offer a compelling vision of business process transformation that is grounded in very real, pragmatic considerations of technology implementation.
A Web Services Primer
Web services are an emerging model for developing and deploying software applications. This approach can improve the flexibility and the reach of existing IT infrastructures. It also creates myriad opportunities for implementing new business models based on the offering and consumption of software. The question remains, however: How do we define the approach that promises these bottom-line results?
The web services label is generic, but software vendors and enterprise customers are developing a relatively consistent understanding of the concept. Web services have two levels of meaning—one conceptual (and business-oriented) and one (technically) specific. Conceptually, web services represent a model in which small pieces of application functionality are available as services to be consumed and combined with other applications over a network. Specifically, web services are a stack of emerging standards that define protocols and create a loosely coupled framework for programmatic communication among disparate systems.
A Common Sense Analogy: The RCA Jack
A common-sense analogy can be drawn to a major milestone in the development of the home electronics industry. Before RCA’s introduction of its device for connecting radio receivers with other devices, stereo systems were closed, monolithic cabinet systems. As more manufacturers began to adopt the RCA jack, stereo components could interoperate without regard to vendor. Consumers benefited from a wealth of choices. No longer locked into one company’s solution, customers could construct a home stereo system on a “best of breed,” component-by-component basis. Enthusiasts could “scale” a home entertainment system with just the right mix of equipment, and replace individual elements as budget and needs dictated.
Clearly, questions of scale, flexibility, and vendor lock-in for enterprise software architectures are of a different magnitude than for home stereo systems, but the benefits of interchangeable components are the key to success for both models.
The Web Services Technology Stack
By intent, web services are not implemented in a monolithic way, but rather represent a collection of several related technologies (see sidebar: Core Web Services Standards). At a bare minimum, any web service entails a connection between two applications—in programmers’ parlance, a remote procedure call (RPC)—in which queries and responses are exchanged in XML over HTTP. This high-level description often is referred to as a services-oriented architecture, or an SOA.
The more generally accepted definition of a web service, however, implies implementation of a stack of specific, complementary standards. Today, three core standards make up the basis of any web services implementation: SOAP, XML and UDDI. These standards, as well as other emerging layers, work together to enable this new architecture.
Figure 1: The Web Services Technology Stack
Today, the core layers that define basic web services communication have been widely accepted by software vendors and customers alike, and these basic web services technologies will be implemented in a very large range of enterprise software products (see figure 1). Higher-level layers that define strategic aspects of business processes remain an open question, and it is possible that divergent approaches will emerge.
A Promise of Business Benefits
In today’s uncertain economic environment, operational efficiency is more important than ever. Management teams are faced with the daunting task of delivering increasing business value with reduced resources. Strategic investments in information technology are a key element of that mandate, but too often, costly technology deployments offer limited returns at substantial risk.
Web services promise to change that dynamic. By allowing businesses to focus on their strategic goals, web services leverage an organization’s existing IT investments to be more responsive to changing business requirements, to reduce risks and deliver predictable results from IT projects, to improve visibility into data-centric operations, and to foster business model innovation.
Increased flexibility. Web services enable companies to make strategic business decisions on reasons of business merit, rather than on technological limitations from existing infrastructures. The vendor-neutral approach embodied by the open web services standards means that companies can select best-of-breed technology architectures and increase supplier choice. Additionally, web services-based transactions significantly reduce costs and enable companies to establish or terminate business partnerships much more nimbly and efficiently (see figure 2).
Figure 2: Web Services Balance Control and Flexibility
Reduced risk. Traditionally, IT has been a significant cost center for most businesses, and only a handful of companies have leveraged technology as a genuine competitive advantage. Web services shift the cost model of strategic IT from front-loaded expenditures and “cliff”-based deployments to one of manageable, incremental investment. This shift reduces risk and frees up working capital for other strategic investments.
Increased visibility. By enabling fluid, realtime connections among operational systems, web services significantly increase managers’ visibility into data-centric business processes. Goldman Sachs estimates that one large manufacturer’s supply-chain initiatives, which leverage web services, could halve the amount of inventory and working capital that currently must be carried on the company’s books. Also, web services could dramatically reduce the company’s operating cost per unit. Not only will internal data become more accessible, but trading partners that share web services-based business processes will also realize a much higher degree of coordination and information sharing. Finally, the process of software development itself will become less opaque, enabling enterprises to predict more accurately when key systems will become operational.
Greater opportunity for innovation. Business managers freed from inflexible technology models can focus on the business process, not the technological implementation. This shift can result in incremental improvements to business operations, and organizations can experiment with a variety of service-based business models that would not have been feasible without an open, collaborative technology infrastructure (see case study: Continental Airlines). By expressing core competencies as software services, a new realm of revenue streams can develop; at the same time, organizations that are not major users of IT can outsource it to companies that can do it more efficiently.
Implications for Enterprise IT
Case Study: Continental Airlines
Situation: Like many businesses in the airline industry, Continental is facing challenges as a result of a slowing travel economy, but the company knows it must continue to provide a high level of service to encourage its customers back on board its planes.
Challenge: How can the company provide real-time data of flight status in multiple contexts, such as interactive devices and customer service agent consoles?
Solution: Continental took the data it already has in its legacy flight operations management systems and “wrapped” it with web services communications standards.
Benefits: The web services offering has helped provide a valuable customer service. A relatively small amount of effort and investment makes the information accessible from devices such as cellular phones or PDAs. The basic infrastructure is scalable and minimizes the amount of custom integration necessary to create usable data feeds. These partnerships might result in a co-marketing agreement that generates direct revenue and goodwill.
More than ever, IT decision-making is tied to economic imperatives, and CIOs must balance technology investments with business realities. Interviews with leading IT executives reveal that enterprises are spending more of their budgets on external projects that generate revenue or improve relationships with customers or partners. Because web services projects can be executed with relatively small overall project scope, such as exposing customer data to customers for self-service access, potential risks are greatly reduced (see case study: Robertson Stephens).
Standardization and efficiency summarize the internal focus. Most internal projects are focused on making existing processes more efficient. A common element of these efforts is standardization on a common software package such as application servers. Many IT leaders describe them as a reaction to a series of acquisitions or an entrepreneurial organization that has moved in too many directions too quickly and without supervision from a centralized IT architecture group.
Managing Diverse Legacy Systems
Existing IT infrastructures are filled with legacy systems and applications that cannot easily connect to one another. Some analysts, in fact, estimate that more than 50 percent of the Global 2000’s operational systems depend on mainframes or other closed systems. Moreover, as an increasing number of corporations merge, or begin to share increasing volumes of operational data with business partners, the challenge of integrating disparate systems only increases.
When businesses first engaged in electronic transactions with trading partners, expensive private data networks and electronic data interchange (EDI) systems were the only available choices. Companies turned to enterprise application integration (EAI) to piece together these disparate systems into a functional framework. Generally, EAI solutions were expensive, dependent on proprietary interfaces, and demanded substantial hardware and staff resources to implement. In fact, Forrester Research estimates that as much as 35 percent of IT staff resources are used for “gluing” together applications. When this “glue” is standardized, the plumbing level of transactions will require relatively little time to develop for web-services enabled systems. This fundamentally changes the cost structure of integration. Additionally, the number of transactions among diverse partners can easily proliferate.
Large enterprise resource planning (ERP) and customer relationship management (CRM) expenditures will experience a higher ROI because they will easily connect with more points throughout the enterprise. A primary benefit of web services is the ability to salvage existing software assets by integrating legacy systems with potentially no recoding necessary.
Once the labor involved in integrating systems has been significantly reduced, enterprises can pursue a best-of-breed approach to software applications and components without as much concern about interoperability, which further increases the organization’s flexibility and strategic options. In these ways, web services work as the missing link for B2B commerce by providing the means to connect to legacy systems, partners and customers in a scalable way.
The Evolution of Enterprise Software
In the context of these ongoing issues in corporate IT, today’s enterprise software market is characterized by two significant trends. Both trends underscore nearly every vendor’s partnerships and competitive strategies, product releases and marketing messages. First is the rise of the J2EE application server and the subsequent market share dtente between Java and Microsoft-centric software architectures. Second is a growing recognition that most of today’s existing software segments are converging on one space: the integration and coordination of disparate technology systems into meaningful business processes.
The Rise of the Application Server and the J2EE-Windows Détente.
Case Study: Robertson Stephens
Situation: The investment bank Robertson Stephens developed a transactional system in Java, but its employee portal was developed on Windows technologies. The transactional system collected research data from external sources, primarily other investment banks with which the company partnered.
Challenge: The bank wanted to make this data available to its institutional sales team, as well as some client-facing groups. Replicating the transactional system on the alternative system would have been costly and time-consuming.
Solution: The bank turned to web services to solve the interoperability complexities. Web services facilitate the flow of information between disparate operating systems.
Benefits: The firm’s institutional sales team had more information without any additional software development, allowing team members to be more productive, and the client-facing team could distribute relevant information to clients in a more timely fashion, increasing customer satisfaction.
The emergence of e-business and Internet-centric development models in the late 1990s prompted the rise of the so-called “application server” as a container to develop server-side functionality. The basic model was simple: The application server provided a runtime environment and development tools for abstracting connections and implementing business logic between the web browser client and the database.
Initially, a number of vendors offered a wide range of proprietary variations on the basic model. However, the vendor-independent J2EE specification has become the dominant framework for developing enterprise software applications. In fact, the application server has become so ubiquitous that it could be described as an “application operating system” that plays a role traditionally held by platform-specific operating systems like Microsoft Windows, Unix and others.
Partly as a reaction to the increasing programming abstraction provided by the J2EE model, Microsoft has gradually incorporated interfaces to once-separate products such as databases, web servers and hardware drivers into a higher level framework now known as “.NET” (pronounced “dot-net”). The .NET framework allows software developed around Microsoft’s Win32 platform to realize many of the same benefits of device and runtime abstraction that J2EE has delivered on other platforms.
In essence, the market for new enterprise software has gravitated toward just two centers: J2EE and Windows. Despite substantive arguments from advocates on both sides, neither side is likely to make inroads into the others’ developer base. These dynamics, although competitive on a case-by-case basis, reflect both an overall stability and a fundamental dtente between the two platforms. We expect the status quo to continue with relatively little change in market share by either primary camp in the short term.
Software Converges on Questions of Integration.
The attention to platform software marketers has grown from stand-alone containers, such as the application server, or from single-server operating systems to include the space of application integration. In part, this is a result of the increasing parity in the market for operating systems and application development frameworks. Vendors of packaged applications such as ERP or CRM software suites see the future of their markets as less about the merits of particular, self-contained software, and more about how well these packages incorporate data from disparate sources and maximize the reach and flexibility of the information they process. As application vendors incorporate interfaces based on web services standards, the proprietary “adapters” that EAI vendors historically have used to develop competitive advantage could quickly become commodity assets. Indeed, the major battleground for shaping the software market is squarely centered on the very challenges of integration and flexibility that proponents of the web services model hope to address.
This increasing pressure on either end of the services management and integration spectrum —from application servers for creation to packaged applications encroaching from the delivery side —leaves a relatively short window of opportunity for independent web services vendors to find a footing and to increase the breadth of their current offerings. Product vendors on either end of this range—for example, stand-alone development tools or portals—will likely suffer the greatest impact from larger vendors, while services management and integration platforms offer the greatest potential for creating viable, independent offerings (see figure 3).
Figure 3: Web Services and Integration Software Battleground
As a result, the rise of web services will drive three clear trends in the realm of enterprise software infrastructure:
- Enterprise software applications will move away from the “suite” architecture toward an unbundled model as customers increasingly look to combine best of breed components using web services.
- Application servers will incorporate increasing amounts of EAI-like functionality and assume a broader focus—moving from a container for components to a manager of interfaces.
- Enterprise Application Integration platforms and new web services solutions will offer two clear alternatives to solving the challenges of managing the interactions among software systems.
Web Services Are Positioned for Growth
As observers of several waves of technology marketing, we know that enthusiasm about adopting web services will race ahead of reality for a year or more. While growth and expectations are perceived as linear, actual adoption will move ahead until a “tipping point” is reached, at which point adoption will surge exponentially. The adoption of web services will follow a direct progression from patterns characteristic of early adopters to those that reflect widespread acceptance for mainstream uses (see figure 4).
Figure 4: Web Services Market Growth—Reality vs. Perception
Phase 1: Grassroots adoption. This first stage, already underway, emphasizes internal integration projects that are driven by cost efficiency and the ability to use existing assets. Companies in this phase are beginning to take advantage of the fact that web services technologies break down monolithic applications into much more discrete components. Still, this phase is mostly one of ad hoc adoption. Adopters are using web services for marketfacing projects in addition to internal projects, many of which were previously considered too expensive. Connections at this stage are built when the known endpoints are clear and the service is simple. Our research confirms that web services are being embraced by the usual early adopters: financial services firms, high-tech manufacturers and telecommunications companies, along with a few less obvious industries such as automotive and insurance. Public web services are already available on the Internet: These freeware services have simple functions such as calculators, and include no security component.
Phase 2: Systematic deployment. Companies begin to shift toward systematic strategies for developing, publishing and using web services. This stage of growth will be characterized by a push toward increasing control and creating transaction-oriented networks. Although some opportunities for establishing new revenue streams will exist, corporations will focus on streamlining existing processes. Vertical, industry-specific web services standards and components will become prevalent.
Phase 3: New business models. Optimists envision new business models that are enabled by web services and defined by truly dynamic business relationships that are formed and dissolved in real time. Enterprises will use web services to create new ways of deploying applications. Knowledge workers will be empowered to create business processes without the need for substantial IT guidance. Early adopter industries may be more receptive to experimenting with this vision than the more traditional sectors.
Vendors Bring Views and Products to Market
The 12 Most Influential Web Services Companies
Be they emerging innovators or major platform vendors, the complementary “coopetition” of software companies is a major reason for the web services model’s early momentum. These 12 leaders stand out for their role in shaping today’s web services implementations.
Platform Leaders
- BEA: Strategic acquisition of Crossgain and market leader J2EE
- IBM: Standards body leader with undeniable enterprise credibility
- Microsoft: Compelling developer tools and ubiquitous market share
- Sun: Chief owner of J2EE specifications; enjoys significant developer loyalty
Significant Influencers
- Bowstreet: Earliest promoter of web services to enterprise customers
- Cape Clear: Early toolkit success and strong financial resources
- Grand Central: Visionary model of shared web services infrastructure
- Iona: Leader in CORBA deployments
- The Mind Electric: Leading toolkit with significant developer buy-in
- Tibco: Leader in EAI market and conceptual support for services model
- UserLand: Co-creator of SOAP and XML-RPC standards
- Xmethods: Testbed and directory for grassroots services deployment
Over the past several years, software vendors have invested aggressively in creating standards and developing web-services architecture (see sidebar: The 12 Most Influential Web Services Companies). IBM has adapted their entire product lines to web services protocol standards. Microsoft has admittedly taken the largest gamble by significantly changing its architecture and developing a series of services (dubbed “My Services”) for immediate consumption. Some mid-tier vendors have fully reinvented their companies, while others have simply enabled their integration or portal offerings with the web services standards. Start-ups have emerged to create innovative approaches to a specific area in the infrastructure.
As might be expected in such a fluid market, software vendors are delivering mixed messages about what web services mean for their customers and how their own products will play a role. Much of this variance stems from vendors’ natural instinct to protect their current franchises, and to maximize the value of the assets they can bring to bear.
For IBM, this means focusing on messaging software and integrating legacy infrastructures. For Microsoft, web services turn the conversation toward developer tools and individual productivity applications. In the case of BEA, they represent a natural extension of the market-leading application server. Web services offer Sun an architecture that uses Java on multiple forms of hardware. And for others, web services advance the cause of models such as EAI, distributed computing and collaborative software design.
Notwithstanding their separate interests, vendors agree on several points when discussing web services and service-oriented architectures. They stress that this is an evolutionary improvement over previous technologies. Indeed, the most remarkable difference between web services and similar past initiatives may be that this wave of innovation champions pragmatism: It is standards-centric, it minimizes complexity and it can be adopted incrementally.
Most vendors concur on the lowest common denominator definition of a web service—a collection of operations that are network accessible through standardized XML messaging, SOAP and WDSL descriptions. Finally, all vendors pay some degree of service to the idea that today’s IT environments are heterogeneous and that interoperability is a fundamental requirement for any new offering.
Adapting to a New Environment
The emergence of web services will have a significant impact not only on software companies and enterprises, but on IT service providers as well. These companies will need to shift away from their traditional models of integrating disparate data sources and applications and toward services based more on process consistency and higher-level business oriented platforms. The companies that make this important shift will be best positioned to provide the services most critical to enterprises that are adapting web services.
Critical Success Factors
Even the most strident advocates admit that web services will not solve every problem. Neither do supporters expect that current software will be discarded blindly in favor of a new approach. Indeed, several factors may play an outsized role in fostering (or hindering) the scale of web services’ success.
Conformance to emerging standards. Without assured interoperability with partners or customers, many of the fundamental technological and economic benefits that web services offer will be lost. Although buy-in from vendors and customers alike has propelled basic web services forward, current standards are still emerging. While top-to-bottom agreement is not essential, adoption of the basic standards that are already well formed is critical. Higherlevel layers that define strategic aspects of business processes must be developed for use on the same scale as the other layers of the stack. For example, the proliferation of graphical environments for creating, using and connecting components will help foster a larger number of available web services, and will help to promote the adoption of web services as a standard application API.
Optimization of data. XML is a lengthy data format, and the very richness of its semantics results in tremendous overhead in transmission. Some estimates measure the size of a typical XML document at 10 times that of an equivalent EDI transmission. Today’s XML parsers, while capable, must still incorporate significant optimizations before the overhead becomes a non-issue for processing latency. Some activities still require faster computing, such as financial engines, and will be unlikely candidates for this new distributed computing approach.
Resolution of security and authentication questions. There is no security standard built into a web service interface yet; companies currently rely on existing firewall, network and application security systems. Therefore, partners must create a dialogue between connecting partners to agree upon the type and level of security—which then limits the scalability of the service. This will not create a functionality barrier for anonymous or strictly internal web services, but implementations of strategic business processes in a highly scalable web services context cannot proceed to their full potential until a standard is adopted.
Experimentation with service-based business models. New business models will be driven by ubiquitous, inter-enterprise collaboration. Companies must allow room for some experimentation to find new trading partners, and for assembly of real-time business processes. Additionally, nearly all current web services are free. Successful pricing models still require substantiation. Nonetheless, even if web services do nothing but make traditional application integration easier, they will have made a significant contribution to IT.
A critical mass of deployments. The benefits of web services are realized in part by the efficiencies of software component reuse. Without a critical mass of services available for consumption, enterprises will need to continue to re-create applications from scratch, and the model ceases to be as effective as it could be. Seeing an opportunity to help solve this problem, in early 2002, many major vendors began shipping toolkits that reduce this challenge. These easy-to-use tools should provide a substantial catalyst to the market and increase the number of web services deployments.
First Steps Using Web Services Today
Companies that have successfully begun to work with web services exhibit several common characteristics. These best practices can be described as “pragmatic first steps.” Vendors, too, should think about how to best enable their customers to begin using web services, not necessarily how to “own” the next great wave of technology spending.
Use what you have and be opportunistic. The economic benefits of a web services approach are predicated on the concept of incremental investment. In practical terms, this means leveraging, not replacing, current infrastructure. It also means taking advantage of current IT skill sets. For example, if your shop is skilled primarily in Java, it is not worth moving to a Microsoft .NET infrastructure (or vice versa) for the sake of a particular web services implementation. Moreover, even though standards are a big part of the web services equation (where standards are already well formed), do not worry about being fully standards-compliant today. It is more important to make use of a subset that works, such as XML-supported systems, than to wait for the “perfect” web services stack to evolve.
Illustrate clear benefits—even small ones. A simple pilot project should solve a real problem and articulate how web services will have a material benefit to the organization (such as strategic redeployment of resources, decreased time to market and improved CRM).
Start small, but plan for scale. Start with a discrete, well-defined project so that the results and benefits are tangible. At the same time, bear in mind that web services’ real value is demonstrated in their ease of reuse and portability. Rather than thinking of pilots as a one-off experiment, design the project with the second and third iterations in mind.
Hold off on transactions that require a high level of security. Regardless of whether you start inside or outside the firewall, remember that transaction-oriented protocols and many security-related issues still must be solved. Start with non-mission-critical processes that can accommodate the hiccups of an emerging technology model’s learning curve. In the interim, XML-supported mature middleware is a good alternative for transactions because they are the easiest to integrate with web services.
Is Your Company Ready? Some Questions to Ask
Consider the question, “Are web services the right approach for your company’s next IT project?” Clearheaded consideration of these openended questions will go a long way toward finding the right answer.
- What areas of business are our core competencies? What drives our competitive advantage? What could be outsourced to partners?
- Have we talked with customers and partners about web services? How can IT increase our value to both constituents? Do our plans present any channel conflict issues?
- What parts of the business are most likely to be “services enabled?” What do we want to use for a proofof-concept?
- Do we have a plan for evaluating the project? What are the metrics and milestones?
- What do we expect vendors to deliver in the web services space? What are our requirements for industry-specific standards?
- Do we understand the impact of increased IT flexibility? How have we managed technology in the past?
Conclusion
Thanks to the incremental investment that is necessary to experiment, and the momentum of the basic protocol standards adoption from both vendors and enterprise customers, web services are making inroads into the foundations of enterprise IT. The ability of web services to incorporate legacy systems and salvage existing assets in a scalable way will strengthen its case for adoption (see sidebar: Is Your Company Ready? Some Questions to Ask). The overwhelming investment from platform vendors assures that web services will change the way IT systems work at the most basic level, but to what degree it changes the way enterprises function is still unclear. Although the full effects are difficult to gauge at this early stage of development, companies are beginning to appreciate the benefits in efficiency and scalability that web services offer in the way of creating applications and connecting systems inside and outside the firewall.
Praveen Madan and Rachel Pyrdol of A.T. Kearney also contributed to this article.
Copyright © 2002 The Stencil Group and A.T. Kearney
Republished with permission from the authors
Brent Sleeper and Bill Robins are co-founders of The Stencil Group and
authored this report in conjunction with management consulting firm
A.T. Kearney. The Stencil Group works with software companies to
understand the business drivers and strategic priorties that shape
their enterprise IT customers’ purchase decisions. The firm focuses on
the bottom-line business impact of technology solutions through
consulting services that include customer needs analysis, product
roadmap and positioning evaluation, and market-facing sales support.
