The Future of SOA-based Applications and Infrastructure
Olin Thompson and P.J. Jakovljevic
For the Big Few vendors (primarily Oracle, SAP, and IBM) focusing on the technology lock-in of users will inevitably have to be replaced by focusing on delivering the best solution for the customer, even if it means using some components from competitors. The ultimate winner in the market will have to provide industry-specific solutions that solve essential problems that others cannot: deep domain experience coupled with industry-specific product functionality which insures successful implementations; an integrated suite of industry-specific products that satisfy current customer requirements; and an enhanced roadmap for customers potential long-term needs (for more information, see If Software Is A Commodity—Can You Still Win Some Competitive Advantage?).
Part Three of SOA-Based Applications and Infrastructure The Next Frontier? series
Despite these large vendors professed dedication to a partner network around their preferred platform, they face a number of challenges in executing the vision, since opening up the software for others to build potentially competitive products will be a major cultural adjustment. In addition, these giants do not have an impeccable track record of sustaining long-term partnerships with ISVs. Some partnerships have not always gone so smoothly in the past (see Gosh, They Kill Partnerships, Don't They?). As these large vendors push more aggressively into the enterprise applications-plus-infrastructure market, they are increasingly bumping heads with many large industry players, some of which are (or should be) important development, consulting, or integration partners.
Also, in an open platform, there is a risk when working with competitors. There will be competition, perhaps even from their own products, which makes opening up to the outside world quite questionable. However, the idea is that everyone contributes and shares in the platform network "pie" and the leaders get feedback from other industry players as it is being built. Still, many things will have to be clarified in such an ecosystem approach, including how much it will cost third parties to access the available services in the repository and what any eventual certification process will entail.
Dispelling the perception of being proprietary while leveraging rich legacy code is another challenge for the likes of SAP. Its legacy code is built on the SAP Advanced Business Application Programming/4 (ABAP/4) language, and need to be extended through a set of Web services description language (WDSL) interfaces as a set of services that can be incorporated into a set of applications.
This is Part Three of a three-part note.
Part One discussed SOA and its impact on business.
Part Two examined SOA as a foundation for a universal desktop for all the Web-based applications of an enterprise.
Multi-Enterprise Service Architecture (MESA) Factor
A successful, and thus winning appli-structure, will not necessarily be the one that comes first to the market. Rather it will be one that features continuously lower IT total cost of ownership (TCO), enables the quick and flexible reconfiguration of business processes, allows different vendor solutions and even legacy systems to plug-and-play impeccably and securely, and allows for product upgrades or enhancements.
Over the past two to three decades, the software marketplace has evolved considerably—from mainframe computing to the client/server model to Internet e-business, and now on to today's SOA computing model. These changes in industry focus have historically driven vast improvements in return on investment (ROI), interoperability, and faster time to market. But, as with any technology, each step has eventually reached a plateau in terms of improved productivity. Overall, one can see that the industry has followed a path of increasing focus on business issues, on less rigid architectures, on greater distribution of intelligence, and on ever-greater flexibility of response, which is a movement toward decentralization and collaboration.
Efforts to increase performance through the integration of applications and processes inside the four walls of an organization have worked well for two decades, but they have also reached a barrier. First of all, large-scale software packages, such as ERP applications, have been commoditized to the point that the playing field has become level. Second, increasing efforts to integrate internally are yielding only marginal improvements in organizational efficiency and the bottom line. Third, the era of cost cutting is reportedly drawing to a close, as businesses once again focus on expansion and growth. Thus, businesses are moving toward delivering data, processes, and applications that may reside outside the enterprise in whole or part. Some of these will, as before, be wholly owned and controlled by specific organizations, while others will be externally hosted, outsourced, or shared. The overall trend, however, is toward service orientation, working well with others in a business context, and moving toward a growth-directed collaboration.
At the same time, there has been a movement in the industry away from data, processes, and applications that are wholly owned, controlled, and reside completely within the single user enterprise. The traditional model for enterprise computing involves applications sitting atop a corporate infrastructure. In this model, all business logic is hard-coded into applications such as ERP or accounting systems, spreadsheets, or word processors. To that end, multi-enterprise computing is about the contact points between organizations, and about the necessary distribution of processes and application logic. Consequently, it weighed down less by giant, hard-to-change enterprise applications.
Back to Achieving Growth
There is a reason for sharing the burden across organization boundaries nowadays: there is simply too much for any single organization to control. It is virtually impossible to stay nimble and adaptive while worrying about applications, networks, the corporate infrastructure, and so on. The only sensible way to respond to all these demands is to federate and cooperate. The differences between the multi-enterprise computing and the traditional enterprise computing models are pronounced, although none of the elements of the multi-enterprise computing paradigm are entirely new. For example, SOA in the multi-enterprise computing paradigm would gain a B2B element and serve a real business need. Likewise Web services will become a community of Web services that can be used to advance collaboration, which demands that application logic be shared and computing tasks be distributed.
No technology should be used just because it has been invented, and that has been the case with SOAs and Web services. A particular business need must be the driver. To that end, many CEOs are now increasingly returning to an emphasis on growth and jockeying for position at the head of the pack. In the forthcoming years, they expect to focus on using technology to drive top-line growth. Many currently see a number of sources for growth, such as more tightly integrated supply and value chains, strategic partnerships, better connections with customers, the ability to roll out new products and services more quickly, etc. For a detailed discussion of how vendors are approaching growth see Achieving Growth: New Accounts versus. Up-selling to Existing Accounts.
All of these have a common thread of relying on the ability to effectively link customers, suppliers, employees, partners, and communities together. In other words, they are all about working well with others. From a high-level perspective, multi-enterprise computing also represents a move from being application-oriented to a focus on business processes, in terms of community building and information sharing. However, these have not yet become the fortes of traditional ERP vendors, but are rather the strengths of the likes of Sterling Commerce, Inovis or General Exchange Services (GXS) which have been filling this roll for thirty years or so. (See GXS Acquires HAHT Commerce for More Synchronized Retail B2B Data).
Former value added network (VAN) services providers that still offer electronic data interchange (EDI) transport and translation services are currently engaged in the most radical re-engineering of their businesses since the onset of Internet EDI over a decade ago. Now they have been formalizing multi-enterprise service architecture (MESA) computing and giving it a coherent framework, which all enterprise applications vendors should observe and adhere to. To that end, traditional EDI and guaranteed message delivery services are being enhanced with a new series of integration and application services designed to enable inter-enterprise business process execution, collaboration, and management for expanding supply chain management (SCM) and customer relationship management (CRM) strategies.
Additional Vendor Services
Also, these vendors' services will now often include data cleansing and syndication, product information management (PIM), cross-platform and inter-enterprise transaction management, document reconciliation and analytics, trading partner BI, and an array of applications to assist businesses with collaborative processes, ranging from dispute and financial process management to demand planning and vendor-managed inventory (VMI).
To achieve these new market requirements, Sterling Commerce recently announced its MESA collaboration strategy and a partnership with Entrust, whose software and services ensure the privacy of electronic communications and transactions across corporate networks and the Internet. Entrust addresses such functions as identification, verification, privacy, and security by authenticating users via smart cards, passwords, and biometric devices, controlling access to information in e-mail, databases, Web pages, and business applications.
On the other hand, GXS, which has a customer base of 40,000 corporate customers with an ecosystem supporting more than 100,000 trading partners, has introduced its Trading Grid architecture that features messaging services, BI services, and application services, with adapters to link to a slew of enterprise applications packages. GXS acquired the EDI services business from IBM, and announced a partnership with the EAI provider webMethods, Inovis has announced the acquisition of a former EDI competitor and PIM provider QRS (see Inovis Delves into PIM by Snatching QRS), and Perfect Commerce has acquired the former Internet marketplace for utilities Pantellos.
Along similar lines, SAP has quietly unveiled IndustrySpeak, a new effort to work with industry standards groups to enable collaborative business processes and scenarios. The emphasis will first be on the electronics/high-tech and chemical industries, by incorporating transaction sets for RosettaNet and Chemical Industry Data eXchange (CIDX). However, the idea is to expand into many more industries over time, such as agrichemicals, oil and gas, automotive, aerospace and defense (A&D), consumer products, mill products, retail, and financial services.
Although the widespread acceptance of Web services inter-enterprise implementations will not happen any time soon, the major players' involvement in leveraging these should prompt large global enterprises to start learning the new protocols, standards, and technologies. Doing so will help companies grasp the potential, underlying business advantage. They should try to understand how developers will leverage Web services, SOA, and business process management (BPM), know what their own ongoing needs are, and what intricacies may arise using these services, such as cultural and standards issues. To casual and power users of business applications on a desktop, conversion to SOA should be fairly transparent. This, however, will not be the case for the army of IT staffers though. The move will represent a significant reworking of the IT infrastructure that is typically a complex mishmash of disparate technologies.
While SOA and Web services may make integration easier through some imposed interoperability, they will not eliminate the need for application integration via adapters, connectors, or other items, since they are not any kind of panacea, nor are they a replacement for event-driven architectures (EDA). Also, although they may provide new business opportunities and create some dynamism and efficiency, they are not going to transform businesses on their own, given that SOA and Web services are only pieces of new technology. They are more flexible and open architecture, but depends on the availability (a critical mass of exposed applications) of services, and whether they can be exposed as services. On the other hand, traditional EDA is more tightly coupled, and in many circumstances will offer higher scalability and performance. As a result, users will ideally need both approaches.
Even if large organizations are the first to dabble with deploying Web services, their impact will ultimately be felt by companies that supply products and services for consumers, regardless of their sizes. Businesses can already create and use Web services that can be reused by other applications, such as services to authorize credit cards or to authenticate a person's identity when logging onto several systems.
Benefiting from Web Services and SOA
Yet, to fully benefit from Web services and SOA, companies should carefully and painstakingly reexamine their business processes and "best" practices and look for the efficiencies a revamped infrastructure could bring. The starting point for building an SOA blueprint is to identify and create Web services around common business reference objects for the entire organization. This development will be largely dependent on the organizational strategic alignment and the industry in case.
While many companies are going to attempt building their own SOA blueprint, others may be driven by more complex cross-company driven SOA blueprint efforts, such as ones already initiated by IBM, SAP, BEA, or the Middleware Company. For example, IBM's SOMA approach uses techniques like domain analysis, process modeling, component-based development, and object-oriented application development, but also includes goal service modeling, which customers may use to determine the optimal granularity of a Web service. Still, the strictness and effectiveness of the innovative but yet unproven process will have to be closely watched down the road.
About the Authors
Olin Thompson is a principal of Process ERP Partners. He has over twenty-five years experience as an executive in the software industry. Thompson has been called "the Father of Process ERP." He is a frequent author and an award-winning speaker on topics of gaining value from ERP, SCP, e-commerce and the impact of technology on industry.
He can be reached at Olin@ProcessERP.com
Predrag Jakovljevic is a research director with TechnologyEvaluation.com (TEC), with a focus on the enterprise applications market. He has nearly twenty years of manufacturing industry experience, including several years as a power user of IT/ERP, as well as being a consultant/implementer and market analyst. He holds a bachelor's degree in mechanical engineering from the University of Belgrade, Yugoslavia, and he has also been certified in production and inventory management (CPIM) and in integrated resources management (CIRM) by APICS.