The E-Business Software Weekly is a series profiling trends and developments in software and applications that support e-business, the Internet, and other electronic communication channels. Look for a new story each week in this space.
Managing Business Processes--Efficiently
During the nearly three centuries that the classical view of atomic physics held sway, matter seemed stable and irreducible. All atoms, physicists claimed, were comprised of three fundamental particles--protons, neutrons, and electrons--that represented nature's lowest common denominator. Atomic matter could be fused or split into these constituents as one might wish, producing outcomes ranging from molecular novelties to the catastrophe of a nuclear explosion, but humankind could peer no more deeply into the underlying matrix of nature than this.
Such a state of affairs placed few practical limits on the probative power of human understanding. To be sure, a few of the Einstein-class equations of the emerging grand unified theories resisted solution, but atomic engineers were more than satisfied with the atom-smashing "chunkiness" of the three-particle view of the world. After all, endeavors ranging from medical radiology to nuclear deterrence required no more granular understanding of matter than the three-particle view afforded, making deeper explorations an expensive and unnecessary intellectual sideshow.
For a while, anyway. Beginning in the post-Manhattan Project 1950s, this once-stable foundation began to shake, and then shatter. With the physicist Murray Gell-Mann's speculations that protons, neutrons, and electrons were themselves assemblages of even more basic particles called "quarks," the atomic world suddenly appeared less like a quaint three-ring circus and more like an anarchic New York intersection. Such theories withstood the test of ever-more advanced particle accelerators at CERN and Fermi, to the point that, by the end of the twentieth century, more than 100 new varieties of sub-atomic elementary particles had been discovered. And they brought with them many practical benefits, leading not only to the resolution of some of Einstein's thornier equations, but giving rise to whole new fields of study, like sub-atomic engineering and quantum-powered computing, that promise to revolutionize day-to-day life every bit as much as did the splitting of the atom half a century before.
Information Technology's Operational Challenge
A similar awakening, as it happens, is just starting to take place in the equally rarefied world of e-business--and the ultimate consequences could be just as profound.
The story starts on familiar ground. Since the advent of modern information technology, technologists and business strategists have designed wide-ranging technical infrastructures that have enabled corporations to dissect and disseminate business-critical information in ways never before possible. Some of these developments led to the creation of stable, replicable high-level technical structures, like client/server computing architectures and electronic data interchange (EDI)--the IT equivalent of protons, neutrons, and electrons--but most of the resulting business processes remained invisible to the naked eye, peculiar to the operation or organization that originated them. As a consequence, each business process required a unique, labor-intensive development effort that proved expensive to complete and difficult to maintain. But there being no alternative, financial officers and IT managers swallowed hard, wrote the checks, and hoped for the best.
The full scale of the resulting operational challenges has become apparent only over the last several years. As Ishmael Ghalimi, CEO of Intalio, Inc., a Burlingame, Calif., based software developer, explains, "The [corporate] enterprise has been using increasingly complex processes throughout the last century to maintain its operations, but we have yet to see an IT infrastructure capable of fully supporting these processes." The reason for this, he says, is twofold: first, "packaged enterprise applications such as Enterprise Resource Planning (ERP) systems manage only those parts of the business process that had to be automated at any cost," and, second, "the deployment of these processes upon some IT infrastructures requires prior organization of enterprise data, a formidable process that has taken several decades to complete."
More than a mere technical quandary, these highly customized undertakings have spawned shock waves that have reached as far as the boardrooms and executive suites. Consider the results of the classic 1999 study "Migrate Headaches" by the Standish Group, a technology research firm based in West Yarmouth, Mass. In examining a wide range of firms across several industries, the research firm found that only 17% of e-business initiatives requiring corporate systems integration were wholly successful. Some 49% overran their budgets, and a full 34% failed, meaning that such projects were twice as likely to fail as succeed. Similarly, a 1999 study by The Meta Group, a Stamford, Conn., based IT research firm, confirmed that e-business integration initiatives in many of the world's largest companies were at "serious risk of failure."
The price tag: billions of dollars, years of effort, and an irreplaceable loss of competitive position.
Managing B2B Integration
The Internet, fortunately, is about to change all this. The Internet, says Silicon Valley based business analyst Ryan Madden, "is a convenient medium for the automation of collaborative business-to-business processes," one that is "much more flexible than traditional EDI." These capabilities will make it possible for enterprises to "integrate their respective business processes over the Internet, enabling them to achieve greater levels of efficiency, create additional revenue opportunities, and enter new markets."
This opportunity, however, comes with new challenges of its own, for business strategists as well as IT managers. "If ERP packages requiring massive process reengineering could only automate [business] processes internal to the company," argues Intalio's Ghalimi, "it is unlikely that any packaged enterprise application will be able to automate the possible interactions between the processes of multiple business partners." There are both theoretical and practical reasons for this. From a theoretical standpoint, "the complexity of integrating multiple processes increases exponentially" with the number of processes involved, thereby preventing the use of any one packaged solution; from a practical perspective, business partners typically possess widely varying IT infrastructures supplied by a number of different vendors, and so "cannot rely on any single solution."
Ghalimi notes that the problem of integrating IT infrastructures in order to enable these seamless business-to-business interchanges will be handled, at the presentation level, by a variety of XML-based initiatives already underway. "Nevertheless," he insists, "there remains the question of integrating these processes with existing IT infrastructures that consist of multiple packaged enterprise applications," e-business systems that soon will be complemented by a host of new applications for the automation of additional business processes, including procurement, fulfillment, and service-chain management.
Because such integration cannot be achieved with packaged enterprise applications or collections of pre-defined processes, says Ghalimi, it will demand "a new kind of enterprise software infrastructure that will allow business analysts to model, deploy, and manage... business processes that span multiple enterprise applications, corporate departments, and business partners..." Twenty years ago, corporations found a way to manage a growing flood of data with then-emerging database management systems. Over the next few years, Ghalimi predicts, companies will likewise find a way to manage "an exploding complexity" of business processes with a comparable technical architecture, an e-business model termed the "business process management system."
Enter Business Process Management
The genius of database management systems was that they enabled enterprises to extract critical business data from the information-processing framework of IT systems in which the data resided, in much the same way that Gell-Mann and his intellectual compatriots liberated the constituent quarks from the elementary particle "shells" that housed them. By separating out data in this way, companies were able to feed and process their data through an infinite variety of data-management systems with only minimal reengineering, creating new and highly valuable business-relevant views of data at little marginal cost.
The motivating premise of business process management (BPM) is much the same: by segmenting business process components from the larger enterprise management systems that contain them, these modularized processes can be rapidly deployed within and among multiple applications and companies, greatly enhancing the replicability and efficiency of Internet-based business-to-business interactions. In the wake of the increasing cost pressures on mainstream companies' e-business operations, this could be a vital development: because the engineering required to connect various business systems has been so resource-consuming, eliminating a sizable portion of these costs would allow business partners to focus on the commercial rather than technical opportunities offered by their relationships, accelerating the creation of productivity-enhancing e-business offerings.
So bright is the promise of business process management that more than 100 companies, including Computer Sciences Corp., Deloitte Consulting, Hewlett-Packard, IBM, and Sun Microsystems, have joined in a global consortium, barely a year old, called BPMI.org, or the Business Process Management Initiative. Administered by Intalio, Inc., the initiative's mission is to promote the use of business process management technologies through the establishment of standards for process design, deployment, execution, maintenance, and optimization. BPMI.org's main role, at present, is the development of open specifications that will enable the rapid development of BPM-based projects within its B2B-focused member companies.
The organization took the first major step along this path three months ago with the release of a specification for what it has christened the Business Process Modeling Language, or BPML. Like XML (the eXtensible Markup Language, profiled in an earlier column), BMPL is a meta-language that describes the common characteristics of business processes in a such a way that these processes can be deployed in a variety of differing computing environments without requiring extensive reengineering or customization. In the same way that XML is used to identify, or "tag," distinctive data types, BMPL is used to tag key business process module types within an overall e-business system. BMPL is complemented by a structured query language, called Business Process Query Language (BPQL), that enables business analysts to easily investigate the status and functioning of business processes operating as part of the e-business system.
Howard Smith, European chief technology for El Segundo, Calif., based IT consulting firm Computer Science Corp., who helped draft the BPML specification, recently told Computerworld that he believes that BPML could unlock business processes in much the same way that SQL did for databases in the 1980s. "Until now, there's been no way to get a view of a business's entire processes, to link them, to layer them," he explained. "BPML will allow us to run those processes independently of applications," thereby enabling collaborative e-commerce to take place "at a deeper, more integrated level than the current application interface model" makes possible.
From Application-Centric to Process-Centric
This application-centric approach to e-business systems at best discourages B2B collaboration and integration, and at worst prevents it altogether. Today, says Bruce Ambler, senior manager for IT software architecture strategy at Lucent Technologies, if a company wants to modify the business processes comprising one of its applications--say, opening up a quote-generation system to accommodate a new business supplier--the company likely would have to change all the processes connected to the target process. "If you go to change any of those things," Ambler recently told Internet Week, "you're obligated to change other things that [the business processes] are hooked to."
By contrast, if business processes were separated out from their host applications, said Ambler, a company could "assemble them and disassemble them more easily to create a string of processes" that could be combined into any desired application configuration, just as atomic engineers now routinely assemble quarks and other sub-sub-atomic particles in order to create new quantum entities. Indeed, this is the ultimate goal of the business process management paradigm. "Business Process Management Initiative members," writes Internet Week's L. Scott Tillett, "envision a day when business processes, like data, can reside in their own management systems," where they can be analyzed and dynamically reconfigured in order "to determine the best way of conducting business..."
Enterprises and their customers stand to gain enormously from this process-centric approach to e-business infrastructure. "The key characteristic of the business process management system," says Intalio's Ishmael Ghalimi, "is its unique ability to be used for mission-critical applications." As a result, BPM implementations offer the prospect of both accelerating time to market and enhancing ongoing operational efficiency--critical financial safeguards in a hyper-competitive era in which even a few percentage points of margin can spell the difference between survival and demise.
ADVERTISING FEATURE
