Home > Articles > Software Development & Management

  • Print
  • + Share This
This chapter is from the book

1.6 -E-Business Meets Aging Hierarchies and Infrastructures

Legacy computer systems process the majority of sales, customer service, billing, distribution, inventory management, procurement, and related revenue-generating, cost-saving and accounting tasks. While most companies would not be able to function without these systems, this was not always the case. Examining the evolution of legacy architectures and the organizational infrastructures upon which they were based provides insights into how to leverage these systems to meet time-critical business requirements while concurrently deploying emerging technologies.

Industries and governments across the board use computer systems that are a reflection of the past. The past, in this case, takes the form of management hierarchies pioneered by companies such as General Motors almost 100 years ago. Companies have historically been divided into functional business units. Each business unit performed specific tasks. Traditionally, this included marketing, sales, production, purchasing, accounting, and other speciality functions. Other business units were created to support unique product lines and international functions.

Business units reported up through a hierarchical management command chain based on these segregated functions. It was this environment into which computers first entered the business world during the 1950s and 1960s.

Businesses were wary at first of using computers at all. IBM founder, Thomas Watson, Sr., once said, "I think there is a world market for maybe five computers"9. This was a far cry from latter-day predictions that the disappearance of computers, by some accounts, would force over half the female population of the United States to become telephone operators.

In most companies, IT was originally under the accounting department and reported up to the chief financial officer (CFO). This decision was based on the fact that early computer systems typically performed accounting and other financial functions. As IT evolved, more systems were developed to automate tasks for other hierarchical business units. The internal IT department (originally called data processing or automated data processing) reporting structure mimicked the management hierarchies used across the enterprise.

Hierarchical subdivisions were extended further as systems evolved throughout the 1970s and 1980s. For example, a sales system would include order processing, sales tracking, sales forecasting, and various other sales functions. The IT sales team would then be subdivided into these subcategories. Figure 1.2 depicts a systems hierarchy chart that reflects both the infrastructure of the IT organization and the systems supporting hierarchical business units.

The specific functions depicted in Figure 1.2 are not of particular relevance. What is relevant is that almost every IT organization uses such a chart to depict how its systems and its people are organized and aligned with the business. This suggests the pervasiveness of the hierarchical management and information systems model.

Figure 1.2Figure 1.2 Hierarchical information infrastructure and systems definition.

The IT organization, and application systems and databases, were segregated silos. Information flow between these systems was at best slow and haphazard. Large mainframe computers would read data from one system and produce files or database updates that could then be processed by the next system. A large-scale legacy environment would repeat this sequential series of steps hundreds of times over the course of a daily or nightly processing cycle.

Hierarchies continued to dominate well into the 1990s as enterprise resource planning (ERP) packages emerged. ERP systems are third-party software packages that became quite popular during the 1990s. These packages were subdivided by the functions they supported, although most ERP systems provided for cross-functional data exchange based on a common architecture that was missing from most homegrown legacy systems. Today, ERP systems represent a small but important percentage of the overall population of legacy applications.

When viewing the evolution of legacy systems over the decades, one case study exemplifies how legacy systems have outlived numerous incarnations of the businesses they were built to support.

Many of the original computer systems built by AT&T are still in place as the telecommunications industry enters the 21st century. AT&T developed a number of systems back in the 1960s and 1970s. One of these systems was the Customer Resource Information System (CRIS), which provided billing and related services. When AT&T broke up in the early 1980s, each RBOC took a copy of the CRIS system and enhanced it as it expanded into regional markets and unregulated industries.

Now several RBOCs have been recombined. At the same time, local phone markets are being deregulated. Many of these companies have also gone global and entered a variety of new markets, such as wireless and broadband. Yet multiple copies of the old CRIS system continue to run day in and day out at these companies.

CRIS and other systems survived multiple business retooling efforts over a period of decades. Can global telecommunications providers thrive in global wireless, broadband, and other new markets using segregated customer management systems and databases based on a regional business model that no longer applies? It is unlikely. Is it cost effective to continue to run four copies of these systems within a same company? No. Is there an easy way to address this challenge? The short answer is no, but issues like this must nevertheless be addressed over the long term.

The introduction of the Internet and e-business is driving the need to more fully integrate business processes across old business and IT hierarchical infrastructures as well as across supply and distribution chains. As this occurs, organizations will need to shift to radically different business models. Businesses now have a new set of requirements.

  • Business processes must flow from the point where they began in one business unit across other business units in less time than legacy systems allow.

  • The transaction-oriented systems that supported segmented business processes should flow from one transaction to another in less time and with less human intervention. This concept is called zero latency, where these events are instantaneous.

  • Data stored in departmental databases should be integrated with data stored in other departmental databases.

  • Cross-functional business process flow, transaction flow, and data flow must occur in a matter of minutes—not hours or days.

As businesses and governments raced to the Web, they began building Web sites to address customer and business partner functionality. Many Web sites and new companies sprung up overnight. The majority of these Web sites had to be rebuilt, which reflected early missteps in assessing how the Web would help businesses expand market share or save money. Little dotcom companies with no infrastructure emerged to challenge established corporations. When they could not sustain a revenue stream, they either folded or were quickly swallowed up by more established enterprises.

In a very short period of time (from 2000 to 2001), the high-tech industry experienced rapid expansion and subsequently rapid compression. All of this was fascinating to watch because it unfolded so quickly, and if you questioned, it you were told that the new economy would be leaving you behind. Then reality sunk in.

As organizations began to Web-enable customers and suppliers, they found that they could not replicate or readily access complex business knowledge embedded in legacy systems and databases. These organizations needed to retool decades of entrenched hierarchical infrastructures in a short period of time. IT is at the earliest stages of this retooling process.

  • + Share This
  • 🔖 Save To Your Account