In-Depth
Modernization with a Twist: Keeping an Organization’s DNA Intact
Could replacing core systems cause an organization’s identity to unravel?
by Joe Gentry
As most students learn in school, DNA is a basic building block of life that contains in its structure all of the history of an organism going back generation upon generation. That unique history is stored in the arrangement of amino acids. The DNA also tells the cells how to behave; it tells the organism how to grow and how to function, and is the unique map to that organism's identity.
Core enterprise systems are the same way. Within these custom-built systems -- which often run on a mainframe and have been modified over many years -- is data that represents the history of how the organization has evolved. These systems also embody the unique combination of data and business rules that differentiate an organization and help maintain the organization's competitive edge in the marketplace.
Untwisting the DNA?
Just as breaking apart strands of DNA is a tricky job in biochemistry, nothing short of a hostile takeover is likely to dismember the DNA of a large organization. However, it is possible to inflict serious damage on an organization's DNA by making unwise choices about the future of core enterprise systems.
Sometimes IT managers believe that "ripping and replacing" core systems is their only option for moving toward a service-oriented architecture (SOA). This is an unfortunate misconception. If you decide to replace core systems -- by rewriting them, substituting a packaged application or outsourcing the functionality instead of preserving and modernizing the systems -- you may damage your organization's DNA.
Attempting to "clone" a system can result in an inferior organism: Sometimes organizations attempt to replace custom-built, mainframe-based core systems with a customized ERP package running on a distributed platform. The problem with this strategy is that installing a solution fundamentally unlike the original and forcing it to perform the same tasks as the original leads to an inferior organism. (Perhaps the first creature to exit a primordial pond and hobble around the beach on its fins could be said to have "walked on land," but just how well could it have walked with limbs better suited to swimming?)
Trying to rewrite the system can lead to "missing chromosomes": Attempting to rewrite a large core system (perhaps written in COBOL, PL/I, or Natural) in a different language (such as Java or .NET) can lead to a similarly unacceptable outcome as trying to replace a core system with a packaged application. We must assume that those doing the rewriting have a near-perfect understanding of the original system in order to replicate its functionality in a different language.
Here are just a few examples: the business logic must be thoroughly understood, the core system's interdependencies with other systems must be known in detail, and how end-users interact with the existing system over the course of many business cycles must be understood.
Failure in any of these areas can lead to recreation of DNA with "missing chromosomes." Just as this condition is a disaster for the organism in biology, it also has a negative impact on the business in IT. In one particular case, the situation in biology is preferable to IT. In biology, one can look at a DNA strand under a microscope and discover which chromosome is missing or defective. In IT, it is sometimes impossible to know an interdepdency or piece of functionality was omitted or misunderstood until users begin screaming. In some ways, replacement can be more challenging than starting from scratch.
Letting key retirees go can risk loss of domain knowledge representing the true "genome map": Any replacement scenario involves technology and human resources. Concern over potential skill loss through retirement of long-time employees can lead to solving the perceived problem by putting systems in place that don't require "legacy" skills. It seems like an easy fix. However, it is very likely that much of the organization's system domain knowledge (or "genome map") resides only in the minds of the experienced employees. Why? Because little or no time has been carved out of their daily activities to write system documentation.
In many cases, these long-term employees have been instrumental in developing and maintaining the core systems. They are familiar with problems that have arisen over the years and they have designed fixes and enhancements. They know the system interfaces across the business. They also know which applications are integral to business operation, which are redundant to other applications, and which are dead weight.
Imagine the prospect of letting these seasoned employees go and trying to replace the core system without the benefit of the domain knowledge they possess. Some elements can certainly be reverse engineered using analysis tools, but much of the replacement task will need to be done via tedious trial and error. Many of the organizational stakeholders will be painfully aware of the hits and misses during the long process!
Application Modernization = DNA Preservation
Whether through imperfect cloning, missing chromosomes, or loss of the genome map, poorly performed replication of the organization's DNA through replacement can seriously impact the mission of the organization and ultimately the organization's financial health. Fortunately, there is a better way: application modernization.
As the successful outcome of a comprehensive application modernization effort, three things will happen: 1) end-users will be able to interact with core systems in a modern way; 2) core systems will be able to interact with each other, and with other systems, in a holistic manner; and 3) core systems will fully support and feed into the ongoing enterprise architecture strategy of the organization.
From Green Screens to Modern Browsers
While the core system itself might still be outperforming its distributed peers by many orders of magnitude, the "green screen" user-interface is the only part of the system many business users will ever see. The first step in application modernization is to "fire the old PR agent" and provide business users with an attractive, browser-based interface to core systems that looks and performs like their favorite PC or Internet-based application. With this relatively simple transformation, the power, reliability, and security of the mainframe-based core system will be matched appropriately with a powerful and more welcoming user interface.
From Data Silos to a 360-Degree View
The existence of data silos throughout the organization is the bane of virtually every industry -- public or private. Yet the requirements of agile decision-making, data warehousing, and business intelligence gathering demand that pieces of information residing in multiple systems can be easily and quickly viewed as a cohesive whole. Therefore, to "see" all of the necessary information from a single vantage point, a single application must be able to "talk" to many kinds of database systems and ask to see the information residing there -- without actually moving the data.
From Inflexibility to Agility
Business users will soon compose their own applications out of individual pieces maintained by IT. As more business processes are modeled and automated, they will begin to incorporate these new user-composed applications as part of the overall process. All of this will facilitate unprecedented levels of responsiveness and flexibility required to compete and meet the needs of stakeholders. Yet, core systems will still be required to deliver non-stop data processing performance, 24/7 reliability, and iron-clad security -- while also supporting these new agile business requirements. How can this happen? The only way this transformation can take place is for core system data and logic to become available as services to feed the new SOA/BPM enterprise architecture. Therefore, core systems will move from inflexibility to flexibility by becoming service-enabled on top of their existing duties.
The Bottom Line
To maintain the company's unique identity and competitive differentiation, the DNA must simultaneously be preserved and evolve. Application modernization is the one approach that enables an enterprise to concurrently implement a strategy of conservation and a strategy of boldness, so it's good that application modernization is usually the most cost effective approach as well.
- - -
Joe Gentry is the chief technology officer and senior vice president of the Enterprise Transaction Systems business line at Software AG.