November 2007
Paul W. Lail, CFPIM, finance manager, capital & initiatives, Verso Paper
In our society, the thirst for information is ever increasing. Along with this thirst, there is a desire for the most current information possible. Every day, I see evidence of these trends in my children's requests to use the Internet. Even though we see the row of encyclopedias, somehow information does not seem as current, relevant or exciting if it has been sitting on a shelf for a few years.
These same trends are also played out daily in business environments and mills. Information for decision-making is requested on an increasingly broad and timely basis, and the impact on business systems is seen in their expanding scope, as well as in the increased need for data sharing between applications and systems.
Integration shapes IT strategy
Systems integration, or the bringing together of disparate systems so that they function together in a seamless fashion, is a big (and growing) business.
A generation ago, IT groups in our industry spent much effort on applications development. As the marketplace and applications have matured, more and more time is now spent on systems integration tasks. An organization's approach towards systems integration work is a major influence in shaping IT strategy and is central to the organization's position on the "enterprise vs best-of-breed" systems debate.
With systems integration tasks becoming progressively more critical to the IT community, sharing significant experiences from such efforts becomes important. Many company IT groups now operate with the "buy bricks, make mortar" approach. Even when IT groups support internally-developed legacy applications, interfacing these legacy systems with other system environments, along with maintaining the interfaces, consumes a growing portion of IT resources.
I went back through past systems integration project notes to assemble the following tips that may prove useful when making integration decisions:
- Don't assume having just one vendor means a well-integrated system. Sometimes a vendor's offering is composed of separate applications cobbled together from past design/development efforts or companies. Inconsistent terminology between different areas of the system is one symptom of this.
- Plan in milestones. System interfaces should be referenced in project milestones. Not doing so risks overlooking an area that puts many overall project dates in jeopardy.
- Test, test, test. It seems obvious, but skimping in this area is a common reason for failure.
- Consider the value added. Integration for integration's sake is not necessarily a good thing. Integration must support a business demand or driver. By looking at the business reasons for integration, a better decision can be made.
- Design for reusability. Before finalizing the design on any system interface, determine if this will need to be done again. Some relatively simple changes in design can allow repeatability and save much rework later.
- Consider data timeliness. How current must the transferred data be? Timeliness of the data must be considered in the interface design, otherwise, the interface may be over or under engineered.
- Monitor interfaces carefully. Design the correct level of monitoring into new interfaces so that if things break they can be easily fixed. If it is time critical, uptime is a primary concern. When monitoring capabilities are built into the design, notification and resolution are enhanced.
- Verify compatibility. Systems and applications may be advertised by vendors as compatible with each other or adhering to a standard interface. Such promises should be verified from references, the marketplace and/or testing.
- Use diagnostic tools. When failure occurs, the ability to diagnose the cause is critical. The more critical and time-sensitive the interface or integration, the more time you should invest in making sure that diagnostic checks can be done quickly and simply.
- Plan for change. It is unlikely that design efforts will yield all future requirements, so the degree to which changes can be incorporated later will pay dividends.
As I assembled these tips, it struck me how the list represents a microcosm of successful project management techniques. The need to plan, develop thorough requirements, understand the needs and capabilities of the integration environment and manage execution of the integration effort parallels what is required to achieve success in managing projects in general.
Pulp & Paper magazine is FREE to qualified subscribers. Click here to find out more.
