[ad_1]

A few years ago, IKEA released a VERY hilarious commercial to promote their line of simple and clean furniture to organize a home or office space. In short, the commercial began with an attractive couple in a rather unattractive and untidy room sharing a single spaghetti dinner. They savored and slurped each noodle as energizingly as one can eat a bowl of noodles. The woman – eyes closed, head on plate, mesmerizing her partner with her technique – was so absorbed in her sensual performance that she accidentally mistook her boyfriend’s dirty shoelaces, which were of course tied to his even dirtier sneakers, for another noodle . (If that’s too gross for you, that might help explain why it was banned in the US.) The commercial faded out, ending with a silent but poignant reminder: “Tidy Up.”

Just like a messy room, a spaghetti (or star) integration can take months or even years to get messy. Traditionally, each department in an organization can use independent systems to meet their own needs, making them autonomous and easier to manage internally. Then, as new requirements arise, an organization grows, and cross-departmental communication becomes more important, point-to-point interfaces evolve to connect each autonomous system together—thus creating a poorly planned “spaghetti” of systems.

So what’s wrong with that you say? Here are my top 3 cons:

1. Just as a messy room is a sign of a messy mind, a disorganized system is a sign of a disjointed IT management team. Choosing the most expedient, quick-and-dirtiest way to integrate is a sign of little collective systems management effort. This lack of foresight leads to complications and problems in the long run… see #2.

2. Changes to a single application or database in a star integration can be costly. Sure, it’s easier to make rapid changes in a single system, giving the organization the appearance of agility. However, if these changes result in a collection of features that are typically not reusable and buggy, the cost of maintaining those new features increases exponentially as your system grows.

3. Business intelligence suffers as enterprise data residing in multiple locations in a spaghetti integration becomes more difficult to consolidate into a single and useful data mart. This makes it more difficult to paint a clear picture of the organization as a whole, be it analytical or operational, and not to be reactionary and inflexible in the long run. But spaghetti integration isn’t all bad. New application functions can be implemented quickly, and since the structure of the system components is not as rigid, there is the benefit of deployment flexibility.

[ad_2]