Silo-Busting – A Primer
Getting off the Blocks
Many companies are unsure when to initiate their silo-busting task. Here’s a primer.
When it comes to silo-busting, the easy part is making the decision to break down silos. Then comes the difficult part: getting started. Companies may have heard of some tools to help, or they may have used some in the past, but these challenges cannot be addressed solely based on a previously used point integration solution. Some companies think that they just need a single tool (either an EAI or ETL product), or they think they just need to write a few custom programs in a simple, well-known language.
Silo-busting is really a much more complex project than just moving data from one place to another. Breaking down the silos is an enterprise-wide initiative even though it often starts off as a series of very small projects.
Initially, a company should look at the big picture and develop a long-term plan that includes an inventory of the current applications, processes, and data; a vision of the end result; and a step-by-step architecture and methodology to get to a single view of their data and streamlined operations.
The GDI Methodology provides an overview of an approach to integrate the data, the processes, the applications, and the projects across an enterprise, and becomes a model for the reuse of business objects.
Beginning with a single project, the approach begins to define the data and the processes in a way that looks at not only the specific needs of a business unit, but also examines the way the data is defined and used in the rest of the organization. With an enterprise view of the organization, each data element builds on other data elements and each process feeds into another process, project by project, to provide a complete view of the enterprise, the interaction of applications and the sharing of data structures among applications. For each subsequent solution, the process layer manages the workflow across the entire enterprise and the data is held as an integrated knowledge base for the enterprise; this knowledge base is the core of the enterprise integration.
GDI Methodology
In each of these initiatives, getting the master data harmonized is one of the key first steps. This activity is the cornerstone to success of each subsequent project and allows each future activity (whether it is driving business intelligence initiatives or putting a common look and feel across applications) to build upon a harmonized and rationalized enterprise landscape and to reuse everything learned in previous projects.
In a nutshell, here are the steps involved:
Identifying the information requirements for current operations and for the analysis of the organization’s data (identifying silos)
Creating a systems portfolio
Developing an approach to retrieve information from silos
Establishing standards (naming, common definitions, format)
Building an information model of an organization's information set containing relationships and rules that represent the semantics of the data and its interaction with other data and processes
Identifying gaps, differences between applications that bridge the silos
Creating an architecture that identifies which data needs to be converted, interfaced, integrated, or consolidated
Determining which technologies are best suited to integrate or consolidate the data and processes
Determining project priorities based on critical needs, high return, similarity of processes and data, level of effort and availability of resources
Creating a long-term implementation plan for enterprise information