Saturday, August 18, 2007

Centralized or Distributed?

I was trawling through some old presentations the other day - when I came across a couple that were given at crisis points in major ECM implementations. What struck me about these was the focus on the architecture of the ECM system. In particular whether it should be centralized or distributed (most of my clients have been very large and often global in nature). The pro's and con's of federated ECM - the issues of replication etc are well understood. But what I think is less understood is the impact that this architectural approach has on ECM as a practice.

By default centralizing or conversely distributing the content management impacts the way that the content is managed - for example if we centralize then we see the following common attributes:

  • Centrally controlled and organized
  • Complex and far reaching processes across the organization
  • Sequential activities
  • Explicit and definitive content/records
  • System managed by an expert
  • Performance may be slow
  • Reliance on corporate (HQ) expertise
Whereas when we distribute the content in a federated architecture we see:

  • Control at local level
  • Simple processes optimized for local needs
  • Parallel activities
  • Distributed and fragmented content/records
  • Various levels of expertise
  • Fast performance at local instances
  • Local understanding of local needs
The decision to distribute or centralize ECM though (in my experience) is almost always decided based on two criteria - technical considerations and distrust/trust.

The technical considerations are obvious really - sometimes federating just doesn't work or becomes too costly to support - likewise centralizing can impact performance etc - but the trust/distrust thing is usually dependent simply on the culture of the firm.

In the new BPM course we stress the need to undertake full As Is and model thoroughly the To Be situations. Looking back on many of my past consulting engagements it has been all too obvious that insufficient study of needs and potential solutions was undertaken, and in nearly every case way too little understanding of the current situation was grasped.

My point (if I have one) is that when making major architectural decisions - we should always be cognizant of the business and cultural aspects that these decisions will impact..........

No comments: