Federation and Integration are two words that all Information Professionals need to become more familiar with.  We are at a crossroads in the computer age and these two terms will be integral to your navigation for some time.  At issue is exactly which fundamental paradigm will be followed for the system design of enterprise networks.

Federation. In networking terms, a federated network is one that is comprised of many smaller, independent networks that only accomplishes its overall objective by effectively employing its sub networks efficiently.  Essentially it describes  a network of networks with each network being capable of independent action.  Critical success factors include being able to specify, monitor, and engineer the interfaces and business rules.  The advantages of this approach are redundancy, optimization of individual functionality, simplicity, and usually design cost.  Its disadvantages are footprint size, increased management functions, and trouble shooting complexity.  Your home is probably a good metaphor for this.  It is currently filled with all sorts of electronics and computers that are all very good at their one function — taken together they form a federated product to produce your entertainment, productivity, and well being.

LAN

Integration. The opposite of the above concept, integration is when sub systems have no independent capability in the absence of the whole.  It is a singular brain — truly a central processing unit — driving a spider web of functions. The advantages to this include lower overall number of components, easier problem identification, and energy efficiency.  Disadvantages include sub optimization,  performance compromise, upgrade difficulty, and design complexity.  The famous single ships computer on Star Trek exemplifies this — a single CPU and memory store that ran everything from the navigation systems to the ships controls, weapons, and even plumbing.

This matters because, in my opinion, most teams designing large enterprises today tend to overestimate the advantages to their selection — integration or federation — and downplay its disadvantages. What this does is invalidate their risk assessments, contingency plans, and most importantly their sustainability and maintenance budget estimates.  They fail to plan for the required level of technicians as well.  In general, I would say that there is a tendency towards overreaching for integrated solutions vice federated ones for no other reasons that initial cost drivers and a misplaced desire to appear technologically forward leaning.  Overall, I believe that sound engineering and a realistic evaluation of risks is the best practice.

How do you see this issue?  Is one old school and one new?  What factors lead you to one paradigm or another?  Happy Thinking.

That is my Information Technology Thought of the Day (ITTOD) for April 17, 2009 ©Scott Coughlin .

Be Sociable, Share!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.