Saturday, February 6, 2016

3.2 Think big: requirements for a Corporate Data Pool

Now we are at a point at which we are committed to consolidate all our data silos into one big database. Let’s assume that we have top management commitment for the necessary organizational changes so let’s not worry about this point for the moment. For easier reading I will from now on call our solution “Corporate Data Pool”.

Corporate Data Pool = central single protected storage of all enterprise data, accessible to everyone in the company

If we made a list of requirements this is what our new Corporate Data Pool should offer – all of these are real, existing customer requirements.
  1. High data quality - current and correct data, no duplicates
  2. High performance - lightning-fast response times (well below one second in any case)
  3. High capacity - capable of handling the expected data growth
  4. Protection of investments - existing applications continue to run with the Corporate Data Pool
  5. Flexibility - suitable for different requirements from existing and future business processes (transactions and analytics)
  6. Lower TCO - much more cost-effective than today's infrastructure (as we are shutting down many silos the cost of operations should drop by at least 20%)
  7. Security - safe and protectable (not everyone is allowed to see everything)
  8. Enterprise-readiness - capable to handle mission-critical applications, scalable
  9. Future-proofness - proven technology from a strong vendor
  10. Business continuity - transition without interruption of productive use of current systems
As a consequence of the above mentioned requirements we would further expect
  • Elimination of all interfaces between data silos and all related data provisioning processes (ETL, batch jobs, nightly data loads, etc.)
  • Smaller data footprint (as we eliminate all duplicates)
  • Less hardware (as we are shutting down silos and as we expect as smaller data footprint overall)
  • Development should be more efficient as we do not need to worry about performance and origin of data anymore
  • A consistent backup & restore should be very easy to do
Reminds you of the wish-list of your kids for Christmas?!

Here comes Albert Einstein again: “We cannot solve a problem with the same thinking we used when we created it”. In other words: the current client-server architecture with relational databases is part of the problem, not of the solution.

But what is the alternative?

No comments:

Post a Comment