Data Vault is not just Theory

it’s unfortunate to hear, but some folks still believe that the data vault model and methodology are just theory.  i’m here to state that the data vault is not theory but really: applied science.   it’s backed by a large number of best practices, solid foundations, and mathematical principles. 

it’s been implemented all over the world at more than 350 customer and government sites, and to my knowledge, it has worked every time.  sure, there are standards, sure there are rules, and yes, these standards and rules can sometimes be hopeful guidelines.  the difference between the dv being theory versus applied science is:  if you break the rules, you can measure the impact and weigh the risks.

with theory, all you get (mostly) is supposition of what might happen, but you are truly unclear.

during the design and architectural construction between 1990 and 1997 i spent many years breaking the rules on purpose, and conduction trial experiments that showed me what happened when the rules were broken.  i ran them as scientific experiments meaning i ran a control, and a non-control for the setup. 

i then tested for a number of possible scenarios including (but not limited to):

  • volume
  • latency (of arrival)
  • key generation
  • maintenance of history
  • attachment of child entities (foreign keys)
  • indexing (compound and individual sets)
  • partitioning
  • process parallelism
  • and many more…

then from the project side i would test possible scenarios including (but not limited to):

  • re-engineering (based on change)
  • downstream impact analysis of up-stream changes
  • ease of design of implementation
  • maintenance costs
  • level of knowledge needed to build new systems
  • complexity ratings
  • automatability
  • repeatability (if it was a success, can we build it again the same way)

i then proceeded to build system after system using these rules and standards.

if you call this a theory, or insist on dividing the model from the methodology, you will be doing the industry an injustice.

if you feel differently, i’d love to hear from you – to comment, please register (for free) below and comment on what you see in the market place.

thank-you kindly,
dan l
ps: check out my coaching area for more answers.

Tags: , , , ,

2 Responses to “Data Vault is not just Theory”

  1. Raphael 2010/12/04 at 5:36 pm #

    Totally agree Dan.
    For those folks out there I would like to assure you that Data Vault works beautifully . I have built DVs and currently working on another one
    DV is not a ‘silver bullet’ but a robust concept applicable to majority of EDW in all industries please contact me if you need referenses.
    Dan’s coaching is the great way to get ahead


  1. Tweets that mention Data Vault is not just Theory -- - 2010/12/03

    […] This post was mentioned on Twitter by Matt Casters, gillespol, Fred Kwakkel, Control Freaks B.V., Daniel Linstedt and others. Daniel Linstedt said: Data Vault is not just Theory: A short set of reasoning as to why the DV is more than just… […]

Leave a Reply