I was recently asked this question, so I thought it might be nice to share a bit of history. This is a very short entry, explaining how DV got it’s name!
In the beginning, I was working for the US Federal Government. My customers included the Department of Defense, National Security Agency, and NASA to name a few. It was the early 90’s and they had a need to build a data warehouse. However, this warehouse was not to be like any other.
You see, most data warehouses at that time were being built in either 3rd normal form, or dimensional models, but that’s not the crux of it. They were – for all intensive purposes, not data warehouses, but instead, information warehouses. In reality, they loaded all the data post business rule execution (ie: post-munging).
This, in turn, made it extremely difficult for anyone to audit the so-called data warehouse. Not to mention, the business users were arguing over business definitions, metadata, and the computations / rules that were applied to the data being “aggregated for the common good / enterprise view”. Well, we all know how that turned out… caused lots of pain and so on, but I digress.
I was asked by the DoD and NSA to construct an auditable historical data store, they wanted a common unified modeling practice (at the time they were training us on Six Sigma, CMMI, PMP, and of course RUP (rational unified process). To have an “auditable data warehouse” was unheard of, people couldn’t or wouldn’t fathom the notion of storing raw data in an integrated fashion. I fought many of the same battles I still see around today (like: what’s the value of the DV instead / over/above a straight PSA?)
Anyhow, that’s beside the point. In order to change the culture, and create a unified modeling view, I had to construct something that would allow the NSA to extend / expand the model based on the core design behind closed doors (in the classified areas), and to inherit all the changes as I went along without disrupting ANY of their additional components that they’ve modified or extended. Yes, near-zero impact was a design challenge right from the start!
Ok, long story short: I started by calling it “Common Foundational Warehouse Modeling Architecture” – CFWMA. It had all the characteristics, met the needs (more stories about how and why), and so on. But CFWMA is not an easily pronounceable acronym (try it, but don’t hurt yourself!).
When I took another look at “what this thing accomplished”, it in fact, did act like a Bank Vault – keeping data safe & secure in the classified world, and in the commercial space, allowing the data (as needed by log in), to be accessible. An additional requirement I had right from the start was to provide for CELL level protection in a relational database space, which doesn’t sound so hard today, but back then – it was tough.
Anyhow, we (my team and I) managed to construct this “bank vault like looking thing” called a “data warehouse”, so I was joking one day with my classified sponsor, and he said: you should call it a “data vault” – it made sense to me, so thus, it was named.
Before September 11th, 2001, I was the only one utilizing the name on the web – but Oracle soon got ahold of it, and after 9-11 they decided it was a cool marketing name for security products. The rest they say, is history.
Now, is the “Data Vault” a place to lock up the data and throw away the key? No, but it does offer secure protection on components of the model that need to be applied. Hope you enjoyed this trip down memory lane.