And then it suddenly became clear to me – for the longest time people have been starting any serious database work with design, design of a schema to be precise. And there was nothing wrong with it – storage was sparse and super expensive, so people had to make sure that no byte was wasted, requiring that the storage format was super efficient. Down to the (bit (for the older ones of us) and the) byte.
Today storage is cheap, no matter if people want to store information on spinning rust (a.k.a. HDD) or SSD or RAM chips. And with that – for most of the new database types we see – the design component has become irrelevant, even stronger – is not even part of many of the new databases we see these days.
So could be the common thread of the new database boom the absence of a design component, the disposition of schema design step that was and is key for the success of any relational database? Let’s take a look:
- Hadoop & friends – A classic example for no design. Just store, cheaply mostly, don’t worry for what purpose and query, just store. No design and thought needed to enter data into a Hadoop database (well we may even say cluster). Check.
- Columnar (often in memory) databases – The whole beauty of columnar database is that you don’t have to worry about about schema. Just feed the name value pairs at storage time. Check.
- Purpose built databases – These are databases built for a specific purpose and they are very good at that. MongoDB as an example comes to mind. Again no design needed, the design has been done before you get the product, and it’s very good at what the database being designed for. Check.
So could there be something the whole ‘No Design Database’ observation? It certainly appeals to people to be able to do that – otherwise we would not witness the success of these new databases. And to use one of the most developed offerings of the relational database world – the data warehouse – we all know how painful and critical the design phase is. Very painful for the business people, who need to imagine all the questions they want the data warehouse to be able to answer – before they can use it. So being able to just store all that ‘stuff’ without getting tons of questions by the technical people, is of huge value to the .. business people.
Is there something to the era of the No Design DataBase? The NDDB? Please comment.