Informix Datablades
Datablades are truly amazing!
Datablades do the same to relational database management systems that performance-enhancing drugs do to international athletes.
They break all the rules.
Up until now the kind of applications that use this technology include:
- police systems
- facial recognition systems
- voice recognition systems
- medical scanning applications
- web-based geospatial applications
- time-series applications in banking and telecommunications
The most important thing that datablades do is to allow you to change the rules under which relational data is stored and indexed.
In doing this, datablades can reduce the amount of space required to store data by more than 50% and make data loading and retrieval more than 20 times faster.
What's the principle behind Datablade technology?
Until the 1980's the computing industry had concentrated its efforts on storing and manipulating alphanumeric data.
During the 1980's relational database management systems emerged and, although these systems introduced a layer of abstraction that enabled businesses to model and access their data more efficiently, they were still focused on storing and indexing alphanumeric data.
The problem with this is that if you create an index in a relational database, you create a binary tree index. This assumes that all data can be sorted by alphanumeric value — for example, 2 is always greater than 1 and 'w' is always greater than 't'. It's so easy, so powerful, and yet so useless for many complex applications.
As we've already said, the most important thing that datablades do is to allow you to change the rules under which certain types of data are indexed. What this requires is that the database understands the type of data being stored and the significance of that data. The database management system must have intelligence.
A good example is a coordinate on a grid. A coordinate is made up of two numeric values that are individually meaningless. This means that you cannot say that (10,2) is greater than (5,16) — a binary tree doesn't work! This is where Datablades come in. They allow you to ask questions such as “Is this point within this area?” or “Which facial image in my database is closest to this one?”. We could not do this using a standard relational database management system; however Informix supports this capability today.
No other database technology except Informix supports this as a core feature, and it's been available since the dark ages.
The official datablade documentation is here.
Oninit® have been developing datablades, bladelets and UDRs since the technology was introduced.