On Information

Concepts in Information Management – blog by Ronald Fuller

A False Hope for Big Data


Big data can provide powerful insights into large data sets. Some experts and scholars have put forward the idea that big data tools and techniques can even be used to replace relational databases for ordinary business use, but this claim offers only false hope for organizations that struggle with relational databases. Here's why:

 datacenter3

Big data systems work with information organized into small 2-part chunks known as key-value pairs (or some related format). For example: Last Name = Fuller; City = Redmond; Car = Honda Accord; Order status = complete. 

 

Organizing information this way is great for things like analyzing trends and detecting patterns. But big data formats cannot be used for ordinary business reporting unless each record is tagged with additional information to tell which other records it is related to. For example:  this address belongs to that person; this item goes with that order, and so forth. Applying these kinds of tags to information in a big data format requires exactly the same kind of discipline and pre-planning as it would if it were organized for a relational database. Big data offers nothing new in this regard.

 

Even when a big data record set includes complete information about the relationships between each pair, big data technologies do not offer anywhere near the flexibility of relational databases for reporting purposes. So any claim that big data presents a plausible alternative to relational databases for general business use is at best misleading and at worse completely false.

 

Add comment

Comments

Subscribe:
  • No comments found