Let's talk about complexity in TM1


Big Data is a big buzzword at the moment, so I thought it would be interesting to explore how "big" data can really get. What happens if we examine a TM1 system, and calculate the number of values that could be theoretically stored in it?

From experience, we know that most cells in a cube will not contain a value, or in technical terms, cubes tend to be sparse, but let's ignore that for a moment.

What would be your guess on where the biggest customer in our Pulse Benchmark Database ended up ? As it turns out, the biggest system in our benchmark database was submitted by a Cubewise customer in Sydney, who has a cube in their TM1 model that can store a massive number of values, specifically: 



(Note that I had to insert a line break so the number could fit on the page!)

This is such a big value I didn't know how to read it out loud, so I asked Wolfram Alpha what it was called. The official English name of this number is:

23 septenvigintillion 623 sexvigintillion 234 quinvigintillion 580 quattuorvigintillion 973 trevigintillion

or, if your background is more scientific you might write it as:

2.36 x 10^85

This certainly seems big, but a physical comparison will give you some perspective on how really enormous it is. This cube has more cells than the visible atoms in our universe, which is just around 10^80.

In everyday business, astronomical numbers like these are not very important to an end user or developer, but this experiment demonstrates a few things:

  1. TM1 can model very complex applications that are not possible with other technologies
  2. Even a system of this size retains the ability to perform real-time calculations and modeling while effectively serving many areas of the business 
  3. Pulse for TM1 helps you document, manage and support an environment like this, and enables you to scale TM1 applications to arbitrarily large sizes and complexity

Stay tuned for more gems from the Pulse Customer Benchmark database!