Skip to content

Day 1: 10:06am – The First Fork – Tabular vs Multidimensional

November 3, 2012

The first fork in the road seems to be picking an OLAP model – Tabular or Multidimensional.

I’m going with Multidimensional.

As I stated in the outset, it’s not my intention to be thorough in my research and decision-making this week. So, this Microsoft White Paper, and Chapter 1 from my Wrox book have given me enough input to choose the Multidimensional model.

Loosely, the factors in my decision are:

The Multidimensional model has been around a lot longer (5 releases of SQL Server), and is described as the ‘Traditional’ model. Presumably this means it is more stable and mature. I’m also hoping there are more resources out there for me to learn it.

The Tabular model is intriguing to me mostly for its newness. The in-memory xVelocity (aka Vertipaq) engine sound like lots of fun, and I know that I’m ignoring an opportunity to build a quicker, simpler solution using the Tabular model, but I can’t help but feel like it is somehow a shortcut that will bite me later on. It also seems to involve a lot of Excel. With all due respect to Excel and its legion of followers, I didn’t come all the way to PASS in Seattle to play in Excel all day. :)

Like I said, I’m making snap decisions to get this built, so there we are – Multidimensional it is.

The 6am breakfast of coffee and Pringles has run their course, so my next decision will be where to get some food.

About these ads

From → Uncategorized

6 Comments
  1. Simon. I really like your idea about going to PASS and tell the story about your intro into the CUBE-thing. I have been round BI for some decades and just want to warn you, the cube is melting ( http://ricol.se/en/the-melting-cube/ ). The authors of “Choosing a Tabular…” are all so involved in the traditional BI-business earning money on the cube, and as many beeing almost blind for changes. The paper should be named “Why MDX is superior to anything else”. For most common BI-solution you will have a toolbox good enough using a tabualr view and you will have a model adapting directly to most changes in the real world. This is the most important aspect partly covered on page 19-22.
    Rickard

    • Thanks for your comments, Rickard. I read your post (“The Melting Cube”), and agree with your sentiment that the cost of hardware resources such as SSDs, RAM, and the availability of in-memory technologies (Vertipaq, ColumnStore indexes, etc) have seemingly begun to change the equations used to determine the need for enterprise-level BI solutions. As I think I mentioned in one of my posts, the Tabular model does seem to have quite a lot of momentum, and maybe it will become the de facto solution for most businesses small and mid-sized businesses. I’m too nascent in my journey with SSAS (and BI in general) to have a more articulate position, but I do know that the need for decision-support systems that SSAS can provide will only be growing as the availability of granular (OLTP-based) data continues to grow.

  2. Michael permalink

    I believe Tabular requires a higher and more expensive SQL Server edition and so will limit its take up by small business, maybe even some meds. too.

Trackbacks & Pingbacks

  1. Day 4: 9:55pm: The Halfway Point « OLAP Sprint
  2. Day 4: 9:55pm: The Halfway Point // simon doubt

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 40 other followers