pombredanne / nolap

NoSQL Online Analytical Processing

NoSQL Online Analytical Processing (NoLAP)

NoLAP is a new flavor of OLAP that extends the OLAP paradigm to benefit from NoSQL' s gain in flexibility.

Here are the top 3 reasons for using NoLAP.

Hypercubes on the Fly

With OLAP, you have to define the shape of your hypercubes in advance. Once the data is in there, you are stuck to that schema.

With NoLAP, while you may have some hypercubes in advance and store them in the database, you do not have to. This is consistent with the NoSQL paradigm where you are not obliged to have schemas upfront or at all. At runtime, you can bring your own hypercube (or even a complete report schema) and retrieve the facts on the fly — all in real time.

Break the Curse of Dimensionality

The curse of dimensionality is often used as a blanket excuse for not dealing with high-dimensional data. For example, MOLAP, and also ROLAP because the number of tables explodes, do not scale up well with the number of dimensions.

With NoLAP, you can efficiently work with such data. It does not matter how many dimensions you have. The data — OLAP measures — is organized in one single facts collection, with optionally one metadata collection that contains dimensional information. Because of NoSQL's heterogeneity, there is virtually no performance penalty. Adding more facts, more dimensions, more hypercubes is as trivial as an insert.

Interoperability between Data Silos

There is no global standard for OLAP. This makes it hard to exchange highly dimensional data between "silos".

NoLAP hypercubes follow the rules specified by the eXtensible Business Reporting Language (XBRL). This allows for exchanging data using XBRL between systems easily.

Links:

• NoLAP in detail: https://github.com/28msec/nolap/blob/master/NOLAP.md
• TPC-H an example implementation of NoLAP: https://github.com/28msec/nolap/blob/master/TPC-H.md
• The biggest NoLAP database, yet: http://secxbrl.info