What is OLAP?
Online Analytics Processing (OLAP) or Multidimensional Analytics Processing (MOLAP) was originally introduced as a revolutionary way to speed up multidimensional analysis and enable businesses to query huge volumes of data. It achieves this by pre-defining dimensions within an OLAP cube – such as date, time, and other metrics used to measure data – and storing pre-processed datasets in order to reduce query time.
How does OLAP work?
OLAP kicks in when datasets get very large or complicated, for example, when JOINs, aggregation, or grouping is needed to perform a query. An OLAP programme will then pre-aggregate and pre-calculate this dataset against its set dimensions, storing it within the OLAP cube. Users can query directly into the cube to get super-fast results.
Disadvantages of OLAP:
- Inflexible analysis: Parameters for the cube need to be established upfront and pre-built, therefore limiting the ability to ask new questions on evolving datasets.
- IT-intensive: OLAP databases are difficult to maintain, requiring highly specialised resources. Any changes to the cube require a lengthy update performed by technical teams, costing time and money, as it delays or prevents insights.
- Limited size: Cubes can quickly become extremely large and complicated as more layers are added, reducing performance speed and responsiveness.
- Contrasting and Ad Hoc Groups: Some datasets or queries can’t be cubed. Contrast analysis or queries on custom groups can’t be performed using OLAP because they haven’t been predefined. For example, counting total customers and querying how many of these customers are performing certain activities.
What is ROLAP?
In contrast to OLAP, Relational Online Analytical Processing (ROLAP) works with raw datasets. It stores unprocessed data in relational database tables and runs queries directly on this data, without pre-aggregation or pre-calculations. ROLAP can handle much larger data volumes than OLAP. Users can run SQL queries and calculations against this data on-demand, allowing for ad hoc analysis, custom groupings and agile insights.
Disadvantages of ROLAP:
- Slow processing times: Processing times become much slower as datasets increase which significantly impacts the cost-benefit of this type of analysis.
- Scalability limitations: Ineffective JOINs operations between large tables restricts the capacity for ROLAP products to scale.
RELATED: For GPUs databases of today, the big challenge is doing JOINs
Brytlyt unlocks the power of ROLAP
The vast amount of data being produced today is restricting both OLAP and ROLAP capabilities. Neither technology can keep up with streaming, real-time, or rapidly expanding datasets. The difference is that OLAP’s pre-processing approach inherently bypasses the potential performance problems of datasets that are too large – giving users a costly, but acceptable, compromise.
Therefore, ROLAP has often been considered more inflexible than OLAP because of the speed and efficiency restrictions, despite the fact it enables a more agile, on-demand approach to data analysis.
Brytlyt’s GPU-accelerated database provides a solution that unlocks all the benefits of ROLAP, with none of the performance costs.
BrytlytDB, our GPU powered database solution, can process billions of rows of raw data in milliseconds to deliver a speed of thought analytics experience. It can be deployed as an end-to-end analytics platform solution, alongside our visualisation tool SpotLyt and intelligent AI tool BrytMind, or integrated as a standalone database product within existing analytics infrastructures.
BrytlytDB is a relational database that works with the responsiveness of OLAP databases. Users get the same experience as cubed processing as well as the flexibility of ROLAP. In addition, we have uniquely unlocked the ability to perform parallel processing for JOINs on GPUs, transforming the speed and cross-database functionality of data processing.
Users can now ask new questions and perform ad hoc analysis using near real-time data. This ability is particularly beneficial for sectors such as the financial services and retail, where there are a variety of different categories and data types which, when analysed together, could reveal interesting trends.
At a Glance
Pros |
Cons | |
OLAP |
|
|
ROLAP |
|
|
Get in touch to see how Brytlyt can empower more agile and fast-paced data analysis, for tomorrow’s data as well as today’s.
