| Open Source Olap |
Article Index for Open Source |
Website Links For Open Source |
Information AboutOpen Source Olap |
|
ROLAP Mondrian (Now Called Pentaho OLAP) - Is a Relational OLAP server written in Java which supports a wide range of the MDX query Language. Cubes and dimensions are defined by creating an XML schema. Mondrian is good for reporting and after a cube has been browsed will cache values so that subsequent access is faster. If Materialised views on a database such as Oracle are used then initial cube viewing will be fast. Since Mondrain essentially sits on top of any relational database it doesn't offer the ability to perform what if type analysis, this could only be done by altering the underlying data, and ensuring the cache was flushed so that modified data was displayed. In November 2005 the Mondrian project was consolidated into the Pentaho Framework, with lead developer Julian Hyde joining the pentaho team to continue the development of Mondrian under the new name Pentaho Olap Server. The server will remain open source, but will now have the backing of a company and hence the added benefit of support. Whilst some may frown at this development, it is nothing to worry about, just think about the licensing of products such as mySQL where you can use the open source product for free or pay for support if you need the assurance should anything go wrong. Indeed many companies would prefer to use products where support contracts can be arranged. Sample Screen shot showing data using the JPivot JSP Tags for Mondrian. MOLAP PALO - An open source MOLAP server whereby values are written back directly to RAM. Useful for planning, budgeting, forecasting, reporting. Palo is a 64 bit open source olap server for windows and linux. It is very easy to use in comparision with traditional tools, and many tasks such as creating dimensions and cubes can be done by using the supplied excel add in. However you shouldn't under estimate this tool due to the fact excel is used, it packs some punch and provides APIs in languages such as C, PHP, .NET & Java. The fact that Palo is considered to be a real-time (RT-OLAP) product conveys several advantages, since calculations such as consolidations are calculated when required and cached. This means that there isn't a long delay in processing cubes and you don't encounter problems such as "database explosion" which can occur with olap servers which store calculated values (indeed this doesn't really help if you are writing values back to the cubes) A screen shot of a demo application created using data pulled from the open source Palo Multi-Dimensional database. The cockpit was created using the free excel addin and uses only Palo and Excel Functions, with no bespoke VBA or any other type of code. Palo has the distinct advantage of being easy to use. Creating dimensions can be accomplished by adding elements and defining the consolidations with the click of a mouse, or by using a simple function in the front end. This ease of use compares favourably with many OLAP products where you have to create star schemas or XML definitions in the case of mondrian. Palo offers a java based plugin for the eclipse platfrom allowing dimensions and cubes to be both modelled and browsed in a platform independent and free environment. Another advantage of PALO is the ability to create an element which has multiple parents within the same dimension, a common requirement in financial modelling. Creating a similar situation with products such as Microsofts Analysis Services is not only difficult but can lead to performance issues, since analysis services can only pre-calculate certain types of hierarchies. Click here to have try a Sample web based PHP Application using Palo SEE ALSO Open Source BI for other pieces of open source BI - ETL, Reporting, Data Mining etc. |
|
|