Slow to generate, under-utilised and inaccurate Business Intelligence (BI) data puts organisations at a competitive disadvantage. In this paper, Daneel de Villiers, managing director and co-founder of De Villiers Walton, explains how the SAP consulting specialist’s back to basics Performance Review for SAP BI identifies the causes of SAP Business Warehouse difficulties and allows organisations to generate better Business Intelligence.
A Performance Review for BI is a low-risk investigation that is typically conducted in response to a performance problem which can arise both at the front end and back end of SAP systems. Front end performance problems usually relate to speed; a user will run a search, or ‘query,’ but must wait a long time for the results to be retrieved. Back-end performance relates to performance issues manifesting during the Extraction, Transformation and Loading (ETL) process.
Factors which may affect front-end performance include issues such as network speed or the wrong version of Microsoft Office tools being used in combination with the SAP Business Explorer (BEx) tools. Additionally, if the user is selecting large volumes of data or the database process for selecting data from the underlying tables is not optimised, this will also affect performance.
Back end performance problems, meanwhile, can also be identified through poor speed of extraction, transformation and loading, for instance when data is uploaded from a source system and its takes an excessive amount of time to complete.
As a first step, we always instigate a client consultation to find out what problems they are experiencing and what they believe may be the root cause. Once we know the ‘symptoms’ we can establish whether we should start at the back or front end.
Front end performance issues
We look at the whole range of factors to determine how to make the front end performance faster and identify the bottle neck preventing faster reporting. Broadly, this process is divided into three layers – the Presentation Layer, Processing Layer and Data Processing. For each of the processing steps, execution and processing times are captured and saved into underlying database tables. By analysing the results in these tables, it is possible to identify where the main problem areas might be.
For performance-related issues in the Presentation Layer, we examine factors such as software compatibility, network bandwidth and the size of the data volume stream to the desktop. The problem could even be as simple as users not having enough resources (memory and processing power) available on their PCs to run their search or ‘query’.
In SAP Business Warehouse, the Business Explorer (BEx) Query Designer is used to define a query through the organisation of fields and values in an InfoCube into a data structure containing rows and columns. Problems within the Processing Layer may arise if queries are poorly constructed. By reviewing the configuration of individual queries, we are able to identify and resolve performance issues at the processing layer.
Within the Data Processing layer, we typically look at the design of the components that provide the data (such as InfoCubes) to check that it is possible to extract it quickly and efficiently. A poorly designed and inefficient InfoCube, for example, will result in long load times when updating data and adding records, a high amount of disk space being consumed, and slow and inefficient data retrieval.
Typically, it is a combination of factors which cause front-end performance problems rather than just one single issue. To properly determine what the causes of the problems are, the Presentation Layer, Processing Layer and Data Processing must each be investigated separately.
Proportionately, correcting an issue in one area may yield a greater performance enhancement than in another, but it is not possible to completely optimise performance without identifying and amending all issues. We therefore identify what we believe are the highest priority areas and begin the process of elimination with these.
Back end performance
The main benefit of improving back end processes is in creating the ability to reliably deliver information to the right people at consistent times – i.e. to make the reporting process both accurate and consistently repeatable.
Like all enterprise software vendors, SAP has different versions of its product. SAP Business Warehouse has gone from version 1.x to the current 7.x. The latest version has been completely redesigned with performance in mind, and contains new objects that are far more efficient, making an upgrade extremely worthwhile. Organisations using Version 3.x objects are processing data using old, inefficient code that could be compared to an old and incomplete tool set.
In BW 7.x, SAP has added a range of new tools and objects for extracting and transforming data from one format to another, and to load or stage it in different layers within the system. Fortunately, SAP BW is backwards compatible which means it has the functionality to incorporate data that has been generated in an older version of SAP BI. Converting the version 3.x objects to version 7.x will immediately result in extraction, transformation and loading (ETL) performance and efficiency gains.
Better data management strategies and reducing the amount of unnecessary data which may be clogging up the system may also yield improvement.
One version of the truth
If an organisation does not have the ability to access consistently reliable data for reporting, users will often return to the source system and extract the data manually into tools such as Microsoft Excel or Access.
In addition to productivity losses, rather than analysing the data, these users can be so focused with producing the results that they have no time to analyse them and make informed decisions, which in turn seriously undermines the ability for everyone within the organisation to work from a single, accurate set of results.
Data warehouses provide a single validated set of data that is always up-to-date and that can be accessed by all users at any point in time through a standard single set of validated reports and have inevitably significantly reduced the risk of inaccurate data.
Before the introduction of data warehouses, many organisations used different versions of data based on a variety of manual processing methods for reporting. A significant amount of management time was taken-up reconciling the differences in the data in order to establish the true position, which delayed the speed at which informed management decisions could be made.
In addition to this process being time consuming, incorrect management information could also lead to incorrect and costly management decisions.
There are therefore obvious benefits of upgrading SAP Business Warehouse and enhancing the existing SAP Business Intelligence. Doing so can enable organisations to generate accurate, credible data and save costs. Most crucially, the ability to access accurate, consolidated reporting also means that business decisions can be made based on the clearest and most truthful reflection of current business performance.