In discussions with customers, the question often arises as to how to proceed with a BW 7.5 and smaller. When should the migration to BW/4 be carried out or should a completely new path be taken? Therefore, we would like to compile some aspects that can support the decision.
We will look at the following aspects:
- Guidelines for new development in BW 7.5 on HANA
How should new BW projects be implemented: Is it already worth using the new objects (aDSO, Composite) or even AMDPs for the development of complex transformations? - Planning in third-party tools
Third-party tools have lost their terror. The first tools are already successfully in use today. Why are these tools not also used for planning, which must be additionally licensed in BW/4? - Still SAP-BW or a completely different tool?
The effort for the migration to BW/4 is very high. Why not switch to another tool? - Why switch to BW/4 HANA?
Guidelines for new development in BW 7.5 on HANA
As soon as your BW release has arrived at BW 7.5 on HANA, the question arises as to which new techniques should be used. At this stage, the "old" and the "new" worlds with multicubes and composite providers coexist on an equal footing. Development can take place in both worlds. Only when migrating to BW/4 does a migration of all old data models become mandatory.
Today, the functionality as well as the stability of the new data models (composite / aDSO etc.) are so mature that the question no longer arises as to which model should be developed in. All new projects should only be developed in the new data model, as the learning effort is low and the simultaneous benefit through new functionality is high. Existing data models, however, do not have to be converted, they will continue to run.
A different question is whether only AMDP processes may be used in transformations. This is desired by some customers, but not absolutely necessary. A transformation into aDSO with ABAP routines is also still executable in BW/4. However, against the background of performance, the transformation with ABAP is not pushed to HANA and therefore runs slower. Before in-house developments are used, alternative formulas or new functionalities (e.g. reading from an aDSO) could solve the problem. These are pushed to HANA and are easier to maintain than convoluted coding. To develop an AMDP routine, a new development language must be learned. This requires appropriate practice for more complex requirements. Then it has to be considered individually whether such a high volume of data is to be expected and what skills are available in the development team as well as in the support of the application.
Planning in third-party tools
In the meantime, even on a BW 7.5, the SAP world is no longer a closed universe. Many companies have third-party tools in use, such as Tableau or Power BI for a fancy reporting interface. This has lowered the inhibition threshold to integrate third-party systems into the BW world. Additional licences are required in BW/4 for the use of planning. With BW 7.5, additional licences were only necessary if planning was to be pushed to HANA on the performance side. Therefore, BW planning will have to compete against the cost of a third-party tool. When introducing external planning tools, there are a few stumbling blocks to consider that are often forgotten in a tool selection and can jeopardise the project and drive up costs shortly before completion.
Data replication for actual and planned figures in and out of the external planning tool
In contrast to a reporting tool, which "only" has read access to the BW data, a planning tool needs its own data storage for the plan figures. In most cases, the actual figures are also stored there. This means that data replication is necessary for the actual figures from BW and often back into BW. This results in another reporting interface. Alternatively, the planned figures are not available until the planning has been completed and replicated again. It should not be forgotten that extensions to the data model also always have to be polished by another system.
Understanding the data model
In addition to the pure physical replication of the data, the data model must also be adapted to the new tool in terms of content and structure. Each tool has slightly different methods and data to store. This is rarely a real problem, but it can easily lead to misunderstandings that delay development.
Authorisation and user matching
User maintenance must also be stored in an additional tool. Often the effort is not the actual user creation, but the updating of responsibilities.
BW or a completely new tool
There are many rumours and fears about the effort required for migration. Many fear a very high effort, similar to a new implementation. This is a reason to question the use of SAP BW and to evaluate other tools.
How high is the migration effort to BW/4 really?
As always, the effort depends on one's own situation. If you have been migrating to BW 7.5 on HANA for some time and have already realised projects with the new tools, there is no longer any migration effort.
The migration tools are now so sophisticated that many models can be migrated automatically. If a system has a lot of historical legacy, the wealth of experience built into the data models should not be underestimated. On the other hand, this can also be an opportunity to "clean up" and make the system leaner through new techniques.
Governance through BW not so present in other systems
In another tool, one or the other solution can certainly be built quickly. However, it should be taken into account that in a BW, the tools for development as well as transports, data loading and control are well established and comfortable. This is not possible in such a lean way in all systems.
Naming conventions and authorisations have also been established for the BW systems in order to distinguish different areas from each other. This comprehensive knowledge has to be set up again in new systems.
What are the reasons for switching to BW/4?
Before an upgrade, therefore, a clear decision should be made in favour of the upgrade before considering different points of view. How does this look with BW/4? Since BW 7.5 is already a quite mature product, there are hardly any general and important new functions. In detail, there are some improvements. Those who work with large amounts of data and sometimes remodel their data model will find some aspects useful, such as data tiering to move older data out of the main memory.
In planning, you can now also plan master data and enter new master data in a planning layout. This enables some new planning scenarios. However, it is mandatory to purchase additional licences if you want to use planning.
Conclusion
It is undisputed that the migration to BW/4 is a big leap that raises many questions. Whether the new functions require a faster upgrade must be weighed up. By 2030 at the latest, the end of the extendend maintenance of BW 7.5, this decision should be made. Especially against the background of the SAP Data Warehouse Cloud on the horizon and the general cloud strategy of SAP...
Hopefully, the blog has provided a few ideas on some of these questions around this complex of topics.
If you have any further questions or need help, please feel free to contact us.
Contact Person
Dr. Ulrich Meseth
Senior Consultant
