Upgrade of SAP BW® and migration to SAP HANA®
Publicly owned energy trading company. The company has about 250.000 customers, both industrial and residential.
In 2014, the company defined a strategic initiative to modernize infrastructure and operations, and address several existing risks:
• System risks - existing SAP BW 3.5 system was nearing end-of-life support
• Business risks - upcoming changes to the market meant agility was needed
• Supplier risks - finding key competencies needed to become easier
• Personnel risks - needed to prevent staff churn and facilitate recruiting
• Strategic risks - future business cases had to be supported
In addition, high licensing costs of the existing RDBMS platform was affecting company profitability. A strategic decision was taken to upgrade and move to the SAP BW on HANA platform.
Upgrade and migration from SAP BW 3.5 to SAP BW 7.3 on SAP HANA.
Over a period of 20 weeks all SAP BW systems - Sandbox, DEV, QA, and PROD - were upgraded from SAP BW 3.5 to SAP BW 7.3, and migrated from the existing RDBMS to run on SAP HANA.
The project was split into three phases: 2 weeks preparation, 16 weeks of SAP BW upgrade and HANA migration, and 2 weeks of post go-live support).
Different partners were involved in installing and configuring the hardware, running ongoing application maintenance, providing operations support and performing the upgrade and migration work.
• Dual development environments had to be managed during the project.
• Backup and storage costs went up with dual systems, before going down after decommissioning of the legacy system.
• SAP HANA appliance versions were delivered with slightly different configurations - since appliances were purchased 4 months apart, they contained different HANA DB versions. An upgrade to the SBX/DEV/QA appliance was necessary to match with PROD appliance.
• Many vendors were involved - HW, SW, maintenance, and migration partners. There were many communication paths and different reporting structures.
• Make good use of your preparation time - secure that connections and user accounts are set up and tested for all parties involved. Run ABAP Test Cockpit (ATC) to find custom code issues.
• Use strong advisors - HANA is a moving target. Use advisors with strong technical skills and who are well connected with SAP. They can get the word from SAP CoE organisations regarding global experiences from customers who have made the journey before you.
• Utilize your SAP support agreement: Technical Quality Checks are available from SAP, and are free of charge for many customers with support agreements. They are executed remotely and require some planning. Take advantage of them and allocate time to analyse and address their findings in your project plan.
• Plan properly and in great detail: Plan for dual maintenance, and communicate system deadlines. Plan for the decommission of existing systems to reach full licence savings potential.
• Involve business users in testing: Key for acceptance - plan for testing and book user test resources well in advance.
• Train your staff: SAP HANA is a rewarding platform to work with, and makes completely new use cases possible. Make sure that your team is well trained and understands the new SAP BW features and architectural considerations.
• Document migration issues: Demand that your migration partner maintains a complete issue log and continuously reviews this with your business and technical experts. All issues should be completely understood. If not resolved immediately, they must be properly documented.
• One team culture - with multiple vendors participating, find a strong project lead and build a collaborative culture. Hold all vendors accountable for the overall success of the project.
Users are reporting higher satisfaction with the new level of performance:
• “A wrong click no longer ruins my morning”
• Data loads are now 30-50% faster.
• Report execution is now 40-50% faster on average.
• All InfoCubes are now always loaded at 7:30 am every day - before most people start their workday. This means no more waiting for data loads to complete, and business analysts can use their time more effectively.
• The DB size went down by 80% (1TB -> 200 GB).
• No significant issues reported in the first 12 months of production use.