BW on HANA TCO - Part 2
By Adam Zeckel
Principal - Business Consulting; Energy, Communications and Services Unit, Infosys
In Part 1, I explored the cost components relative for potential IT cost savings by moving to Business Warehouse (BW) on HANA. In this part I will expand on the three major cost categories: labor, software, and hardware.
The first significant piece of savings comes from labor. With BW on HANA, there are fewer developed objects due to streamlined architecture and faster load times during development, testing, and deployment.
Estimates from SAP and several existing customers show that this adds up to a 20% reduction in effort required to develop and maintain the data warehouse on an annual basis. Because many of these efforts include contracted workers, that savings is easily converted to the bottom line.
The second major area of savings comes from hardware. The growth in SAN space required is largely due to the "old" architecture standards for BW. Data is stored in many different layers on the data warehouse. PSAs are growing and there are layers for staging, corporate history, levels of integration and aggregation. Also, IT organizations often support multiple "production sized" environments for testing, training, performance, etc. The growth of data is then amplified across the landscape.
For BW on a legacy database, this is business as usual. If data were to corrupt or an error were to be found, it's much better to have an intermediate step to recover from as opposed to having to pull millions of records from the OLTP system - a process that can take weeks and affect the performance of the source.
With HANA as the database, such a high level of data redundancy is no longer needed. Loads run much faster for recovery. Also, the query performance is dramatically improved so users can query directly from staging layers without needing the additional layers of aggregation implemented for performance. Other features like indexes and aggregates are also not required.
Not only is the compression rate in HANA much higher, but near-line storage is a perfect complement to HANA. A mature BW environment with 5+ years of data can utilize near-line storage to limit the net add of transaction data into the warehouse.
With data growth, additional database servers are required to meet loading window SLAs. Similarly, BW Accelerator blades must be added to keep up with the growth of data. Moving to HANA allows organizations to avoid these costs.
Finally, software costs associated with moving to BW on HANA are certainly more than in the steady state BW environment. This includes both software licensing for HANA and for near-line software. However, organizations can avoid the cost of additional BWA licensing required to keep up with data growth. It's important to remember that software is only about 20% of the TCO equation.
With all the components of TCO for BW added together and compared to BW on HANA, it's clear that the future state in HANA will provide significant savings for IT organizations. Depending on the size of the implementation, moving to BW on HANA can have a payback period using IT benefits alone of less than 3 years. This analysis doesn't even include the less tangible business benefits that will undoubtedly enable business processes that are interrupted or delayed by timely access to data.