This blog explores the benefits of using one database to support the needs of an application (ERP, for example) while simultaneously supporting the needs of complex analytics. SAP’s Suite on HANA (SOH) and S4HANA offerings use one database to simultaneously support the application and complex analytics.
This is not a new concept, but the technology available to support the single database approach today is far superior to the one available when ERP systems were first adopted. There are several advantages:
1. Reporting off of the latest information in the application (real time information)
2. A simpler and less expensive system (lower cost of ownership)
3. Less effort to bring new reports to end users (a higher focus on innovation via quicker availability of information)
When ERP systems were first brought to market in the 1990s, one of the complaints was ‘we put all of this data in these systems , but we can’t get it out’. Everyone knew the information was in the system and they wanted to pull the information out in a way that the systems were not designed to support. This brought the advent of operational reporting. Operational reports were run on the application database. These reports used the single database approach and had the advantage of running on the most current data set. Over time, the sheer magnitude of data and the complexity of the reports started to considerably slow the database and thereby the application (remember the phrase “the system is slow, accounting must be running their reports”). The database could not handle supporting the application while simultaneously returning complex queries. This led to the birth of the data warehouse.
The data warehouse would solve this issue. When the application was not being used, typically at night, data was exported out of the application database, into a data warehouse. There were inherent advantages and disadvantages to this approach. Since the data had been copied out of the application database into the data warehouse, now complex queries could be written that would not affect the application. The disadvantages being that not all of the data could be copied into the data warehouse and that the latest information in the data warehouse was old, and finally, there was the added cost of a new additional data warehouse infrastructure and application. With all of these disadvantages, this was still the best option available.
Basically, if the database were fast enough to support both the application and complex analytics simultaneously, there would be no need for separate data warehouses and data marts. Creation of layers and layers of redundant data would be eliminated – leading to better data quality, simplified landscape and thus a lower TCO. That is the value proposition of both SAP’s Suite on HANA (SOH) offering, as well as the latest offering, S4HANA.
The key to HANA is not just that it is a true in-memory database that eliminates performance issues; it is also in the technology that is built into it. For example, it can simultaneously process OLTP and OLAP – thereby bridging the chasm that got created between the application systems and the data warehouses. There are other features that provide multiple advantages, including the elimination of the upkeep and feeding of the separate data warehouse, faster delivery of new reports (please refer to my blog ‘Is it time for a new data warehouse’ at http://goo.gl/p5JcRw for a deeper discussion on faster delivery of new reports), as well as running the analytics on the most up to date information. It’s time to take a new approach to analytics, especially if it is one that is proving to be better, faster, and cheaper.