Businesses in their digital transformation journey are eager to explore new domains at great speeds and are faced with enormous challenges, such as the following:
- A need to thoroughly analyze enormous datasets from a wide range of detailed data sources to generate accurate and real-time contextual information for decision making
- A need to bring diverse networks of business users into their business processes with a large number of stakeholders
- The next wave of analytics requires executing complex models that can capture several perspectives and mash them up into easily consumable information cubes.
In the past, analytical applications underwent a variety of compromises in order to enhance their practicality, e.g. standard reports over aggregate data. Such concessions, however, are no longer needed in the era of SAP HANA. Businesses have the need to run real time reports from very detailed data with a possibility to drill down into line items without the pre-fabrication of information cubes. New sources of data are being added for 360-degree exploration. Thus, the biggest challenge is to embrace complexities arising simultaneously in all these multiple dimensions while preserving the real time experience.
The following example shows how smart collaboration can result in unprecedented performance and value; this is exactly what was achieved at packaged foods company ConAgra. Both SAP and ConAgra took on the challenge to deliver the required performance with a minimum amount of hardware, thus targeting huge savings in this area as well as that of maintenance. The result of this effort was SAP Total Margin Management running on HANA to answer complex questions instantly – over enormous amounts of data, even under severe constraints.
The project at hand primarily consisted of multi-user execution of complex calculation-based queries over half a billion records of data covering billions of planning points, execution of ‘what if‘ scenarios, and incremental updates of data. The introduction of HANA at ConAgra has eliminated a number of boundaries that traditional systems encountered in the past, and made high performance analytic functions possible using in memory columnar techniques.
In a complex application as such, it was necessary for the Customer Innovation & Strategic Projects team to understand platform capabilities, and to optimally tie them to ConAgra’s business requirements. The following characteristics of SAP Total Margin Management were key to delivering performance:
- Fast execution planning reports over a data of billions of planning points
- Ability to support large analytic loads
- High throughput rates during data import and other transfer jobs so they can run within stipulated time windows.
Furthermore, optimal performance of a HANA analytics solution, such as SAP Total Margin Management, generally depends on a number factors, including:
- Data volume – primarily size of key figure and planning object tables
- Load on the system -number of concurrent users
- Complexity of key figures
- Rate of modification of data
- Nature of the reports frequently called by users
A business as complex as ConAgra pushed the boundaries on all these fronts. Under these circumstances, performance improvement by a factor of over 100 was ultimately made possible by systematically applying sound principles of software performance and relational theory. A “tuning” and “optimization” type effort could have in contrast only brought a small incremental improvement. This approach has progressed the customer from waiting for results for hours, to receiving reports within seconds – all while taking on a much larger active user base.
While the details of various methods used in performance transformation at ConAgra could well be the topic of another detailed blog, here are some examples for illustrative purposes:
- Streamlining planning levels
- Optimal handling of hierarchies
- Pushing aggregations to OLAP engine
- Reduction of data flowing through the calculations
- Root optimization and other methods designed to unleash the power of HANA
While the platform does deliver formidable performance at all times, it takes a special ingredient to achieve optimal outcomes that suit specific needs. Benefits of great application performance and the powerful characteristics of HANA are fully realized when there is collaboration between SAP and the business – as was experienced during the co-innovation effort with ConAgra. For example, it was important to reorganize data and application workloads in a way that the system could tap into the capabilities of the platform in the most optimal fashion. I believe that this implementation of HANA must rank very high in the list of performance-optimized implementations of any HANA application to date. The learnings from this implementation are easily applicable to most HANA applications dealing with massive amounts of data and complex analytics.
This collaboration between ConAgra and the Customer Innovation & Strategic Projects team at SAP has been immensely successful and the key factor in making it so has been the willingness of ConAgra to take a fresh look at all aspects in co-innovating on this solution in their quest to deliver a high-performing solution for their end users.
P.S.: David Tomlinson, Sr. IT Director at ConAgra Foods, was featured in a HANA Spotlight Call just earlier this week, speaking about SAP Total Margin Management and the collaboration with SAP Customer Innovation & Strategic Projects. Click here to listen to the recording.