Merck & Co., an innovative, global healthcare leader that is committed to improving health and well-being around the world, has experienced an explosion of data since it implemented SAP Business Warehouse (SAP BW) nearly a decade ago. At the time, Merck had roughly 2TB of data in its enterprise data warehouse run on a third-party database, and used SAP Business Warehouse Accelerator (SAP BW Accelerator) for operational reporting, which was more than sufficient to enhance query performance. This was still true two years later, when steady growth into new markets increased Merck’s data footprint to 8TB.
With continued expansion and new business ventures, however, Merck’s accelerator functionality was soon on overdrive with roughly 15TB of data flowing through its enterprise data warehouse. This created pressure to provide the business with trusted and timely data, to correlate SAP data with non-SAP data, and to reduce the ever-increasing data load time to run queries.
A rapidly increasing data volume was happening in parallel with an ongoing Merck initiative to break down data silos in an effort to create a single source of truth. The realization that common data and processes were key to a business transformation was the core of this initiative, named “Collaborating as One Merck to Enable Transformation” (COMET).
“We wanted an easier and faster way to get to the information we needed for analysis,” says Selva Neelamagam, Associate Director of Information Management and Solution at Merck. “Because we kept all our data live in one instance instead of offloading it, the data loading performance and speed had begun degrading. The accelerator did improve query performances, but not in all cases — so that was the main driver for looking at another platform to accelerate and liberate our data.”
We were able to create virtual views that could enable business users to perform their analytics without IT’s help, no longer limiting what they could see.
— Nancy Cronce, Associate Director of Information Management and Solution, Merck
A Case for SAP HANA
In deciding to move to a new data warehousing platform, SAP HANA was the logical choice for Merck for several reasons. Because the company was on the latest SAP BW version optimized for SAP HANA, an SAP HANA migration would not entail a time-consuming re-architecture of its underlying SAP BW code. Also, with a mature SAP BW instance, a move to an entirely new data warehousing platform wasn’t a best-case scenario. In addition, Merck already had a taste for bringing the power of in-memory processing to data warehousing; SAP BW Accelerator —like SAP HANA — uses in-memory processing to accelerate query runtimes. The difference, however, is that the accelerator software is limited to accelerating reporting on selected data views. Conversely, SAP BW optimized for SAP HANA runs the entire SAP BW application on the SAP HANA database, meaning data load and calculation performance are optimized in addition to query performance. Because it works as a transactional and analytical database, SAP HANA satisfies Merck’s business transformation requirement for harnessing enterprise data for insight to drive business results.
While the accelerator software was adequate for what Merck refers to as "hindsight analytics," it became clear that SAP BW running on SAP HANA would provide the business with a real-time analytics environment that could leverage all relevant data, integrated across all applications, to achieve superior business outcomes.
The new platform, which would let users be more proactive in how they interacted with data, required the business to view enterprise data warehousing from an analytics perspective rather than from a strict operational reporting perspective.
“We wanted to enable advanced analytics while liberating the SAP data for business users to be able to consume it how they wanted,” says Nancy Cronce, Associate Director of Information Management and Solution at Merck. “SAP HANA was the best of both worlds; we could take advantage of what had already been built, give it the ability to accelerate, and operate in a simpler environment.”
In 2014, Merck built a business case for SAP BW running on SAP HANA that focused on increased performance and financial benefits. In 2015, the company executed its migration plan with a lift and shift off its existing SAP BW database to SAP HANA. To balance cost and performance, Merck used the SAP BW Near-Line Storage (NLS) interface for SAP Sybase IQ, which separates frequently accessed data from rarely accessed data. Because near-line data is stored in a compressed form, it puts fewer demands on resources in the SAP BW system by compressing seldom-used transactional data.
“With our database close to 15TB, we identified the items that could either be deleted or removed,” Neelamagam says. “The first phase of the project was moving those items to the NLS system, which reduced our database size to a more manageable position.”
Merck moved roughly 50% of its SAP BW data into NLS, which has a 10:1 compression ratio. With this phase complete, Merck initiated the database migration, moving the other half of its SAP BW data into SAP HANA (at 4:1 compression) to complete the five-month project in August 2015. Neelamagam and Cronce attribute an eventual successful cutover to top-level executive support as well the use of SAP’s rapid deployment solutions, which is a pre-configured package for migrating data from an existing data warehouse to SAP BW running on SAP HANA. This package contains proven software deployment plans, methodologies, and implementation strategies that help ensure migration projects are completed on time and within budget.
“It was an extremely useful resource,” says Neelamagam. “We had a very knowledgeable resource from SAP to help us through that process; I’d call it the core part of this migration project that helped us move forward. It was time well spent because we were able to manage any issues we found upfront to ensure a smooth cutover.”
We wanted an easier and faster way to get to the information we needed for analysis.
— Selva Neelamagam, Associate Director of Information Management and Solution, Merck
Unlocking Enterprise Data
Because Merck opted for a big-bang cutover rather than “greenfield,” some optimization activities occurred post-migration. Being on the latest SAP BW version did not preclude some architecture simplification for optimizing SAP BW for SAP HANA. What this meant for the roughly 5,000 active users was that, while they almost immediately noticed faster data loads and query runtimes than before, it took them longer to realize that process optimization for SAP HANA meant they could actually look at data as never before.
“People knew how quickly they could get to the data in the new system, but now they realize they can do things differently,” Neelamagam says. This realization speaks to what Cronce and Neelamagam meant about liberating data from the enterprise data warehouse, and it supports the company’s overarching platform strategy for COMET, which is to minimize the influence of IT in delivering pre-built reports.
A specific use case of analytics with SAP HANA for commercial and marketing purposes highlights the impact of the new platform at Merck. When SAP BW was on a third-party database, it was difficult to provide a published layer out of the existing solution, and thus the business started to request pulls of sales information at a granular level — customers, products, and country level. “It can take a long time to pull all of that data out, and there just aren’t enough hours in the day,” Cronce says. “With SAP BW running on SAP HANA, we were able to create virtual views of the information that could be consumed by business users and enable them to perform their analytics without IT’s help, no longer limiting what they could see.”