Leading electronics manufacturer Atmel partners with SAP and Tech Mahindra to transform its business intelligence (BI) landscape with SAP HANA – resulting in a 360-degree view of information across all functions.
As a leading global manufacturer of microcontroller solutions (MCUs), electronics firm Atmel recently entered the market for the Internet of Things (IoT) with the launch of a range of touchscreen controllers and touch-sensing solutions. The company unveiled its portfolio of IoT solutions at Electronica in November 2014. The IoT market is estimated to reach $7.1 trillion by 2020, according to IDC, and Atmel wants to be a well-positioned player.
Atmel, based in San Jose, California, embarked on a major IT transformation under the leadership of Sabyasachi Bose, Chief Information Officer and Vice President of Information Technology, to keep pace with the company’s business strategy and plans for growth. An overhaul of the company’s BI system with an accompanying migration to SAP HANA was among three major transformation projects initiated in 2014.
As a result of the success of this SAP HANA project, reports now take two to three seconds compared to 10-15 minutes previously. Atmel also rolled out analytics solution Tableau on top of SAP HANA, which has opened up a very powerful visualization and user experience at the company. Queries that once took four to six minutes are now ready in one to two seconds. This has brought tremendous benefits to Atmel’s leadership team and adoption is 100%.
“Typically the state of business intelligence is extracting data from transactional systems, transforming, creating data warehouses and retrospective reporting,” says Bose. “Most companies are unable to get past this stage to transform data into information. Atmel was no exception, but with the implementation of SAP’s BI on SAP HANA and adding a semantic layer aided by Tableau analytics, we have enabled the business to make better informed decisions.”
Reduced complexity, better UX with business intelligence
Pankaj Tiwari, Senior Director, Global Business Apps at Atmel, laid down the objectives of the BI initiative. His prerogative was to increase existing reports’ performance at least by factor of 10, enrich how the information is being consumed, and improve the user experience. “We did not want to redesign the entire existing BI landscape and take several years to complete it. Rather we wanted to hide all the complexity from business and enable them for self-service analytics and reporting within six months,” says Tiwari.
The existing architecture at Atmel was unstable, and needed 24×7 monitoring and support to keep it running. Even then, it still could not meet the basic reporting needs of the business. After evaluating various options, SAP HANA stood out as the clear winner to help meet Atmel’s objectives. The prime responsibility of the transformation process was entrusted to Taru Gupta, Senior BI Manager, Atmel.
Tech Mahindra brings its SAP HANA expertise
After careful consideration, in Feb 2014 Atmel decided to partner with Tech Mahindra, an SAP partner that is one of the largest systems integrators operating globally in the SAP space.
In 2013, Tech Mahindra received recognition for having the highest number of consultants certified on SAP HANA. Currently, it has more than 300 people certified on SAP HANA and has 11 SAP HANA customers operating across spectrum of industries, including many top brands. The company is investing heavily into SAP HANA, with focus on ABAP on SAP HANA, BW on SAP HANA, BPC on SAP HANA, CRM on SAP HANA, SCM on SAP HANA, enterprise SAP HANA, SAP HANA side car, SAP Business Suite on SAP HANA, managed cloud delivery with specialization in SAP HANA Enterprise Cloud and SAP HANA Cloud Platform.
Journey to the finish line
The project kicked off April 21, 2014, with a combined global team from Tech Mahindra and Atmel. Project strategy was to accomplish an upgrade of SAP BW from 7.0 to 7.4, and a migration from the existing database to SAP HANA using the SUM DMO approach. In addition, Atmel had to split its ABAP and JAVA stacks for the new landscape. The project scope was defined to include the major 181 queries. Architecture was to partition one of the SAP HANA boxes into two instances for development and quality assurance, and one dedicated appliance for production. Testing requirements were clearly spelled out from the start.
The project was filled with twists and turns, and the occasional surprise. The very first dry run in the SAP HANA dev environment could not proceed past the 10% SUM migration tool. This issue was traced back to inconsistencies in Atmel’s SAP BW system tables that may possibly have been the result of a project in 2008. TechMahindra resolved the issue by reimporting the missing support pack stack components.
Initial testing revealed that performance had not improved as expected. This was a setback for the team. Gupta closely examined the application server hardware, its configuration, and the infrastructure around the whole BW – SAP HANA landscape. Such detailed analysis proved to be an absolute win, resulting in a performance gain of up to 200%. Tiwari and Gupta architected the semantic layer for Atmel’s business processes and implemented it on SAP HANA. This helped improve the processing time at the database layer by another 100%.
Special testing was performed for the finance department, wherein, Gupta’s team validated the department’s 25 golden queries over a period of 5 years. The team proceeded month by month and drilled down to the lowest possible details, typically with 30-40 key figures in every report. The testing effort was gigantic, but worthwhile for user acceptance.
As the go-live date neared, Gupta decided to do one last dry run, but this time on the SAP HANA production appliance. This decision turned out to be a savior for the project, as a faulty mother board was discovered and quickly replaced.
Atmel went live with SAP HANA on October 20, 2014, according to plan. All of its existing reports saw an immediate performance improvement by a factor of 10-30 times.
“This was one of the most complicated projects I have managed,” says Gupta. “We had to pass the ball several times between hardware provider, SAP, Tech Mahindra, and Atmel before we could even identity the root cause [of an issue], let alone fixing it. Having everyone involved, coordinating, not taking your eyes of the goal, and management (Pankaj) support was the key to success. We went live without a single incident.”
One-team concept drives collaboration
Fundamental to the project success was the one-team concept that drove the collaboration. To achieve success with SAP HANA, the team required very strong collaboration between hardware, software and implementation partners. VT Reddy, Global Head of SAP Analytics and HANA Practice in TechMahindra, says, “It’s a combination of Tech Mahindra, SAP, and Atmel. It was a complete collaboration.”
User community enjoys new reporting capabilities
Atmel recognized immediate ROI from the performance gain on the reports. Users can get data within seconds to minutes; plus, year-over-year and quarter-over-quarter trending is now feasible. As a result, executives now have the ability to study data patterns, which is proving extremely useful in decision making. Atmel has also seen additional benefits from freed-up data storage and related hardware costs.
In post-migration testing, Atmel found performance improvement of as much as 300% for some queries. Below (click on the image to see the detailed information at full scale) is a graphical representation showing the time taken by queries (in blue) before the migration to SAP HANA and the time after the migration was done (shown by orange lines).
Employee hours have been freed up to devote to more meaningful tasks. Rather than trying to squeeze data from the system for monthly reporting, employees are now focusing on trending and analyzing the data more holistically.
Atmel is now working on providing enhanced visual analytics and dashboard capabilities to its users. For dashboards, the company has found success in implementing Tableau on SAP HANA. The new system is able to analyze millions of records in seconds and display the results, which was unthinkable before.
Top image: Shutterstock