FirstGroup Drives Better Performance and Business Intelligence Using SAP HANA on AWS

Over the past 12 years, First Rail invested in SAP Business Intelligence (BI) as its core platform for management information, reporting, and analytics.

“Because we’re determined to continuously improve our train experience and provide rail customers with more real-time information, we’ve prioritized the development and optimization of the channels our rail customers use during their interactions with us—including our website and our mobile application—as well as data collection,” says Waugh.

First Rail built its environment on SAP Business Warehouse (SAP BW), a model-driven data warehousing product, to manage and drive insights from data.

Initially, the company chose to run everything on Oracle and hosted with a private cloud provider. Soon, the FirstGroup team encountered scalability challenges as it sought to process and run analysis on its data.

"Our scaling needs were significant because of the volume of data we were looking to process,” says Tom Grant, head of rail business intelligence for FirstGroup. “We made a technology-roadmap decision to move forward with BW and upgrade it to run on HANA.”

FirstGroup decided to upgrade SAP BW to run on SAP HANA for two reasons: HANA’s performance and statistical capabilities. “HANA’s in-memory database and architecture—which would give us the ability to rapidly process more data and make decisions based on many more data points at once—as well as its statistical capabilities made it a natural fit for us, particularly given our use of other SAP applications,” says Grant. “The question then became: Where do we want to run and host BW on HANA?”

After starting with an evaluation of its existing provider’s capabilities, the FirstGroup team began to explore the benefits of migrating BW to HANA on Amazon Web Services (AWS). “We had positive experiences running an SAP HANA One environment on AWS but only knew a little about what was possible,” says Grant.

While cost analysis was crucial, the FirstGroup team also wanted to understand what benefits it could gain by moving to AWS beyond simply running SAP systems on AWS. “We were keen to explore additional cloud-based services with real-time data capabilities, in complement to our SAP investment and platform. The art of the possible made AWS a very compelling provider to us,” says Grant. Once the team chose to migrate to AWS, it began to work closely with AWS Professional Services.

At the beginning of the engagement, AWS Professional Services held a candid discussion with FirstGroup about the company’s options for migration and running the platform following the migration. “Working with AWS Professional Services was an excellent choice,” says Grant. “We chose to work with AWS Professional Services so we could migrate to and land on AWS with confidence that we were well-architected.” For the AWS Professional Services team, taking advantage of the SAP HANA on AWS Quick Start Reference Deployment was vital to meeting the aggressive migration timelines required by FirstGroup.

FirstGroup and AWS Professional Services migrated SAP BW, SAP Portal (used for business intelligence), and Business Objects Business Intelligence (SAP BOBI) software to HANA on AWS. The project went live on time, within budget, and with zero defects in the AWS migration portion.

“The exceptionally complex heterogeneous database migration, which included moving from Oracle to HANA and thus changing our operating system from Windows to Linux, and the application upgrade to support running on HANA, took only a little over five months,” says Grant.

FirstGroup runs its 4 TB HANA analytics workload on four Amazon Elastic Compute Cloud X1.16xlarge instances, each of which is powered by Intel Xeon E7 8880 v3 processors, offering 64 vCPUs and 976 GiB of DRAM-based memory. It’s a scale-out architecture, in which HANA has two deployment modes: scale-out for BW and scale-up for ERP. “We’ve gone with a four-node scale out using the X1 platform; the underlying services for this are Amazon Elastic Block Store for persisting that database to disk, and Amazon Simple Storage Service as a long-term backup,” says Grant.

As a result of the increases in performance that running SAP HANA on AWS provides, FirstGroup has been able to gain richer insights and encourage more data analysis amongst its employees. “We originally migrated to HANA on AWS to drive better forecasting, optimizing, and decision-making around pricing for our revenue-management group,” says Grant. “But by moving all of our SAP systems to HANA, we’ve taken hundreds of users of our BI platform and planted them on HANA, giving them a massive performance improvement and changing the way they can analyze data and the depth of exploration they can achieve. Within four weeks of moving to AWS, we noticed a 25 percent uptick in query executions.”

As soon as FirstGroup was live on AWS, users saw a 20x improvement in SAP application performance compared to the on-premises environment, and a 7x improvement in performance during peak times. “Immediately following the migration, we moved from 80 percent of our data sets being available at the start of the working day to up to 95 percent of business data available for analysis at the start of the working day,” says Grant.

The company’s backup and recovery capabilities have also improved thanks to its use of AWS. In the past, it took 12 hours to run backups on legacy infrastructure; FirstGroup is now able to do a full database backup on AWS every night in three hours.

Thanks to the help of AWS Professional Services, FirstGroup was busier with new projects in its first year than anticipated. “The support we received from AWS has been quite impressive across AWS teams,” says Grant.

As the company explores what is possible using AWS, its plans involve exploring the analytics capabilities AWS provides and using AWS-native services to drive further insights using data. “Today, we’re looking at how we can use data—from trains and from operational systems—in real time to make better operational decisions, particularly during disruption,” says Grant.