A System Architect’s View: Hospitality Company Re-Platforms Enterprise Data Warehouse to SAP on Azure

In today’s IT landscape, when the need for rapid deployment of resources has created a boom in the cloud computing industry, IT leaders find themselves adjusting to the ‘new normal,’ determining what the future may look like and how the organization will quickly adapt to the next unexpected challenge. The good news is that cloud deployments give IT departments the ability to set up and provision equipment in a matter of hours, versus the possible days or weeks it could take to find physical hardware, making cloud a central feature of rapid deployment solutions. So when a hospitality company approached Protiviti with a monetarily punishing deadline and a need to re-platform multiple complex systems onto SAP, a cloud-based solution was the only approach that could get the job done.

Just before the pandemic hit, Protiviti worked with a popular hospitality company to implement this solution. It involved re-platforming the data warehouse from a system hosted by a separate third-party, to a company-owned solution in the Azure cloud space utilizing SAP® HANA. The data warehouse would take information from the client’s main casino management system, as well as other supporting source systems, land it in the data warehouse, then use native HANA views to recreate the logic in memory and expose it to the front end users through SAP Analytics Cloud (SAC) and SAP BusinessObjects (BOBJ). This project had multiple aspects that needed to be managed and monitored efficiently in order to meet the specified deadlines.

The first steps of the project involved setting up the infrastructure to handle all of the necessary server and networking requirements in the Azure cloud. This involved setting up five servers: one each for HANA, BOBJ, Data Services, a repository server and an SAC connectivity server. The HANA server utilized an M series Virtual Machine (VM) from Azure that has been specifically approved by SAP for production usage with SAP HANA. Without this specific VM, SAP will not consider the installation to be a certified SAP HANA install and may not offer support if things go astray. The other servers were either D or Ds series, which are production ready for smaller workloads. The non-HANA servers do not require nearly as much computing power as the HANA DB, which creates a situation where performance and price can be tuned to desirable levels.

After the server provisioning was completed, the networking portion of the infrastructure needed to be configured. A single virtual network (VNet) was used for all of the servers, with individual subnets consisting of 32 IP addresses per subnet within the VNet. These subnets allowed for grouping of the individual development, QA and production environments while ensuring the machines could still communicate on the same VNet. The servers were set up to only utilize a private IP address for security purposes. This required Protiviti to work with the client’s networking team to create a route-based virtual private network gateway (VPN gateway) connecting the Azure virtual network to the on-site client network. This allowed for anyone on the client’s network or VPN to have access to the machines in the Azure Virtual network. The use of a route-based VPN gateway allows for multiple site-to-site connections, which could come in handy if multiple sites need to be connected to the Azure network. This completed the infrastructure setup of the landscape for a single tier using Azure.

With the infrastructure completed, the servers were ready to install the various SAP software that would run on each of them. At this point, the installation process follows the normal steps a regular SAP install would take. Information Platform Services (IPS), SAP BusinessObjects (BOBJ), SAP Data Services (BODS) and SAP HANA were all provisioned on the respective servers over a week’s time and were ready to handle ETL, model and report development. After a few ‘hello world’ tests to ensure everything was working as expected, full development began to move the data out of the legacy data warehouse and into the new data warehouse in Azure. From here, models were created using the report output from the legacy reporting system on Cognos. The generated Cognos queries were parsed and reorganized by the project team to create queries that could be replicated through HANA model development and used for the complex reports that were necessary for day-to-day operations.

Through this rapid cloud deployment process, the client was able to stand up a new environment quickly and efficiently while minimizing costs, and also saw a large increase in query speed and reporting responses, reducing runtimes of one report from more than eight hours to 90 seconds. In addition to the increase in speed, the team created an infrastructure that could be easily redeployed within a matter of hours instead of weeks. Additional deployable templates could be easily created for a single machine or the entire landscape, then deployed through the Azure portal or Azure PowerShell if a programmable solution was required.

Cloud deployment of SAP solutions provides a fast, scalable and repeatable process for deploying full SAP stacks in the cloud. Utilizing SAP certified Azure Virtual Machines to house the SAP servers allowed for a quicker install and configuration for the client, helping to ease the strain created by the looming deadlines they provided. This could not have been completed without the collaboration of the client’s IT team, their third-party vendors and the deep SAP expertise from Protiviti. Through this process, we created a means to quickly deploy new landscapes in the cloud, peer the cloud network with the client’s on-site network and facilitate data movement from multiple sources, through the cloud network, and into the HANA data warehouse.

About the Authors

Mickeal Taylor
Senior Consultant
Technology Consulting – Enterprise Application Solutions