SAP and AWS – Joint Reference Architectures to maximize utilization and investments
SAP and AWS have been partnering since 2008 to innovate on behalf of our customers to bring the flexibility and agility of the AWS platform to manage SAP’s cloud applications and workloads. As a logical progression to this partnership, SAP, and AWS have joined forces and have come up with a set of Reference Architectures to tackle practical Business scenarios. These Joint Reference Architectures (JRA) built by SAP & AWS will be the foundation for building new applications, analytical dashboards, or machine learning models for our customers. SAP & AWS are committed to updating and improving the architecture with new services and features released by SAP and AWS in the future.
Joint Reference Architecture (JRA)
The JRAs have been grouped under three logical pillars based on the key business benefits to maximize the return on customers’ investments. The architecture diagram above depicts a typical data flow across different layers for an SAP on AWS setup. These architectures can be enhanced by customers to fit their business needs.
- Data-to-Value Architecture – Architecture patterns outlined in the data-to-value pillar help customers to blend SAP data with non-SAP data. By securely federating and blending data from SAP and AWS sources such as Amazon Redshift, Amazon S3, and Amazon Athena through SAP Data Warehouse Cloud, customers can do more efficient analysis, sales planning, financial planning, and training of machine learning models.
- Integration and App Development – In the Application Layer, the integration and app development pillar provide a backbone for building business-ready applications either by extending the existing ones or creating new ones with custom logic and business rules. These applications can now access data stored in external data stores such as Amazon Aurora and if required, can also use predeveloped content or customized integration protocols to interface with other business applications.
- Platform Foundation – Platform Layer: Once these applications have been built, they should also be made highly available to provide the best customer experience. And there is also a need for customers to provide an intuitive and simple interface to access their applications to increase adoptability. As highlighted in the above architecture, SAP BTP services, such as the SAP Launchpad service, can act as an entry point for these applications. And AWS services like Amazon Route53 ensure a seamless experience by providing resilience and high availability that complements SAP’s Business Technology Platform.
Data is one of the most valuable assets of any business. Customers usually have their data hosted in various AWS cloud services with their mission-critical data residing in business applications like SAP S/4HANA Cloud and BW/4HANA. It is crucial that customers have access to this data in a consolidated, real-time and secure environment so it can be easily consumable for practical applications like Analytics, Machine Learning, Enterprise Planning, etc. The set of architectural patterns listed below recommends the optimal ways to federate data leveraging SAP BTP and AWS cloud services:
1. Assemble Distributed Data
One of the major requirements that SAP customers face is to figure out the best approach to combine the distributed data from SAP and non-SAP systems. Now with data federation-based architecture customers can bring together the data from all the sources, to maximize value and draw insights from it. Consider a customer scenario where IoT data is stored in Amazon Redshift and there is a need to combine this data with business-critical processes and master data into SAP Data Warehouse Cloud. Now, this use case pattern can be addressed by data federation from Amazon Redshift into SAP Data Warehouse Cloud.
A similar pattern can be extended to use cases when data is stored in Amazon Simple Storage Services (S3) and accessed via Amazon Athena. Customers can now federate this data into SAP Data Warehouse Cloud to derive meaningful insights. This use case has been covered with a recommended solution in the blog for Data Federation with SAP Data Warehouse Cloud from Amazon Athena. The highlight from the above scenarios is the ability for customers to federate data securely while avoiding any replication making it more economical and performant.
In both, the approaches data federated from AWS to SAP Data Warehouse Cloud can be visualized in SAP Analytics Cloud.
2. Machine Learning to Maximize value from the Federated Data
Machine learning is one of the key pillars of digital transformation. To set up an effective machine learning algorithm a comprehensive insight into how processes work and what data is being consumed is required. With a service like Amazon SageMaker, customers can now draw insights and set up the baseline for the ML algorithm. Additionally, SAP’s FedML library will enable this by facilitating access to the right data while avoiding replications. SAP’s recommended approach on how to utilize Amazon SageMaker to build Machine Learning Models on federated data to train, predict and write back results to SAP systems has been discussed in the blog.
3. Enterprise Planning
As an operational or financial planner, you want to have access to your data so you can analyze, plan, and forecast. This data might reside in SAP (applications like Analytics cloud and Data warehouse cloud) or non-SAP applications. This architectural pattern will allow you to have a consolidated view of the data and extend it into SAP Analytics Cloud (SAC) to allow effective enterprise planning. More details on this approach with implementation steps have been discussed here: Enterprise Planning with AWS data using SAP Data Warehouse Cloud and SAP Analytics Cloud
4. Federate SAP Data into AWS
For use cases, when data is predominantly stored in an AWS data service but still needs to blend with SAP’s business-critical data, we recommend this architecture pattern where data from sap applications like SAP Data Warehouse Cloud will be federated into Amazon Athena. Details on this architecture along with implementation steps have been discussed in this blog: Querying SAP Data Warehouse Cloud from Amazon Athena using Amazon Athena Federated Query
Customers who have leveraged their existing investments in SAP and AWS have maximized the return. The reference architectures and guides on how to federate data to perform business-critical operations like Enterprise Planning and Machine Learning are helping them to derive more value. These guidelines will ensure the following key benefits:
- Securely federate data both into and from SAP sources
- Avoid building complex data replication pipelines
- Real-time access to data
- Reduced data storage costs
Integration and App Development
SAP Cloud Application Programming Model (CAP) framework has been widely adopted by many customers to build enterprise-grade services and applications. Many a time, customers face situations when these CAP applications must consume not only SAP data (stored in SAP S4/HANA for example) but also non-SAP data stored in other databases like Amazon Aurora. Ensuring that such CAP applications not only perform efficiently but are also durable and resilient is very important. Also enhancing these CAP applications with other practical features like notifications has also been discussed below:
1. Durable and Resilient applications
A CAP application with SAP S/4HANA data integration along with non-SAP data stored in Amazon Aurora can be made more durable and resilient by implementing a replica at the database level, especially for read operations. This read replica implementation of Amazon Aurora along with Amazon Route53 outlines the architecture customers can leverage to build highly available CAP applications. The core idea is about creating a read replica for Amazon Aurora, preferably in a different region, and augmenting with auto-scaling policies. More details about this implementation with solution architecture and implementation steps have been explained in this blog: Distributed Resiliency of SAP CAP applications using Amazon Aurora (Read Replica) with Amazon Route 53
2. Real-time Notifications
Real-time event-based notifications like Email, SMS or push notifications could prove vital for critical applications that demand immediate actions. This blog details one such scenario with a recommendation on how SAP Event Mesh can be combined with Amazon SNS to send notifications based on events triggered from SAP S4/HAHA
This section will come in relevant for all customers looking to build enterprise-grade SAP Cloud Application Programming (CAP) model-based applications with distributed data. The JRAs in this section will elaborate on how to make these CAP applications resilient and durable thus delivering the following benefits:
- Durable application, that is highly available
- Reliability of distributed data
- Resilient applications with trusted Integration
- Reduced latency
- Real-time push Notifications that guarantee faster response times
1. High Availability
As the trends indicate, high availability is an implicit expectation from the customers of any critical service. A highly available solution could make a difference in providing a seamless customer experience. The architecture outlined in the blog explains how to make SAP BTP Launchpad service Highly Available (HA) by using Amazon Route53 service. This concept can be extended to other BTP services, like the SAP Cloud Platform Integration, too. Details about that will be shared in the upcoming blogs.
2. Geography-based content delivery
The HA scenario discussed above and the Amazon Aurora read replica architecture discussed under the Integration and App Development section, both, utilize Amazon Route53 to act as the front-end DNS resolution tool. Amazon Route53’s traffic control policy provides customers the flexibility to build redirection rules based on geographic location. This feature equips customers to build solutions that can deliver content based on end users’ location providing the flexibility to avoid any latency and to deliver relevant local content.
The architectures in this section address one of the mandatory requirements of every business-critical service – High Availability (HA). Recommendations on how to make your BTP services like SAP Launchpad highly available to ensure the following advantages:
- Stability and Availability of the service
- Geo-location-based content delivery
- Positive customer experience
- Load balancing and distribution
What to expect next?
The partnership between SAP and AWS will be one of the highlights of TechEd 2022 with a session DT200 dedicated to the same. AWS re:Invent 2022 will also have a joint session to draw attention to these joint reference architectures which for sure will truly benefit our customers and partners.
In the future, SAP and AWS teams will continue to collaborate and identify more use cases and services to be combined into reference architectures. We will continue to publish all these findings and more related materials in the future.
The above reference architectures are the results of teamwork and contributions from both SAP and AWS. I would like to thank the following SAP team members for their efforts: Sangeetha Krishnamoorthy, Weikun Liu, Mahesh Kumar Palavalli, Shanthakumar Krishnaswamy, and Sandesh Shinde. And Sivakumar N, Martin Frick, Uwe Klasing, and Anirban Majumdar for their support and guidance.
And special thanks to the following AWS team members for their inputs and feedback: Sunny Patwari, Sabareesan Radhakrishnan, Rajesh Chigurupati, Amrish Patel, Renga Sridharan, and Soulat Khan.
If you have any questions, please leave a comment below or contact us at email@example.com.
Comprehensive and amazing blogs, thank you for sharing.
I really would like to get a comprehensive step by step that partners could leverage when considering building new solutions. Please also consider resource/cost efficient solutions so that we can boost the SME adoption. Thank you very much and congrats for this blog.
Fault prediction in Object Oriented Software using Neural Networks