introduction
The ability to adopt cutting-edge technologies and innovate is not a luxury, but a prerequisite for organizations seeking continued success. We hear from thousands of customers running their business-critical SAP workloads on Amazon Web Services (AWS) and exploring new ways to build innovative solutions to enhance their business operations. With the recent emergence of mainstream generative artificial intelligence technologies, SAP customers are looking to leverage Amazon generative AI services such as Amazon Bedrock to accelerate innovation within their business processes such as order-to-cash, procure-to-pay, hire-to-terminate, and design-to-produce. To this end, earlier this week, AWS and SAP announced an expansion of their strategic partnership to include new capabilities for generative AI. Amazon Bedrock models are now available in SAP Generative AI Hub within SAP AI Core. This enables companies leveraging SAP ERP to build generative AI-powered application extensions using SAP Business Technology Platform (SAP BTP) and foundation models (FM) on Amazon Bedrock while maintaining a clean core. This blog provides technical guidance from AWS and SAP in the form of a “Joint Reference Architecture” (JRA) for customers in the SAP ecosystem considering using FM on Amazon Bedrock.
Building Generative AI with AWS
Customers are looking for guidance on how generative AI can support their overall enterprise modernization strategy, but at the same time, they also expect these solutions to meet their high standards of security and infrastructure reliability for SAP workloads. To help customers modernize, automate, and innovate with generative Gen-AI technologies, we launched Amazon Bedrock as a fully managed service that offers a wide range of high-performance foundational models (FMs) from leading AI companies such as Anthropic, AI21 Labs, Cohere, Stability AI, Mistral, Meta Llama, and Amazon Titan. Amazon Bedrock provides a comprehensive set of capabilities for building Gen-generative AI applications, simplifying development while maintaining privacy and security. Key features include customizing models with your own data, fine-tuning for specific tasks, and Retrieval Augmented Generation (RAG) that uses your company's knowledge base to increase response accuracy. Bedrock also supports building intelligent agents that can automate tasks by interacting with enterprise systems. Strong security measures such as data encryption and compliance with standards such as SOC and ISO ensure a secure environment, making Amazon Bedrock ideal for developing innovative AI-driven applications.
AWS also offers other generative GenAI service infrastructure such as AWS Trainium and AWS Inferentia to enable customers to train AI models and run inference effectively on the cloud at low cost. Additionally, developers can leverage Amazon Q Developer, an AI coding companion, to improve developer productivity by generating real-time code suggestions based on the developer's natural language comments and previous code within the integrated development environment (IDE).
AWS and SAP Partnership
SAP and AWS have partnered since 2008 to help customers run SAP applications more efficiently and innovate faster. Together, SAP and AWS have created a set of reference architectures to address practical business scenarios under the umbrella of modernization, bringing the power of AWS services to SAP customers through the SAP Business Technology Platform (SAP BTP). By adopting a clean core model, these Joint Reference Architectures (JRA) provide a framework for application development and integration extensions to SAP S/4HANA that automate and optimize business processes. Built by SAP and AWS, this JRA provides joint expert guidance to build new scalable applications, analytical dashboards, or machine learning models for customers. AWS and SAP also plan to adapt the JRA to new services and capabilities released by SAP and AWS in the future.
Integrating Amazon Bedrock Generative AI Models with SAP – A Joint Reference Architecture Overview
SAP has been an early adopter of generative AI technology with the release of Generative AI Hub in SAP BTP. Generative AI Hub in SAP AI Core provides purpose-built AI development tools, enterprise-grade access to leading AI foundational models, and robust data control for creating AI-powered applications. Integrating Amazon Bedrock generative AI models through Generative AI Hub in SAP AI Core gives SAP customers access to foundational model families such as Anthropic Claude3 and Amazon Titan. This joint reference architecture enables SAP customers to accelerate their adoption of generative AI and modernize key business processes built on SAP solutions. These innovations can be used as integrated components in embedded use cases within SAP’s RISE and intelligent scenario lifecycle management capabilities or directly in parallel on SAP BTP. Customers can build generative AI solutions using Generative AI Hub in SAP AI Core and AWS services to further enable custom AI capabilities within SAP’s portfolio of cloud solutions and applications. This can provide new insights and optimizations across various business functions, including finance, human resources, etc. SAP and AWS plan to expand the use of Amazon Bedrock capabilities in the Generative AI hub to enable embedding AI capabilities within SAP's portfolio of cloud solutions and applications, including use cases across finance and product lifecycle management.
Let's take a closer look at the architecture and the different components involved.
Amazon BedrockAmazon Bedrock allows you to choose from a wide range of large language models (LLMs) provided through APIs. This JRA uses Amazon Titan and Claude from the Anthropic family of Foundational Models (FMs), pre-trained on AWS with large datasets. These are powerful general-purpose models built to support a variety of use cases. You can use them as-is or privately customize them with your own data.
SAP AI Core Generates AI Hub: The SAP AI Core service exposes AI assets such as large-scale language models to customers and provides a unified interface for SAP applications running in the SAP BTP ecosystem. This JRA uses the Generative AI Hub in SAP AI Core as an access and lifecycle management layer to manage access to Amazon Bedrock and provide endpoints for applications to consume the underlying models. Through the Generative AI Hub, SAP applies a host of content filtering, SAP-specific risk mitigation, and safety guardrails centrally, providing a compliant approach to protect against potential business and legal risks at scale across the SAP ecosystem.
The Generative AI Hub in SAP AI Core gives developers instant access to a wide range of large language models (LLMs) from various providers through a governed commercial and legal framework. This access allows developers to orchestrate multiple models. The Generative AI Hub also connects to the vector capabilities of SAP HANA Cloud, enabling developers to reduce model illusions and incorporate contextual data as embeddings to deliver more customized results for specific use cases.
Storing and Retrieving DataSAP HANA Cloud is a multi-model database management system that helps you build and deploy large-scale intelligent data applications. The SAP HANA Vector engine can support Retrieval Augmented Generation (RAG) for better results from LLM.
application development: The Cloud Application Programming (CAP) model is an approach to developing cloud applications using SAP Build. CAP provides a more structured and seamless framework for data modeling and enhanced integration with other services. CAP provides developers with a range of open source and SAP frameworks to accelerate innovation and streamline development. In this JRA, we use CAP as the entity layer for our application fronted by an SAP UI5 front end.
Authentication and Identification: Large language models require as much data as possible, but ignore user authentication at query time, making integration difficult. To avoid this, use SAP BTP and SAP Identity Provisioning services to manage the SAP and non-SAP identity lifecycle for model prompts.
These components used together enable customers to build scalable and reliable full-stack generative AI-powered SAP applications leveraging Amazon Bedrock and SAP BTP services. This JRA pattern is not only adaptable to a wide range of business process extensions within an enterprise-grade SAP landscape, but also helps maintain the clean core approach recommended by SAP.
Conclusion
In this blog, we discussed guidance from SAP and AWS to provide customers with a reference architecture that enables them to leverage Amazon Bedrock's generative AI capabilities with SAP BTP. This architecture enables you to complement your SAP workloads with Large Language Models (LLM) and Transformer model capabilities to harness the power of your SAP data, resulting in improved insights and operational efficiency at a lower cost. SAP and AWS are currently working to define specific use cases that leverage Amazon Bedrock's generative AI capabilities through SAP BTP to solve business problems and improve customer experiences through automation and innovation. Stay tuned.
Check out the following blogs published by our joint team in 2022-2023 regarding joint reference architectures for SAP workloads on AWS:
SAP and AWS – Joint Reference Architecture to Maximize Utilization and Investment
AWS and SAP – Joint Reference Architecture for IoT Scenarios with Amazon Monitron
To learn more about AWS for SAP, Amazon Bedrock, please see the AWS product documentation, and if you require additional expert guidance, please contact your AWS account team to speak to a local SAP specialist solutions architect or AWS Professional Services SAP specialization.
Join the SAP on AWS discussion
In addition to your account team and AWS support channels, the AWS for SAP solutions architecture team regularly monitors AWS for SAP topics and has started a re:Post to discuss and answer questions to help customers and partners. If your question is not support-related, consider joining the discussion in the re:Post and adding it to the community knowledge base. To learn more about the thousands of active customers running SAP on AWS, see the AWS for SAP page.
credit
The AWS and SAP partnership on the Joint Reference Architecture is the result of close collaboration and contributions from organizations across SAP and AWS. We would like to thank the following individuals for their expertise, support, and guidance:
- Team AWS: Sunny Patwari, Yuva Athur, Ganesh Suryanarayanan, Spencer Martenson, Steve DiMauro, Soulat Khan.
- Team SAP: Madankumar Pichamuthu, Weikun Liu, Daniel Zhou, Sivakumar N, Anirban Majumdar
