Deploy a Scalable OwnGPT Model on Oracle Cloud Infrastructure for ERP Workloads
The deployment of OwnGPT on Oracle Cloud Infrastructure (OCI) has demonstrated accelerated model development, improved productivity, and strong return on investment (ROI).
The objective for this reference architecture was to enable users and functional consultants to interact with Enterprise Resource Planning (ERP) data using natural language through OwnGPT, a domain-trained AI model deployed on OCI, initially focused on Oracle E-Business Suite R12 (Accounts Payable and Receivable modules), with extensibility to other modules and ERP platforms.
Our initial deployment utilized A10 Tensor Core GPUs, providing accelerated model training and inference. Integrating Oracle HeatWave MySQL further enhanced database performance, while OCI Object Storage ensured secure and efficient data management. Leveraging Oracle E-Business Suite Vision as the primary data source enabled OwnGPT to process structured enterprise data effectively.
Deploying OwnGPT on OCI’s GPU instances highlighted the vital role of infrastructure in fostering AI innovation. OCI's flexibility in deploying GPUs across Bare Metal and Virtual Machine environments allowed for tailored configurations to meet evolving business needs. Additionally, Oracle’s low-latency Remote Direct Memory Access (RDMA) networking accelerated distributed model training, while OCI Block and OCI Object Storage provided scalable data management for large datasets.
Architecture
This is a production ready OCI deployment of an AI-powered ERP solution (OwnGPT), designed for secure, scalable, and modular AI-based query resolution. It integrates Oracle E-Business Suite (ERP) with components like Airflow, Oracle Base Database Service, GPU compute for GPT, and storage while connecting securely to both on-premises users and OCI services.
- AI-Ready (GPU-powered) for LLM workloads
- Modular and layered design separates load balancing, app, and AI logic
- ERP-integrated and tightly linked with Oracle Oracle E-Business Suite and ready for AP, AR insights
- Secure and scalable with WAF, IAM, NAT Gateway, and OCI-native security features
- Highly Available and uses fault domains to protect against hardware failures
The following diagram illustrates this reference architecture.
Users Groups are internal users or business analysts who will access the application via web or VPN. The Internet Gateway provides access to the OCI environment from external networks (e.g., employee web browser access). All resources exist in a compartment called OWNTESTHUB and are distributed across three Fault Domains for high availability. The entire architecture is built inside a secure virtual network. Oracle Cloud Infrastructure Web Application Firewall (WAF) protects from threats like SQL injection, DDoS, and so on.
The Load Balancer distributes incoming traffic across backend web, API servers for scalability and availability. WEB API (VM.Standard.E5.Flex) hosts the front end and REST APIs for the OwnGPT application. AIRFLOW (VM.Standard.E5.Flex) manages task orchestration, scheduling jobs (e.g., model training or ETL). OWNGPT (VM.GPU.A10.2) has a GPU-accelerated VM that runs the OwnGPT model, ideal for real-time AI inference.
Oracle Autonomous Data Warehouse stores ERP related structured data (AP, AR modules) for fast querying. Oracle E-Business Suite APP (VM.Standard.E5.Flex) uses backend services for Oracle E-Business Suite (likely running connectors or middleware). OCI File Storage provides shared storage, useful for Airflow DAGs, app artifacts, or logs.
OCI Identity and Access Management controls user roles, policies, and secure authentication. Service Gateway connects to OCI services like OCI Object Storage without going over the internet. OCI Object Storage is used for storing model checkpoints, training datasets, logs, or archives. The NAT Gateway enables private subnet VMs to initiate outbound internet traffic securely. Response Generation Model extracts relevant data, interprets context, and generates a human-readable response without requiring manual report generation.
- Oracle E-Business Suite
Oracle E-Business Suite comprises a set of products (human capital management, order management, procurement, and logistics) that supports evolving business models, drives productivity, and meets the demands of the modern mobile user.
You can provision Oracle E-Business Suite on Oracle Cloud Infrastructure or migrate Oracle E-Business Suite environments from their data center to Oracle Cloud Infrastructure, creating multihost, secure, and high-availability topology.
- Oracle Autonomous Data Warehouse
Oracle Autonomous Data Warehouse is a self-driving, self-securing, self-repairing database service that is optimized for data warehousing workloads. You do not need to configure or manage any hardware, or install any software. Oracle Cloud Infrastructure handles creating, backing up, patching, upgrading, and tuning the database.
- OCI Object
Storage
OCI Object Storage provides access to large amounts of structured and unstructured data of any content type, including database backups, analytic data, and rich content such as images and videos. You can safely and securely store data directly from applications or from within the cloud platform. You can scale storage without experiencing any degradation in performance or service reliability.
Use standard storage for "hot" storage that you need to access quickly, immediately, and frequently. Use archive storage for "cold" storage that you retain for long periods of time and seldom or rarely access.
- File
Storage
Oracle Cloud Infrastructure File Storage provides a durable, scalable, secure, enterprise-grade network file system. You can connect to OCI File Storage from any bare metal, virtual machine, or container instance in a VCN. You can also access OCI File Storage from outside the VCN by using Oracle Cloud Infrastructure FastConnect and IPSec VPN.
- Identity
and Access Management
Oracle Cloud Infrastructure Identity and Access Management (IAM) provides user access control for OCI and Oracle Cloud Applications. The IAM API and the user interface enable you to manage identity domains and the resources within them. Each OCI IAM identity domain represents a standalone identity and access management solution or a different user population.
Recommendations
- VCN
When you create a VCN, determine the number of CIDR blocks required and the size of each block based on the number of resources that you plan to attach to subnets in the VCN. Use CIDR blocks that are within the standard private IP address space.
Select CIDR blocks that don't overlap with any other network (in Oracle Cloud Infrastructure, your on-premises data center, or another cloud provider) to which you intend to set up private connections.
After you create a VCN, you can change, add, and remove its CIDR blocks.
When you design the subnets, consider your traffic flow and security requirements. Attach all the resources within a specific tier or role to the same subnet, which can serve as a security boundary.
Considerations
When implementing this reference architecture, consider these options.
- AI Processing
OwnGPT processes the query using trained context from AP, AR modules.
- Output Delivery
Response is returned in real-time or near real-time via UI or chat interface.
- Technical Metrics
- Response time: Average 1.2 seconds, 95th percentile 2.5 seconds, complex queries 3.7 seconds max.
- Throughput: Peak 120 concurrent users, 500 queries per minute, 99.97% API success rate.
- Resource utilization: CPU 42% average (78% peak), memory 4.2GB (7.1GB max), storage efficiency 12MB per company for vector data.
- Scalability: Linear scaling up to 50 companies, stable performance with 100,000+ document chunks, 25+ simultaneous connections.
- Business Impact
- Productivity: 68% reduction in search time, 42% fewer repeated technical queries, saving 3.5 hours per employee weekly.
- Accuracy: 92% on company-specific questions, 87% on technical queries, 95% accuracy in source identification.
- User Satisfaction: 4.7/5 rating, 94% adoption among eligible employees, 76% reduction in support tickets.
- ROI: Break-even in 4.2 months, 327% ROI over 12 months, estimated $142,000 annual savings for mid-sized deployments.
Explore More
Review these additional resources to learn more about the features of this reference architecture.
- How OCI supports demanding AI workloads with scalable GPU clusters and low-latency networking
- Access tutorials, workshops, and tools to build AI and generative AI solutions on OCI
- Oracle Cloud Infrastructure Documentation
- OCI Block Storage
- OCI Compute
- Well-architected framework for Oracle Cloud Infrastructure
- Oracle Cloud Cost Estimator
- Cloud Adoption Framework