diegoecab/oci-deal-accelerator
1
1
Fork 2
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
4552ef7226acb2092eb79b7861bcd3b6a7630ee3
oci-deal-accelerator/kb/diagram/assets/archcenter-refs/oci-generative-ai-llm-platforms/_description.md
root b30a4f0d32 Diagram generation: ref-arch-driven procedure + spec validator + KB enrichment 
The diagram path now follows a documented standard procedure (lookup
the closest Oracle Architecture Center reference → confirm components
→ author absolute_layout → spec validator → render → visually verify)
and ships persistent guardrails so layout regressions can't recur.

Persistent procedure changes (apply to all users, all sessions):
- tools/diagram_spec_validator.py — geometry checks (CONTAINER_TOO_THIN,
  CONTAINER_PADDING_VIOLATION, LABEL_OVERFLOW_PARENT) run BEFORE either
  renderer (drawio + PPTX). Catches the subnet-collapse / label-overflow
  bugs that the post-render drawio validator missed.
- tools/oci_diagram_gen.py + tools/oci_pptx_diagram_gen.py — call the
  spec validator before emitting any output. Adds mysql / mysql_heatwave
  type aliases.
- tools/archcenter_pattern_lookup.py — scores against cached page
  descriptions (not just the 1-line summary), supports --queries for
  multi-fragment composition, and applies synonym expansion via
  kb/architecture-center/synonyms.yaml so "LB HA cross AD" matches
  "load balancer high availability availability domain".
- kb/architecture-center/synonyms.yaml — canonical synonym table
  (load balancer, autonomous database, data guard, …) used by the
  lookup scorer.

KB enrichment:
- tools/archcenter_description_fetcher.py + 121 cached _description.md
  under kb/diagram/assets/archcenter-refs/<slug>/. Removes the runtime
  dependency on docs.oracle.com when authoring specs and feeds the
  pattern-lookup scorer.
- 110+ cached .drawio / .svg / .png references for offline reuse,
  plus the OCI Toolkit v24.2 import (kb/diagram/assets/oci-toolkit-drawio).

Documentation:
- docs/skill/output-formats.md — new "Standard diagram-generation
  procedure (MANDATORY)" + geometry rules + the new validator entry.
- SKILL.md option 2 — references the mandatory procedure.
- README.md — describes the spec validator, archcenter_pattern_lookup
  and description fetcher, and updates the KB-health table.

Tooling that backs the procedure (cumulative across recent sessions):
tools/archcenter_case_runner.py, archcenter_batch_driver.py,
archcenter_zip_downloader.py, drawio_visual_validator.py,
drawio_fidelity_eval.py, harvest_drawio_icon.py, import_oci_library.py,
oci_pptx_diagram_gen.py, oci_pptx_render.py, refresh_pptx_icon_index.py.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 21:15:21 -03:00

8.4 KiB
Raw Blame History

Enable secure and scalable self-service platforms for generative AI and LLMs within OCI

Summary (catalog)

Enterprise GenAI platform with AI CoE governance. NVIDIA MIG for fractional GPU allocation, OCI DevOps for CI/CD, Oracle Database 23ai for vector storage. IAM-based environment segmentation per team.

Architecture (fetched from source)

Architecture

This architecture illustrates how Oracle Cloud Infrastructure (OCI) supports end-to-end generative AI workflows across development, integration, and user interaction.

Flow A: Integration

  • Customer applications

  • Oracle Integration

  • OCI Object Storage (buckets)

  • OCI Events detection

  • OCI Streaming and OCI Connector Hub

  • OCI Functions (logic execution)

  • Oracle Process Cloud Service (inference by GPUs)

  • Data layer ( Oracle AI Database 23ai and buckets)

Flow B: User interaction

  • End-user interfaces (Apex)

  • Applications ( OCI GenAI Agents , OCI Speech , Oracle Digital Assistant )

  • Oracle Process Cloud Service (inference by GPUs)

  • Data layer ( Oracle AI Database 23ai and buckets)

Flow C: Development and sandbox

  • External model sources

  • Code security validation

  • Development and testing

  • Automation pipeline to production

The following diagram illustrates this reference architecture.

Description of the illustration ai-llm-workflow-architecture.png

ai-llm-workflow-architecture-oracle.zip

Architecture overview by functional domains

  • Development and training (self-service workspace)

The architecture is structured under a centralized compartment for LLM operations:

  • Data Science provides an integrated workspace for model development, Jupyter notebooks, and pre-built ML frameworks. Includes quick action tools for model deployment and job execution.

  • Model deployment hosts virtual machines (VMs) for model testing and deployment. Users can validate models here before moving them into production.

  • Playground is a GPU-accelerated environment (Flex VMs, A10, A100, LS40) offering isolated and high-performance compute resources for custom and third-party models (for example, Hugging Face). It serves as the experimentation zone for Bring Your Own LLM (BYOLLM) workflows.

  • Application and function layer

  • OCI Speech and language APIs offer ready-to-consume services for transcription, NLU, and entity extraction.

  • OCI Functions is used for real-time transcription, NLP, and serverless execution of AI pipelines.

  • APEX front-end and monitoring tools provide interfaces for user interaction, analytics, and governance.

  • OCI GenAI Agents and Digital Assistant enable conversational experiences using enterprise data and integrated LLMs.

  • Processing (production layer)

  • OCI Kubernetes Engine (OKE) supports containerized deployment of production models and inference services.

  • OCI Generative AI provides API-based access to Oracle-hosted or custom, fine-tuned LLMs, supporting secure and scalable enterprise use cases.

  • GPU infrastructure (H100 and RDMA support)

  • Bare metal GPU instances (H100 with RDMA) enable multi-node, distributed training and inference with high-throughput, low-latency communication, ideal for massive LLM workloads.

  • Optimized for Kubernetes and NVIDIA Multi-Instance GPU (MIG) technology, this setup enables GPU orchestration and dynamic resource sharing, allowing fractional GPU allocation and multi-user scheduling across teams.

  • Data and knowledge layer

  • Oracle AI Database 23ai, enhanced with support for vector and semantic search, acts as the retrieval layer for Retrieval-Augmented Generation (RAG) workflows.

  • OCI Object Storage buckets store unstructured data, embeddings, documents, and model artifacts.

  • MLOps (production model pipeline)

  • The architecture includes a CI/CD pipeline for promoting models from the playground environment to production. Currently represented by OCI DevOps is OCI's native, fully-managed, continuous integration and continuous delivery (CI/CD) service that enables organizations to automate the deployment of machine learning models from experimentation to production.

  • Integrated build pipelines with Git.

  • Automated deployment to VMs or containers.

  • Native integration with OCI Artifacts Registry , OCI Functions , and OCI API Gateway .

  • Integration and security layer

  • OCI Object Storage buckets act as the central storage for models, training data, inference outputs, and embeddings.

  • OCI Events , OCI Streaming , and OCI Connector Hub enable event-driven orchestration and service integration across the environment.

  • Oracle Identity Cloud Service , IAM policies, OCI Logging , and security lists provide robust governance, authentication, access control, and compliance capabilities across all OCI services.

  • Oracle Integration is a pre-built middleware platform that enables secure and seamless integration between on-premises systems and cloud services, supporting real-time data synchronization, API orchestration, and process automation across heterogeneous applications.

The architecture has the following components:

  • Availability domains Availability domains are standalone, independent data centers within a region. The physical resources in each availability domain are isolated from the resources in the other availability domains, which provides fault tolerance. Availability domains don’t share infrastructure such as power or cooling, or the internal availability domain network. So, a failure at one availability domain shouldn't affect the other availability domains in the region.

  • Bare metal Oracle’s bare metal servers provide isolation, visibility, and control by using dedicated compute instances. The servers support applications that require high core counts, large amounts of memory, and high bandwidth. They can scale up to 192 cores, 2.3 TB of RAM, and up to 1 PB of block storage. Customers can build cloud environments on Oracle’s bare metal servers with significant performance improvements over other public clouds and on-premises data centers.

  • Compartment Compartments are cross-regional logical partitions within an OCI tenancy. Use compartments to organize, control access, and set usage quotas for your Oracle Cloud resources. In a given compartment, you define policies that control access and set privileges for resources.

  • OCI Connector Hub Oracle Cloud Infrastructure Connector Hub is a message bus platform that orchestrates data movement between services on OCI. You can use connectors to move data from a source service to a target service. Connectors also enable you to optionally specify a task (such as a function) to perform on the data before it is delivered to the target service.

You can use OCI Connector Hub to quickly build a logging aggregation framework for security information and event management (SIEM) systems.

  • Dynamic routing gateway (DRG) The DRG is a virtual router that provides a path for private network traffic between VCNs in the same region, between a VCN and a network outside the region, such as a VCN in another OCI region, an on-premises network, or a network in another cloud provider.

  • OCI FastConnect Oracle Cloud Infrastructure FastConnect creates a dedicated, private connection between your data center and OCI. FastConnect provides higher-bandwidth options and a more reliable networking experience when compared with internet-based connections.

  • High-performance computing High-performance computing is designed for workloads that require cluster networking and high-speed processor cores for massively parallel workloads.

  • Internet gateway An internet gateway allows traffic between the public subnets in a VCN and the public internet.

  • On-premises network This is a local network used by your organization.

  • Region An OCI region is a localized geographic area that contains one or more data centers, hosting availability domains. Regions are independent of other regions, and vast distances can separate them (across countries or even continents).

  • Route table Virtual route tables

Reference in New Issue View Git Blame Copy Permalink