Lab 1: Deploying an AIFabric

This lab guides you through deploying an AIFabric resource in the MX-AI ecosystem, which includes setting up agents that can interact with each other and with services.

You will learn how to create an AIFabric, define agents, connect them to services and empower them with an LLM of your choice.

AIFabric Custom Resource

The following YAML snippet defines an AIFabric resource where the agents leverage OpenAI APIs to generate their responses.

myfabric.yaml

apiVersion: odin.trirematics.io/v1
kind: AIFabric
metadata:
  name: myfabric
  namespace: trirematics
spec:
  services:
    - name: ui
      image: hub.bubbleran.com/orama/services/ui:2025_06a1
    - name: observability-db
      image: hub.bubbleran.com/orama/services/observability-db:2025_06a1
  llms:
    - name: openai-model
      provider: openai
      model: gpt-4.1-mini
      api-key: <your-api-key>
  agents:
    - name: orchestrator-agent
      role: orchestrator
      image: hub.bubbleran.com/orama/agents/orchestrator:2025_06a1
      llm: openai-model
    - name: athena-agent
      role: worker
      image: hub.bubbleran.com/orama/agents/athena:2025_06a1
      llm: openai-model
      service-refs:
        - observability-db

In the above example, you may change the name of the services, llms and agents, but you should avoid changing the image field, as otherwise the odin-operator will not be able to find the correct Docker images to deploy.

To create the AIFabric, run:

brc install aifabric myfabric.yaml

Next, you can use the UI service to interact with the agents. To do so, SSH to the BubbleRAN cluster with local port forwarding on port 9900:

ssh -L 9900:localhost:9900 <your-user>@<your-cluster-ip>

Then, activate port forwarding for the UI service in Kubernetes:

kubectl -n trirematics port-forward service/myagent-ui 9900:9900

At this point, you can access the UI by navigating to http://localhost:9900 in your web browser.

Asking a question in the UI will trigger the orchestrator-agent, which will then delegate the task to the appropriate worker agents based on their capabilities. In this case, the only worker agent is the athena-agent, which will use the OpenAI API to generate a response.

The UI will display the steps taken by the agents while processing the request.

Example Questions

After deploying a Network and one or more Terminals, you can test the solutions with questions like:

• "What is the TDD configuration of the network?"
• "How many UEs are available?"
• "What is the IMSI of the UE named 'ue1'?"

In general, since the athena-agent is based on a RAG architecture, you should try to ask questions containing keywords that are present in the Network/Terminal configuration yaml files.