What happened:
The cloud-edge collaborative inference for LLM example fails on Windows during runtime execution. The pipeline crashes while loading the EdgeModel due to a missing vllm dependency.

What you expected to happen:
The example should either run successfully on Windows or clearly document Linux-only dependencies and fail early with a clear error message.

How to reproduce it:

git clone https://github.com/kubeedge/ianvs.git
cd ianvs
python -m venv ianvs-env
ianvs-env\Scripts\activate

pip install -r requirements.txt
python setup.py install

mkdir workspace
ianvs -f examples\cloud-edge-collaborative-inference-for-llm\benchmarkingjob.yaml

Anything else we need to know?

OS: Windows 11
Python: 3.10.11
CPU-only machine

Error: ModuleNotFoundError: No module named 'vllm'

Source: examples/cloud-edge-collaborative-inference-for-llm/testalgorithms/query-routing/models/vllm_llm.py

vllm does not support Windows and is not documented as a required dependency

Proposed Fix

Fix 1 (Recommended & Realistic): Run the example on Linux

• vllm is a hard dependency for the EdgeModel used in this example.
• vllm currently does not support Windows.
• Running the example on Linux (native Linux, WSL2, VM, or cloud VM) allows successful installation and execution.

Suggested environments:

• Ubuntu 20.04 / 22.04
• WSL2 (Ubuntu) on Windows
• Cloud VM (AWS / GCP / Azure)

Fix 2 (Documentation Improvement): Clearly document OS constraints

• Explicitly mention in the example README that:

  • vllm is required
  • The example is Linux-only

• Add a note in Quick Start or example README:

  "This example requires vllm and currently does not support Windows."

Fix 3 (Optional, Future Improvement): Provide a fallback EdgeModel

• Allow disabling vllm via config
• Provide a HuggingFace-only / CPU-compatible EdgeModel
• This would improve accessibility for Windows and CPU-only users