Horizontal vs Vertical Scaling Quiz

What you'll be asked

1. Which of the following best describes 'Horizontal Scaling' in the context of GenAI inference systems?
2. When serving a Large Language Model (LLM) that exceeds the VRAM capacity of a single GPU, which scaling approach is most appropriate?
3. Which scenario provides the strongest justification for choosing Horizontal Scaling over Vertical Scaling?
4. What is a primary drawback of relying solely on Vertical Scaling for a GenAI application?
5. In a distributed inference architecture using Kubernetes, if you implement a Horizontal Pod Autoscaler (HPA) based on GPU utilization, what is the most critical requirement for the system to remain performant?

Answers and explanations come after you submit.