MLA-C01 Question 413
Single answerYou are deploying a machine learning model for real-time inference on AWS and want to optimize for latency and cost. The model has low computational requirements but needs to handle a high volume of requests with minimal latency. Which instance type would be most appropriate for this use case?
- A
Compute Optimized (C5) instances
- B
Inference Optimized (Inf1) instances
- C
Memory Optimized (R5) instances
- D
General Purpose (M5) instances
Show answer and explanation
Correct answer: B
Explanation
Inference Optimized (Inf1) instances are specifically designed for machine learning inference tasks. They leverage custom-built AWS Inferentia chips, which provide high throughput and low latency at a lower cost compared to other instance types. For real-time inference with low computational requirements, Inf1 instances are the most suitable choice.
- A. Incorrect.
Compute Optimized (C5) instances are designed for applications requiring high computational power, such as batch processing or scientific simulations, but they are not specifically optimized for inference tasks.
- B. Correct.
Inference Optimized (Inf1) instances are designed to handle machine learning inference workloads with low latency and high throughput, making them cost-effective for real-time inference use cases.
- C. Incorrect.
Memory Optimized (R5) instances are suited for workloads requiring large amounts of memory, such as big data analytics or in-memory databases, but they are not ideal for inference tasks with low computational requirements.
- D. Incorrect.
General Purpose (M5) instances offer a balanced mix of compute, memory, and networking, but they are not specialized for inference workloads and may not provide the optimal performance or cost efficiency for this use case.