Amazon EC2 Auto Scaling vs Amazon Elastic Inference comparison

Amazon EC2 Auto Scaling helps you maintain application availability and allows you to automatically add or remove EC2 instances according to conditions you define. ... Dynamic scaling responds to changing demand and predictive scaling automatically schedules the right number of EC2 instances based on predicted demand.

Amazon Elastic Inference allows you to attach low-cost GPU-powered acceleration to Amazon EC2 and Amazon SageMaker instances or Amazon ECS tasks to reduce the cost of running deep learning inference by up to 75%. Amazon Elastic Inference supports TensorFlow, Apache MXNet, and ONNX models, with more frameworks coming soon.
In most deep learning applications, making predictions using a trained model—a process called inference—can drive as much as 90% of the compute costs of the application due to two factors. First, standalone GPU instances are designed for model training and are typically oversized for inference. While training jobs batch process hundreds of data samples in parallel, most inference happens on a single input in real time that consumes only a small amount of GPU compute. Even at peak load, a GPU's compute capacity may not be fully utilized, which is wasteful and costly. Second, different models need different amounts of GPU, CPU, and memory resources. Selecting a GPU instance type that is big enough to satisfy the requirements of the most demanding resource often results in under-utilization of the other resources and high costs.
Amazon Elastic Inference solves these problems by allowing you to attach just the right amount of GPU-powered inference acceleration to any EC2 or SageMaker instance type or ECS task with no code changes. With Amazon Elastic Inference, you can now choose the instance type that is best suited to the overall CPU and memory needs of your application, and then separately configure the amount of inference acceleration that you need to use resources efficiently and to reduce the cost of running inference.