A mobile game's backend runs on on-premises VMs exposing a REST API. Business logic uses multiple API keys and reads player session data from central file storage. Load varies and has peaks causing insufficient capacity and high latency when fetching session data. The API model must remain unchanged. Which cloud architecture meets needs for variable load handling and low-latency access?

This option provides a fully serverless and scalable solution. Amazon API Gateway handles the REST API, automatically scaling to meet variable load. AWS Lambda functions execute the business logic, also scaling on demand and eliminating the need for managing EC2 instances. Amazon DynamoDB with on-demand capacity offers low-latency access to player session data and automatically adjusts throughput based on usage, addressing the high latency and insufficient capacity issues of central file storage. The other options are less suitable: NLB and EC2 instances require manual scaling and management, which doesn't fully address variable load. Aurora Serverless is a good database but might not be the optimal choice for high-throughput, low-latency key-value access compared to DynamoDB for session data. ALB and Lambda functions are good for scaling, but ALB is typically used for HTTP/HTTPS traffic balancing to EC2 instances or containers, whereas API Gateway is specifically designed for managing, securing, and scaling APIs. AWS AppSync is for GraphQL APIs, not REST APIs, so it doesn't meet the requirement to keep the API model unchanged.