A lift-and-shift migration left a single Auto Scaling group where each EC2 instance runs web, database, and Redis cache components, causing variable response times and overloaded instances. The company wants to separate components to improve availability and performance. Which architecture change will satisfy this requirement?

The correct option effectively separates the components and leverages appropriate AWS services. An Application Load Balancer (ALB) is ideal for HTTP/HTTPS traffic for the web application, and an Auto Scaling group ensures scalability and high availability. Migrating the database to Amazon Aurora with Multi-AZ provides a highly available, scalable, and durable relational database solution, addressing the previous single-instance bottleneck. Creating an Amazon ElastiCache (Redis OSS) cluster is the best practice for a managed, scalable, and high-performance Redis cache, eliminating the need to run Redis on EC2 instances. Incorrect options: Using a Network Load Balancer (NLB) for the web application is less suitable than an ALB, as ALBs offer more advanced routing and HTTP-specific features. Using an ALB for Redis is also inappropriate; Redis is typically accessed directly or through ElastiCache. Using an NLB for the web application is suboptimal. Creating an Auto Scaling group for Redis in a single Availability Zone defeats the purpose of high availability and is less efficient than ElastiCache. Using an NLB for the web application is suboptimal. Creating a target group with DNS targets pointing to an ElastiCache hostname is an unnecessary and complex workaround; applications should connect directly to ElastiCache endpoints.