Q2 — AWS SAP-C02 Ch.2

Question 2 of 75 | ← Chapter 2

Q152. A solutions architect is redesigning a three-tier application that a company hosts on premises. The application provides personalized recommendations based on user profiles. The company already has an AWS account and has configured a VPC to host the application. The frontend is a Java-based application that runs in on-premises VMs. The company hosts a personalization model on a physical application server and uses TensorFlow to implement the model. The personalization model uses artificial intelligence and machine learning (AI/ML). The company stores user information in a Microsoft SQL Server database. The web application calls the personalization model, which reads the user profiles from the database and provides recommendations. The company wants to migrate the redesigned application to AWS. Which solution will meet this requirement with the LEAST operational overhead?

Correct Answer: B. Export the personalization model. Store the model artifacts in Amazon S3. Deploy the model to Amazon SageMaker and create an endpoint Host the Java application in AWS Elastic Beanstalk. Use AWS Database Migration Service (AWS DMS) to migrate the SQL Server database to Amazon RDS for SQL Serve

Explanation

To migrate the three-tier application with the least operational overhead, the recommended solution would be: B. Export the personalization model. Store the model artifacts in Amazon S3. Deploy the model to Amazon SageMaker and create an endpoint. Host the Java application in AWS Elastic Beanstalk. Use AWS Database Migration Service (AWS DMS) to migrate the SQL Server database to Amazon RDS for SQL Server. Explanation: Option B provides a solution that minimizes operational overhead while leveraging AWS services for scalability and managed services: 1. Personalization model: The personalization model is exported and stored in Amazon S3, which provides a highly durable and scalable storage solution for model artifacts. 2. Amazon SageMaker: The model is deployed to Amazon SageMaker, which is a fully managed service for building, training, and deploying machine learning models. SageMaker provides managed infrastructure and scalable endpoints for serving the model. 3. Java application: The Java application can be hosted in AWS Elastic Beanstalk, which is a fully managed service for deploying and scaling web applications. Elastic Beanstalk abstracts the underlying infrastructure, reducing operational overhead. 4. AWS Database Migration Service: The SQL Server database can be migrated to Amazon RDS for SQL Server using AWS DMS. This service simplifies the database migration process by handling schema conversion and data replication with minimal downtime. By using these AWS services, the company can take advantage of managed services that handle infrastructure provisioning, scalability, and operational tasks, reducing the operational overhead compared to managing on-premises infrastructure. Option A is incorrect because AWS Server Migration Service (SMS) is primarily used for migrating virtualized on-premises servers to AWS. It does not support migrating physical servers or the personalization model. Additionally, using AWS DMS to migrate the SQL Server database to Amazon RDS for SQL Server is a better approach for managing the database. Option C is incorrect because AWS Application Migration Service is not designed for migrating individual VMs or the personalization model. It is focused on application-level migration scenarios. Option D is incorrect because containerizing the personalization model and Java application using Amazon EKS would introduce additional complexity and management overhead. The solution would require managing and scaling the Kubernetes cluster, which is not necessary for this scenario. Therefore, option B is the recommended solution for migrating the three-tier application to AWS with the least operational overhead.