Oracle Cloud Infrastructure 2022 Architect Professional — Question 42

A SaaS startup that hosts its application on-premises is experiencing rapid growth. Due to cost and flexibility considerations, they have decided to run some module of their application in the cloud with a multi-cloud approach.

The proposed solution architecture has the application entry point on-premises, where a load balancer redirects the request to the appropriate module and provides a failover mechanism. In each location, the module instances are exposed via a public IP.

You have been asked to review the above architecture and avoid any ‘single point of failure’.

How can you change your architecture to meet the above requirement?

Answer options

• A. Create a DNS zone in Oracle Cloud Infrastructure (OCI) for your application. Create a Traffic Management Steering Failover policy for each module. Add an Answer Group for each datacenter location with a type A answer to point to the module public IP. Adopt the policy domain as the entry point for each module.
• B. Create a DNS zone in Oracle Cloud Infrastructure (OCI) for your application. Create an A record for each module public IP. Create a Load Balancer Service instance with a Virtual Hostname associated with each A record domain name. Configure different weight to each backend server.
• C. Create a DNS zone in Oracle Cloud Infrastructure (OCI) for your application. Add a CNAME record for each module. Create a Load Balancer Service instance with the least connection policy. Configure the path routing to redirect the traffic based on the CNAME records.
• D. Create a Traffic Management Steering policy with IP Prefix Steering policy type for each module. Add an Answer Group for each datacenter location with a type A answer to point to the module public IP. Setup the rule to redirect the traffic to the appropriate datacenter location.

Correct answer: A

Explanation

Option A is correct because it implements a Traffic Management Steering Failover policy, which ensures that traffic is redirected to available modules in case of failure, thus avoiding a single point of failure. Options B and D do not specifically address failover mechanisms, and while they provide redundancy, they do not ensure automatic failover like option A. Option C utilizes CNAME records and a least connection policy, which may not adequately handle failover scenarios.