Free Amazon ANS-C01 Actual Exam Questions - Question 4 Discussion
pressure, and location data through the MQTT messaging protocol. The hardware sensor modules
send this data to the company's on-premises MQTT brokers that run on Linux servers behind a load
balancer. The hardware sensor modules have been hardcoded with public IP addresses to reach the
brokers.
The company is growing and is acquiring customers across the world. The existing solution can no
longer scale and is introducing additional latency because of the company's global presence. As a
result, the company decides to migrate its entire infrastructure from on premises to the AWS Cloud.
The company needs to migrate without reconfiguring the hardware sensor modules that are already
deployed across the world. The solution also must minimize latency.
The company migrates the MQTT brokers to run on Amazon EC2 instances.
What should the company do next to meet these requirements?
Maybe B. Using Global Accelerator in front of the NLB can help direct traffic to the closest AWS region, reducing latency for global users. Since the sensor modules are hardcoded with public IPs, BYOIP paired with Global Accelerator lets them keep those IPs while benefiting from faster global routing. A only uses an NLB without Global Accelerator, so it might not handle worldwide latency as well. ALB (C) isn’t great for MQTT because MQTT uses TCP, not HTTP. CloudFront (D) is mostly for web content, so it doesn’t fit MQTT use cases here.
It’s A because using an NLB with TCP listeners supports MQTT directly, and BYOIP lets you keep the same IP addresses without reconfiguring hardware. Global Accelerator isn’t needed if latency is managed by keeping IPs consistent.
If the company doesn’t own those IP addresses, BYOIP isn’t an option, so A and B might not work. Also, CloudFront (D) is mainly for HTTP/HTTPS, so it seems off for MQTT traffic.
B vs A? Global Accelerator could reduce latency for worldwide customers.
B seems right, with Global Accelerator to reduce latency globally.