A media streaming company is launching two new services on Amazon EC2. The first is an analytics application that must run continuously 24/7 with a highly predictable CPU utilization pattern over a three-year period. The second is a video transcoding microservice that processes user-uploaded media files; these jobs are stateless, store intermediate progress in Amazon S3, and can be interrupted or restarted without data loss. Which two Amazon EC2 pricing models should the company select to meet these requirements at the lowest cost?
- Reserved Instances to host the continuously running analytics applicationAnswer
- Spot Instances to run the video transcoding microserviceAnswer
- COn-Demand Instances to host the continuously running analytics application
- DSpot Instances to host the continuously running analytics application
- EDedicated Hosts to run the video transcoding microservice
Answer
The company should select Reserved Instances for the continuously running analytics application and Spot Instances for the video transcoding microservice.
For the analytics application, Reserved Instances are correct because they offer significant discounts for a committed three-year, steady-state, 24/7 workload. For the transcoding microservice, Spot Instances are correct because the workload is stateless and fault-tolerant, allowing the company to leverage spare EC2 capacity at the lowest cost.
Step-by-Step Solution
Key Concept
Selecting optimal Amazon EC2 pricing models based on workload predictability, duration, and fault tolerance.
Estimated Time:1m 30s