Which of the following solutions is MOST cost-effective?
Keep the website on T2 instances. Determine the minimum number of website instances required during off-peak times and use Spot Instances to cover them while using Reserved Instances to cover peak demand. Use Amazon EC2 R4 and Amazon EC2 R5 Reserved Instances in an Auto Scaling group for the video analysis application.
Keep the website on T2 instances. Determine the minimum number of website instances required during off-peak times and use Reserved Instances to cover them while using On-Demand Instances to cover peak demand. Use Spot Fleet for the video analysis application comprised of Amazon EC2 C4 and Amazon EC2 C5 Spot Instances.
Migrate the website to AWS Elastic Beanstalk and Amazon EC2 C4 instances. Determine the minimum number of website instances required during off-peak times and use On-Demand Instances to cover them while using Spot capacity to cover peak demand. Use Spot Fleet for the video analysis application comprised of C4 and Amazon EC2 C5 instances.
Migrate the website to AWS Elastic Beanstalk and Amazon EC2 R4 instances. Determine the minimum number of website instances required during off-peak times and use Reserved Instances to cover them while using On-Demand Instances to cover peak demand. Use Spot Fleet for the video analysis application comprised of R4 and Amazon EC2 R5 instances.
Explanations:
This option suggests using T2 instances for the website, which are not suitable for high-demand periods due to their performance limitations. While it proposes using Spot Instances and Reserved Instances for the video analysis, the fixed nature of T2 instances could lead to scalability issues and higher costs during peak demand.
This option recommends using T2 instances for the website and combining Reserved Instances for the base load with On-Demand Instances for peak demand. The use of Spot Fleet for the video analysis application allows for cost savings during sporadic, compute-intensive tasks while taking advantage of the lower prices of Spot Instances. This approach maximizes cost-effectiveness while accommodating demand fluctuations.
Migrating to Elastic Beanstalk and using C4 instances does improve performance and scalability, but it still suggests using On-Demand Instances for the base load, which is less cost-effective than using Reserved Instances. The use of Spot Fleet for video analysis is positive, but the overall cost savings would be diminished due to the reliance on On-Demand pricing.
While this option suggests using Elastic Beanstalk and R4 instances, which are more powerful and suitable for scaling, it still recommends using Reserved Instances for off-peak and On-Demand for peak demand. This approach may lead to higher costs than necessary, especially since Spot Instances are a more cost-effective solution for the video analysis application. Additionally, R4 instances may not be needed if the workload is sporadic.