Below question answer rationale need to understand.
The marketing department at your company has been using a homegrown approach to content management for the company’s customer facing website. Recent content publishing issues have resulted in a directive to move to Drupal for all content management going for ward. The IT budget is already stretched, and there is no capital available to purchase new servers, so the Drupal solution will be implemented on AWS. The current on-premises content management system runs on six Linux servers, each with four CPUs and sixteen gigabytes of memory. There are also two MySQL database servers, each with four CPUs and thirty-two gigabytes of memory. Spikes in volume are common during the fourth quarter of the year with increases up to fifty percent during these months. Which architecture will provide the most cost effective solution?
Answer : Purchase four EC2 m5.xlarge Reserved Instances on a 3-year term. Configure Auto Scaling with Drupal AMI launch configurations. Implement one Amazon Aurora db.r5.large instance and one Amazon ElastiCache Memcached r5.2xlarge instance.
Given Answer Explanation :
Since the company has deployed six on-premises content management servers to handle seasonal spikes in volume up to fifty percent, four servers will be needed during most of the year. So we should purchase four EC2 Reserved Instances, and have Auto Scaling add up to an additional two when needed. Since content management systems are mostly read-intensive, we can save on the sizing and cost of our Aurora nodes, and improve performance by implementing larger ElastiCache nodes. The rich feature set of Redis is not needed for this use case, so we’ll go with Memcached for simplicity. Running Drupal on AWS Amazon EC2 Reserved Instances Pricing Amazon Aurora Pricing Amazon ElastiCache Pricing.
My Understanding & Question :
1 Linux Servers 6 Servers 4 cpu 16GB
2 Database : 2 Servers 4 cpu 32GB
50 % spike in volume means –
Linux 6 Servers -> 3 Servers. But answer in 4 servers RI and 2 on-demand, Aurora DB, Elastic Cache
But explanation is (handle seasonal spikes in volume up to fifty percent, four servers will be needed during most of the year.)
I need your help to explain how the calculation of reserved instances has made ?