A customer has a website which shows all the sales available in leading online and brick-and-mortar stores.
Currently, the website’s web, application, and database tiers are hosted on five large, on-demand EC2 instances to manage day-to-day traffic. However, from late November through the end of December, the website traffic quadruples the normal rate for various periods, sometimes a few minutes, and sometimes a few hours. During the peaks in activity, the website requires up to 20 large instances.
The level of activity is not predictable. It is also prone to large, sudden spikes at random times based on when various temporary sales are available. The site’s traffic can quadruple in a matter of minutes.
Which option is the most cost-effective and achieves better performance to handle these peaks in traffic reliably?
Explanation
AWS provides an on-demand, scalable infrastructure. Amazon EC2 allows the user to launch On-Demand instances and the organization should create an AMI of the running instance. When the organization is experiencing varying loads and the time of the load is not known but it is higher than the routine traffic it is recommended that the organization launches a few instances beforehand and then setups AutoScaling with policies which scale up and down as per the EC2 metrics, such as Network I/O or CPU utilization.
If the organization keeps all ten additional instances as a part of the AutoScaling policy sometimes during a sudden higher load, it may take time to launch instances and may not give optimal performance. This is the reason it is recommended that the organization keeps an additional five instances running and the next five instances scheduled as per the AutoScaling policy for cost-effectiveness.
Additionally, the website has established a baseline of normal workload outside of the peak season, so they can invest in Standard reserved instances to manage this workload and save dramatically on compute resource costs.
Explanation
AWS provides an on-demand, scalable infrastructure. Amazon EC2 allows the user to launch On-Demand instances and the organization should create an AMI of the running instance. When the organization is experiencing varying loads and the time of the load is not known but it is higher than the routine traffic it is recommended that the organization launches a few instances beforehand and then setups AutoScaling with policies which scale up and down as per the EC2 metrics, such as Network I/O or CPU utilization.
If the organization keeps all ten additional instances as a part of the AutoScaling policy sometimes during a sudden higher load, it may take time to launch instances and may not give optimal performance. This is the reason it is recommended that the organization keeps an additional five instances running and the next five instances scheduled as per the AutoScaling policy for cost-effectiveness.
Additionally, the website has established a baseline of normal workload outside of the peak season, so they can invest in Standard reserved instances to manage this workload and save dramatically on compute resource costs.