Solutions to Optimize AWS Costs without Service Impact

Correct Answer - A, E.

AWS documentation in https://docs.aws.amazon.com/aws-technical-content/latest/cost-optimization-laying-the-foundation/cost-optimization-pillars.html has introduced the pillars for Cost Optimization.

Please check this document for related details.

Option A is CORRECT: Because this is a way to dig out existing CloudWatch data and get recommendations for the right size of EC2 instances.

Please refer to https://aws.amazon.com/compute-optimizer/ for the implementation.

You can take advantage of the recommendations in Compute Optimizer to reduce costs by up to 25%

Compute Optimizer analyzes the configuration and resource utilization of a workload to identify AWS compute resources, such as Amazon EC2 instances, that might be under-provisioned or over-provisioned.

Compute Optimizer then recommends you to right-size your AWS compute resources and switch to a different instance size or instance type to save costs.

Option B is incorrect: Because Direct Connect may cost much more than VPN connections.

It also depends on which kind of services are used before deciding on Direct Connect or VPN.

However, in most cases, Direct Connect is not cost-efficient.

Option C is incorrect: Because HDD (st1) is actually cheaper than SSD (gp2)

So using SSD is not cost-efficient.

Option D is incorrect: Because spot instances cannot ensure the service.

So they are not recommended for production environments.

Option E is CORRECT: Because unused EC2 instances can be stopped to save cost.

A script or a pipeline can automatically execute this task to stop and start the instances appropriately.

As a consulting company, you can recommend the following two approaches to optimize the costs for your customer without impacting the service:

A. Use Compute Optimizer to choose the optimal Amazon EC2 instance types based on your utilization data:

Compute Optimizer is an AWS service that helps optimize the costs and performance of Amazon EC2 instances. It analyzes the utilization data of EC2 instances and provides recommendations on the optimal instance types that best fit your workloads, which can help you reduce the costs without sacrificing the performance.

By leveraging Compute Optimizer, you can identify the underutilized or overutilized EC2 instances, and switch them to the appropriate instance types to match the workload demands and reduce the overall costs.

B. To save the cost of the VPN connections between VPC and on-premises servers, replace the VPN connections with a Direct Connect that is cheaper and also more reliable:

Virtual Private Network ( VPN) connections between VPC and on-premises servers provide secure communication over the internet, but the cost can be high, especially if the traffic volume is large.

AWS Direct Connect is a dedicated network connection between AWS and on-premises data centers that provides a faster, more reliable, and cost-effective way to access AWS services. By using Direct Connect, you can bypass the public internet and connect directly to AWS resources, which can reduce the data transfer costs and improve the network performance.

By replacing VPN connections with Direct Connect, you can reduce the costs of data transfer, which can be particularly significant if your customer's workload involves large data volumes.

C. For non-production environments, use the default General Purpose SSD (gp2) instead of EBS Throughput Optimized HDD (st1) storage as SSD costs half the price:

Amazon Elastic Block Store (EBS) provides persistent block-level storage volumes for Amazon EC2 instances. EBS volumes come in different types and sizes, each with different performance characteristics and costs.

General Purpose SSD (gp2) and Throughput Optimized HDD (st1) are two common EBS volume types. Gp2 provides a balance of price and performance, while st1 is designed for high throughput workloads and offers lower costs per GB.

For non-production environments, where performance is not critical, you can save costs by using the default gp2 storage, which costs about half as much as st1 storage.

D. To save cost, use spot instances instead of on-demand instances for Jenkins that manages the production CI/CD pipelines:

Spot instances are unused EC2 instances that are available for lower prices than on-demand instances. AWS provides these instances through the Spot market, which enables users to bid on spare EC2 capacity.

For Jenkins instances that manage production CI/CD pipelines, you can save costs by using spot instances instead of on-demand instances. Since Jenkins instances usually perform non-critical workloads, you can bid on spot instances to achieve significant cost savings while still ensuring the required capacity is available.

E. Turn off certain non-production instances during weekends when no one uses them and turn on these instances on Monday morning:

Many customers have non-production instances that are only used during working hours. By turning off these instances during weekends, you can reduce the running time and save costs.

AWS provides various automation tools that allow you to schedule the start and stop times of your instances, such as AWS Lambda and Amazon CloudWatch Events. You can leverage these tools to automate the start and stop of your non-production instances during weekends, which can significantly reduce costs without impacting the service.

In summary, by recommending the above approaches, you can help your customer optimize the costs of AWS usage while ensuring the service is not impacted.