4 of the 6 questions we have for this objective, shown with the answer and the reasoning.
Determine a cost optimization strategy to meet solution goals and objectives
Aldermoor University gives each of 120 research groups a sandbox AWS account in an OU named Sandboxes. Each group may spend up to $500 per month. When a group's forecasted spend exceeds the limit, the group lead must be notified, and when actual spend reaches the limit, creation of new resources in that account must be blocked automatically, without the central team intervening.
Which solution meets these requirements with the LEAST operational overhead?
- ACreate an AWS Budgets cost budget for each sandbox account with a forecast-based alert to the group lead and an actual-spend threshold that triggers a budget action to apply a service control policy denying resource creation to the account.AWS Budgets supports alerts on forecasted and actual spend and budget actions that automatically apply an IAM policy or SCP when a threshold is reached, which enforces the limit without manual work.
- BCreate an Amazon CloudWatch billing alarm in each account that emails the group lead when estimated charges exceed $500.A billing alarm only notifies on actual estimated charges. It cannot alert on a forecast or block resource creation.
- CHave the central team review AWS Cost Explorer each week and suspend groups that exceed $500.Manual weekly reviews require central intervention and allow overspending between reviews.
- DEnable AWS Cost Anomaly Detection for each account and subscribe the group lead to anomaly alerts.Cost Anomaly Detection flags unusual spending patterns, not a fixed monthly limit, and it does not block new resources.
Why: Expenditure controls combine awareness, through alerts, with enforcement, through automated actions. AWS Budgets can alert when forecasted spend is expected to exceed a budget and can run budget actions that apply IAM policies or service control policies when actual spend crosses a threshold. This meets both the notification and the automatic blocking requirements without central intervention. Billing alarms, Cost Explorer and Cost Anomaly Detection improve visibility but do not enforce a limit.
Determine a cost optimization strategy to meet solution goals and objectives
Abernethy Genomics runs analysis jobs on Amazon EC2 instances in private subnets in us-east-1. Each month the jobs read about 60 TB from Amazon S3 buckets and write 10 TB of results to Amazon DynamoDB tables, all in us-east-1. All of this traffic flows through a NAT gateway, and the Cost and Usage Report shows NAT gateway data processing as the largest networking charge. The company wants to reduce this cost without making the subnets public.
Which two actions should the solutions architect take? Each correct answer presents part of the solution.
- ACreate an S3 gateway endpoint and add it to the route tables of the private subnets.Gateway endpoints for S3 have no hourly or data processing charges, and traffic to S3 in the same Region no longer passes through the NAT gateway.
- BCreate a DynamoDB gateway endpoint and add it to the route tables of the private subnets.DynamoDB also supports gateway endpoints at no additional charge, which removes the DynamoDB traffic from the NAT gateway.
- CCreate interface VPC endpoints for S3 and DynamoDB in every Availability Zone.Interface endpoints remove NAT processing but add hourly and per-GB charges. For same-Region access from inside the VPC, gateway endpoints are the lower-cost option.
- DDeploy one NAT gateway per Availability Zone instead of a single shared NAT gateway.Per-zone NAT gateways improve resilience and avoid cross-zone traffic but still charge data processing on every gigabyte, so the main cost remains.
- EEnable S3 Transfer Acceleration on the buckets.Transfer Acceleration speeds long-distance transfers over the AWS edge network for an additional per-GB fee. It does not bypass the NAT gateway and would raise cost.
Why: Data transfer modeling starts by identifying which path each byte takes and what each hop charges. Here the expensive hop is NAT gateway data processing on traffic that never needs the internet, because S3 and DynamoDB are in the same Region. Gateway VPC endpoints for S3 and DynamoDB route that traffic privately and carry no additional charge, which removes most of the NAT cost. Interface endpoints, extra NAT gateways and Transfer Acceleration solve different problems and add cost.
Determine a cost optimization strategy to meet solution goals and objectives
Linmouth Games distributes game patches of 2 to 8 GB directly from a public Amazon S3 bucket named patches-public in us-west-2 to players worldwide. Data transfer out to the internet is now 250 TB per month and is the largest line on the bill, and players in Asia report slow downloads. Most downloads in any week are for the same few patch files. The company also wants to stop serving the bucket publicly.
Which solution will reduce data transfer cost and improve download speed?
- AEnable S3 Transfer Acceleration on patches-public and give players the accelerated endpoint.Transfer Acceleration adds a per-GB fee on top of data transfer out and does not cache content, so repeat downloads of the same patches are charged in full every time.
- BServe the patches from Amazon EC2 instances behind a Network Load Balancer in us-west-2.Data transfer out from EC2 is billed at similar rates to S3, and the company would add instance and load balancer costs without improving latency for distant players.
- CReplicate patches-public to buckets in ap-northeast-1 and eu-west-1 with S3 Cross-Region Replication and direct players to the nearest bucket.Replication adds inter-Region transfer and storage costs, and players still download directly from S3 at internet data transfer out rates without caching.
- DCreate an Amazon CloudFront distribution with patches-public as the origin, restrict the bucket to the distribution with origin access control, and serve downloads through CloudFront.Data transfer from S3 to CloudFront is not charged, edge caching serves the popular patches close to players so they are fetched from the origin only once per edge cache, and CloudFront volume-tiered delivery pricing typically costs less at this scale than serving the same bytes directly from S3. Origin access control keeps the bucket private.
Why: For large, popular downloads, a content delivery network reduces both latency and cost. Amazon CloudFront caches objects at edge locations, so repeated requests are served without returning to the origin, and AWS does not charge for data transferred from S3 origins to CloudFront. CloudFront also offers volume-tiered pricing and, with origin access control, lets the bucket stay private. The distractors either add per-GB fees, add Regional copies without caching, or move the same egress to a more expensive tier.
Determine a cost optimization strategy to meet solution goals and objectives
Tarrant Logistics spends a steady $40,000 per month on compute: Amazon EC2 instances of several families in us-east-1 and eu-west-1, Amazon ECS tasks on AWS Fargate, and AWS Lambda functions. Over the next year the company plans to move most EC2 workloads from x86 to Graviton instances, shift some services from EC2 to Fargate, and possibly move a workload to ap-southeast-2. Finance wants to commit for 3 years to reduce cost but does not want the commitment to stop applying as the architecture changes.
Which pricing model should the solutions architect recommend?
- AStandard Reserved Instances for the current EC2 instance types in each Region.Standard RIs are tied to an instance family and Region and do not apply to Fargate or Lambda, so moving to Graviton, Fargate or another Region would strand the commitment.
- BEC2 Instance Savings Plans for each instance family currently in use.EC2 Instance Savings Plans give deeper discounts but apply to one instance family in one Region and do not cover Fargate or Lambda.
- CCompute Savings Plans with a 3-year hourly commitment sized to the steady baseline spend.Compute Savings Plans apply automatically to EC2 usage regardless of family, size, operating system or Region, and also to Fargate and Lambda. The discount continues as workloads change architecture, compute service or Region.
- DConvertible Reserved Instances that can be exchanged as instance types change.Convertible RIs can be exchanged, but exchanges are manual, are limited to EC2, and do not cover Fargate or Lambda usage.
Why: Commitment-based discounts trade flexibility for price, so the right choice depends on how much the workload will change. Compute Savings Plans offer the most flexibility: a dollar-per-hour commitment that applies across EC2 families, sizes and Regions, as well as Fargate and Lambda. EC2 Instance Savings Plans and Standard RIs give larger discounts but only within a family and Region. Because the company expects processor, service and Region changes, flexibility is the deciding constraint, and sizing the commitment to the steady baseline avoids paying for unused commitment.