A biotechnology firm operates a genomics platform on AWS. The architecture consists of the… — AWS Certified Solutions Architect - Associate

A biotechnology firm operates a genomics platform on AWS. The architecture consists of the following components:

1. A web application portal hosted on Amazon ECS on AWS Fargate that runs continuously to provide researchers with search capabilities. The portal requires a predictable baseline of $4\text{ vCPU}$ and $16\text{ GB}$ of RAM at all times.
2. A batch processing system hosted on Amazon ECS on AWS Fargate that executes alignment algorithms on raw sequence files. These jobs are highly parallelized, stateless, can be interrupted without loss of progress due to checkpointing, and typically take $6$ to $8$ hours to complete.
3. An Amazon RDS for PostgreSQL database instance that stores metadata and query logs, running continuously $24/7$ .
4. An Amazon S3 bucket where raw sequence files are uploaded, processed by the alignment algorithm within $24$ hours, and then immediately deleted from S3.

Which combination of purchasing and storage strategies will provide the most cost-effective solution for these workloads?

A
Purchase a single Compute Savings Plan sized to cover both the baseline Fargate compute for the web application portal and the Amazon RDS PostgreSQL instance, run the batch processing system on Fargate Spot, and store the raw sequence files in Amazon S3 Standard.
Purchase a Compute Savings Plan to cover the baseline Fargate compute for the web application portal, run the batch processing system on Fargate Spot, purchase an Amazon RDS Reserved DB Instance for the database, and store the raw sequence files in Amazon S3 Standard.Answer
C
Migrate the batch processing jobs from Fargate to AWS Lambda, purchase a Compute Savings Plan to cover the web application portal and Lambda execution costs, purchase an Amazon RDS Reserved DB Instance for the database, and store the raw sequence files in Amazon S3 Standard.
D
Purchase a Compute Savings Plan to cover the baseline Fargate compute for the web application portal, run the batch processing system on Fargate Spot, purchase an Amazon RDS Reserved DB Instance for the database, and store the raw sequence files in Amazon S3 Standard-Infrequent Access (S3 Standard-IA).

Answer

Purchase a Compute Savings Plan to cover the baseline Fargate compute for the web application portal, run the batch processing system on Fargate Spot, purchase an Amazon RDS Reserved DB Instance for the database, and store the raw sequence files in Amazon S3 Standard.

Purchasing a Compute Savings Plan is the most cost-effective choice for the continuously running Fargate web portal because it offers significant discounts on Fargate compute for a committed usage. Since the batch processing jobs are stateless, checkpointed, and can tolerate interruptions, Fargate Spot is the most economical model for containerized batch workloads, offering up to a 90% discount compared to On-Demand prices. Amazon RDS has its own dedicated Reserved DB Instances, which provide substantial discounts for continuous 24/7 database workloads; Compute Savings Plans do not apply to RDS. Finally, storing raw sequence files that are processed and deleted within 24 hours in Amazon S3 Standard avoids the 30-day minimum storage charge penalty associated with S3 Standard-IA.

Step-by-Step Solution

Analyze the web application portal requirements.

Since it runs continuously with a predictable baseline of

4\text{ vCPU}

and

16\text{ GB}

of RAM, a Compute Savings Plan is the most cost-optimal purchasing model for Fargate, providing discounts up to

66\%

in exchange for a committed usage.

Web application compute runs continuously and requires a stable baseline.

Analyze the batch processing system requirements.

Since the jobs are stateless, parallelized, and checkpointed to tolerate interruptions, they can be run on Fargate Spot to save up to

90\%

compared to On-Demand pricing.

Spot instances/Fargate Spot are designed for fault-tolerant and interruptible workloads.

Analyze the database requirements.

Since the RDS PostgreSQL database runs continuously

24/7

, purchasing an RDS Reserved DB Instance is the correct optimization strategy. Compute Savings Plans do not cover RDS.

RDS databases are not eligible for Compute Savings Plans, so dedicated Reserved Instances must be used.

Analyze the S3 storage requirements.

Raw sequence files are processed and deleted within

24

hours. Amazon S3 Standard has no minimum storage duration and no retrieval charges, whereas S3 Standard-IA has a minimum

30

-day storage charge. Storing them in S3 Standard is more cost-effective.

Short-lived data (deleted within

30

days) incurs a cost penalty on S3 Standard-IA due to its minimum storage duration policy.

Key Concept

Selecting cost-optimized AWS compute purchasing models (Compute Savings Plans, Spot, Reserved Instances) and storage tiers based on workload patterns.

Estimated Time:2m 30s

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics