Question #971
A developer has AWS Lambda functions that require access to dynamically updated machine learning models and training datasets. Separate teams manage the model development and dataset curation. Both teams must be able to modify and deploy new versions of their respective assets independently. The Lambda functions operate within the company's central VPC.
Which solution will provide the Lambda functions with consistent access to the latest models and datasets while maintaining team autonomy?
Attach an Amazon Elastic Block Store (Amazon EBS) volume to the Lambda functions using EBS Multi-Attach in the central VPC. Grant the Lambda execution roles access to the EBS volume. Modify the Lambda code to read assets directly from the mounted volume.
Package the models and datasets in the Lambda deployment package's /tmp directory. Update the Lambda code to load assets from the /tmp directory during execution.
Create an Amazon Elastic File System (Amazon EFS) file system with mount targets in the central VPC. Configure the Lambda functions to mount the EFS file system. Grant the Lambda execution roles permissions to access the EFS resources.
Deploy an Amazon FSx for Lustre file system with VPC mount targets. Configure Lambda functions to mount the FSx file system. Update Lambda execution roles to permit access to the FSx resources.
Explanation
Option C is correct because Amazon EFS (Elastic File System) is a managed, scalable network file system that supports concurrent access and dynamic updates. By mounting EFS in the central VPC, Lambda functions can access the latest models and datasets stored in EFS without requiring redeployment. Separate teams can independently update their respective assets (models and datasets) in EFS, maintaining autonomy. EFS also integrates with IAM roles, allowing granular permissions for each team.
Other options are incorrect because:
- A: EBS volumes cannot be directly attached to Lambda functions and lack shared access capabilities.
- B: Storing assets in Lambda's /tmp directory is ephemeral and requires redeployment for updates, violating team autonomy.
- D: FSx for Lustre is optimized for high-performance computing, which is unnecessary here, and adds complexity/cost compared to EFS.
Key Points:
1. EFS enables shared, dynamic storage accessible to Lambda via VPC.
2. Teams can update assets independently without redeploying Lambda.
3. IAM roles enforce access control for team autonomy.
Answer
The correct answer is: C