L1, L2, and L3 Constructs in AWS CDK - krdheeraj51/aws-labs GitHub Wiki

Introduction

AWS Cloud Development Kit (CDK) provides three levels of constructs:
L1 (Low-Level Constructs), L2 (High-Level Constructs), and L3 (Patterns).
These constructs define different abstraction levels for AWS infrastructure provisioning using CDK.

---

L1 Constructs (Low-Level Constructs)

L1 constructs are direct CloudFormation resource representations in AWS CDK.
They provide a low-level abstraction and require detailed configuration.

Characteristics:

• Directly maps to AWS CloudFormation resources.
• Uses Cfn prefix (e.g., CfnBucket for S3).
• Requires explicit definitions for properties.

Example:

from aws_cdk import core
from aws_cdk import aws_s3 as s3

class MyStack(core.Stack):
    def __init__(self, scope: core.Construct, id: str, **kwargs):
        super().__init__(scope, id, **kwargs)

        bucket = s3.CfnBucket(self, "MyL1Bucket",
            bucket_name="my-l1-bucket",
            versioning_configuration={"status": "Enabled"}
        )

• L1 constructs are low-level constructs that directly map to AWS CloudFormation resources.

Characteristics:

• L1 is more verbose: Requires explicit property definitions.
• Example: A VPC in L1 may require 282 lines of code.

---

L2 Constructs (High-Level Constructs)

L2 constructs offer a higher-level abstraction over L1 constructs, providing intelligent defaults to simplify resource creation.

Characteristics:

• Simplified API with sensible defaults.
• More concise and easier to manage than L1.
• Reduces complexity in defining AWS resources.

Example:

from aws_cdk import aws_s3 as s3

class MyStack(core.Stack):
    def __init__(self, scope: core.Construct, id: str, **kwargs):
        super().__init__(scope, id, **kwargs)

        bucket = s3.Bucket(self, "MyL2Bucket",
            bucket_name="my-l2-bucket",
            versioned=True
        )

• Same S3 bucket, but with a much simpler syntax.
• More concise: The same VPC in L2 may need just 3 lines of code instead of 282.

L3 Constructs (Patterns)

• L3 constructs, also called patterns, are predefined solutions for common use cases. They combine multiple L2 constructs into ready-to-use architecture patterns.

Characteristics:

• Combines multiple L2 constructs into one pattern.
• Even higher-level abstraction than L2.
• Useful for common AWS architectural patterns.

Same S3 bucket, but with a much simpler syntax

• More concise: The same VPC in L2 may need just 3 lines of code instead of 282.

---

L3 Constructs (Patterns)

L3 constructs, also called patterns, are predefined solutions for common use cases. They combine multiple L2 constructs into ready-to-use architecture patterns.

Characteristics:

• Combines multiple L2 constructs into one pattern.
• Even higher-level abstraction than L2.
• Useful for common AWS architectural patterns.

Example:

from aws_cdk import aws_apigateway as apigateway
from aws_cdk import aws_lambda as lambda_

class MyStack(core.Stack):
    def __init__(self, scope: core.Construct, id: str, **kwargs):
        super().__init__(scope, id, **kwargs)

        handler = lambda_.Function(self, "MyLambda",
            runtime=lambda_.Runtime.PYTHON_3_9,
            code=lambda_.Code.from_asset("lambda"),
            handler="index.handler"
        )

        api = apigateway.LambdaRestApi(self, "MyApi",
            handler=handler
        )

• Preconfigured API Gateway with Lambda.
• Combines multiple resources (API Gateway, Lambda, IAM Roles).
• Other examples:
  • aws_cdk.aws_ecs_patterns.ApplicationLoadBalancedFargateService
  • aws_cdk.aws_apigateway.LambdaRestAPI

Key Differences Between L1, L2, and L3 Constructs

Feature	L1 Constructs	L2 Constructs	L3 Constructs
Abstraction Level	Low	Medium	High
Mapping	1:1 with CloudFormation	CloudFormation with built-in defaults	Predefined patterns for use cases
Naming Convention	Uses Cfn prefix (e.g., CfnBucket)	Uses direct AWS SDK-like naming (e.g., Bucket)	Pattern-based (e.g., LambdaRestAPI)
Code Complexity	High (detailed configuration required)	Medium (intelligent defaults)	Low (fully managed patterns)
Use Case	Full control over AWS resources	Simplified configuration with defaults	Pre-built solutions for common patterns

---

When to Use Each Construct?

• L1 Constructs: When you need full control and a 1:1 CloudFormation mapping.
• L2 Constructs: When you want simplicity with best-practice defaults.
• L3 Constructs: When you need predefined architectures for common workloads.

---

Conclusion

Understanding L1, L2, and L3 constructs helps in choosing the right abstraction level for AWS CDK projects.

• L1 offers full control but is verbose.
• L2 simplifies configurations with defaults.
• L3 provides reusable patterns to speed up deployments.

By using the appropriate construct level, developers can reduce complexity, improve maintainability, and accelerate cloud infrastructure deployment.