프라이빗 리소스를 위한 VPC 내 Lambda

By default, AWS Lambda functions run in a network managed by AWS. This network provides internet access but isolates your function from your private AWS resources.

To securely access resources like databases (e.g., Amazon RDS, DynamoDB tables in a VPC) or internal services that are not publicly available, your Lambda function needs to operate within your own Virtual Private Cloud (VPC).

A Virtual Private Cloud (VPC) is like your own isolated, private network within AWS. You define its IP address range, subnets, and network gateways.

• Subnets: Divisions within your VPC. Private subnets host resources that shouldn't be publicly accessible.
• Security Groups: Act as virtual firewalls, controlling inbound and outbound traffic for resources within your VPC.

Imagine you have a database that holds sensitive customer data. You wouldn't want it exposed to the public internet.

If your Lambda function needs to read from or write to such a database, or connect to an EC2 instance, or an internal API endpoint that resides only within your VPC, then your Lambda function must also be placed inside that VPC.

When you configure a Lambda function to run within a VPC, AWS creates an Elastic Network Interface (ENI) for your function in the specified subnets.

This ENI provides your Lambda function with a private IP address and allows it to communicate with other resources in your VPC, just like an EC2 instance would.

When attaching Lambda to a VPC, you specify:

• Subnets: At least two private subnets in different Availability Zones for high availability. Lambda functions will be deployed across these.
• Security Groups: One or more security groups to control what network traffic is allowed to and from your Lambda function. Ensure they permit communication with your private resources.

In the AWS Management Console, when creating or configuring a Lambda function, you'll find a "VPC" section under "Advanced settings" or "Configuration".

Here, you select your desired VPC, choose at least two subnets (for redundancy), and assign appropriate security groups. AWS handles the ENI creation automatically.

This simple Python Lambda function doesn't actually connect to a database, but it shows the structure of a function that could run within a VPC and interact with private resources. If it were in a VPC, it could initiate a connection to a private database.

Try running this example:

If your Lambda function in a private subnet needs to access the internet (e.g., to call external APIs, fetch updates), it won't have direct access.

You'll need a NAT Gateway (Network Address Translation Gateway) deployed in a public subnet within your VPC. Traffic from your private subnets will route through the NAT Gateway to reach the internet.

For accessing certain AWS services (like S3, DynamoDB, SQS) from a Lambda in a private subnet, you can use VPC Endpoints instead of a NAT Gateway.

VPC Endpoints allow your Lambda to communicate with these services privately, without traversing the public internet, which can be more secure and cost-effective.

While powerful, running Lambda in a VPC has implications:

• Cold Starts: The time it takes for Lambda to create an ENI can sometimes increase cold start latency.
• IP Address Management: Each ENI consumes a private IP address from your subnet, so ensure you have sufficient IP space.
• Network Overhead: Managing subnets, security groups, and potentially NAT Gateways adds configuration complexity.

You've learned why and how Lambda functions can operate within a VPC. Let's test your understanding.

In this lesson, you learned that configuring Lambda functions within a VPC enables them to securely access private resources like databases and internal services.

We covered how Lambda uses Elastic Network Interfaces (ENIs) to connect to subnets and security groups, and the considerations for internet access (NAT Gateway) and private AWS service access (VPC Endpoints). This setup is crucial for building secure, enterprise-grade serverless applications.