[AWS] Deploy C++ Code for AWS Lambda

Ref: C++ Code Samples for AWS Lambda

Ref: Create a Highly Scalable Image Processing Service on AWS Lambda and API Gateway in 10 Minutes

 

一、Python版本效率

Ref: OPENCV: COMPARING THE SPEED OF C++ AND PYTHON CODE ON THE RASPBERRY PI FOR STEREO VISION

Final comparison of frame capture speed with C++ and Python code with the respective recommended settings:

	1280x480 (stereoscopic)	640x240 (stereoscopic)
C++	40 fps	90 fps
Python	15 fps	20 fps

 

 

二、C++版本实操

Ref: https://github.com/daniel-fudge/aws-lambda-cpp-local-build【针对Ubuntu的版本，可用】

Ref: How to Use C++ with AWS Lambda Runtime. [实操]

Ref: Introducing the C++ Lambda Runtime [code]

 

  老头儿就是简单地按照“实操”的命令行都执行了一遍 by AWS Cloud9。

 

• 环境搭建 

(1). 配置基本的依赖库，yum安装各种依赖。

(2). cmake主要升级到例如3.12：

*** Download CMake from: https://cmake.org/download/
wget https://cmake.org/files/v3.12/cmake-3.12.3.tar.gz

*** Compile from source and install
tar zxvf cmake-3.*
cd cmake-3.*
./bootstrap --prefix=/usr/local
make -j$(nproc)
make install

*** Validate installation
cmake --version

cmake version *.*.*
CMake suite maintained and supported by Kitware (kitware.com/cmake).

(3) Download and compile the runtime 需要自己编译特定机器的runtime。

 

• 准备自定义 C++ 函数

./aws-lambda-cpp/ 编译环境。

./hello-cpp-world/ 代码。

// main.cpp
#include <aws/lambda-runtime/runtime.h>

using namespace aws::lambda_runtime;

invocation_response my_handler(invocation_request const& request)
{
   string json = request.payload;
   return invocation_response::success(json, "application/json");
}

int main()
{
   run_handler(my_handler);
   return 0;
}

最后，make之后再执行如下步骤，打包成一个 hello.zip。CMakeLists.txt 中已定义好了名字。

$ make aws-lambda-package-hello

 

  准备好了 zip，可以开始 CICD 了呢。

 

 

三、CICD - AWS CLI

• Create the role for Lambda function

1) 准备role规则文件：trust-policy.json

{
 "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Principal": {
        "Service": ["lambda.amazonaws.com"]
      },
      "Action": "sts:AssumeRole"
    }
  ]
}

2) 创建命令。 

$ aws iam create-role \
--role-name lambda-cpp-demo \
--assume-role-policy-document file://trust-policy.json

3) 可以看到新添加的role了呢。 

4) 并获得了role的 Amazon Resource Names：“Arn”: “arn:aws:iam::<account_id>:role/lambda-cpp-demo”

 

• Create the Lambda Function

上传.zip作为lambda的运行内容。

hello-cpp-world/build $ aws lambda create-function \
--function-name hello-world \
--role <specify the role arn from the previous step> \
--runtime provided \
--timeout 15 \
--memory-size 128 \
--handler hello \
--zip-file fileb://hello.zip

 

• Test lambda

通过命令触发。

$ aws lambda invoke --function-name hello-world --payload '{ }' output.txt

{
  "StatusCode": 200
}

 

 

四、CICD - Terraform

• Python 对比版本 

Ref: The most minimal AWS Lambda + Python + Terraform setup

简单的代码。

import json

def handler(event, context):
    print("Received event: " + json.dumps(event, indent=2))

打包成zip。

zip lambda lambda.py

定义 Terraform 文件：main.tf。

provider "aws" {
  region = "us-east-1"
}

variable "function_name" {
  default = "minimal_lambda_function"
}

variable "handler" {
  default = "lambda.handler"
}

variable "runtime" {
  default = "python3.6"
}

----------------------------------------------------------

resource "aws_lambda_function" "lambda_function" {
  role             = "${aws_iam_role.lambda_exec_role.arn}"
  handler          = "${var.handler}"
  runtime          = "${var.runtime}"
  filename         = "lambda.zip"
  function_name    = "${var.function_name}"
  source_code_hash = "${base64sha256(file("lambda.zip"))}"
}

resource "aws_iam_role" "lambda_exec_role" {
  name        = "lambda_exec"
  path        = "/"
  description = "Allows Lambda Function to call AWS services on your behalf."

  assume_role_policy = <<EOF
  {
    "Version": "2012-10-17",
    "Statement": [
      {
        "Effect": "Allow",
        "Principal": {
          "Service": "lambda.amazonaws.com"
        },
        "Action": "sts:AssumeRole"
      }
    ]
  }
  EOF
}

 

• Cpp 版本 

/* 暂时没发现 demo code，但核心应该还是 zip；runtime 的类型也要改变一下 */

 

End.