Introduction

This is a preview release of the High-level enhancements for AWS in C++, which consists of high-level abstractions for AWS services in the style of the C++ standard library. Included examples show how to read and write S3 objects using IOStreams, call AWS lambdas in the style of thread functions and demonstrate how STL algorithms (currently only transform) can take advantage of AWS Lambda concurrency to speed up an example program from 15 minutes to 9 seconds.

As an experimental preview release, there are many gaps in coverage that we will fill over time and expect that interfaces will change. Note that this is provided on an AS-IS basis. See Section 7 of LICENSE for more information.

Requirements

For users

• zip
• CMake 3.15+
• AWS SDK for C++ 1.9+
• AWS CLI for uploading the AWS Lambda functions.
• A C++20 compatible compiler. (C++17 may work)
• AWS Lambda Runtime for C++
• TL Expected for compilers that do not yet support std::expected

Additional dependencies for developers

• GoogleTest 1.11+

To make changes to the codebase and contribute changes please read our development guide

Installation

The installation follows Modern CMake idioms. A quick start script to build and install the projects is below for reference.

The following script assumes the project code was downloaded and the shell is at the project source code directory.

mkdir build
cd build
cmake ..
cmake --build . 
cmake --install . 

Alternatively, a native installation package can be created using cpack for distributing to multiple machines after built once.

mkdir build
cd build
cmake ..
cmake --build . 
cpack

cpack will generate a bundle package with the installation artifacts generated by cmake. For example, in MacOS, it creates a awslabs_enhanced_cpp-0.1.0-Darwin.tgz file that can be expanded in the installation location.

Usage

See Quick Tutorial for usage instructions. Examples are in the doc/examples directory.

Basic Features

RAII classes

RAII classes AwsApi for initializing the API, Logging for initializing logging.

S3 stream abstractions

The is3stream and os3stream libraries can be used to read and write data in s3 in a similar style to the use of fstream to read and write files.

Lambda abstractions

The lambda_client.h header facilitates calling lambdas from C++ similarly to other callables. See lambda_add_example.cpp for a simple example. The central_limit_theorem.cpp example shows how to use the transform function template to take advantage of AWS Lambda's built-in concurrency to efficiently transform an iterator rannge.

The aws_lambda.h header facilitates the creation of lambda functions. See lambda_add_fn.cpp for an example. To package the lambda function for the cloud, you can use aws_lambda_package_target from the AWS Lambda Runtime for C++ as shown in the CMakeLists.txt in the examples directory.

To deploy the lambda function from the CLI, you can use the command

cli_deploy_cpp_lambda lambda_name iam_role zipped_lambda_package

You can also use your preferred Infrastructure as Code technique,and there is an (under development) script for AWS SAM deployment included as well.