coreHTTP Basic S3 Download Demo
Single Threaded Vs Multi Threaded
There are two coreHTTP usage models, single threaded and multithreaded (multitasking). Although the demo on this page runs the HTTP library in a thread, it is actually demonstrating how to use coreHTTP in a single threaded environment (only one task uses the HTTP API in the demo). Whereas single threaded applications must repeatedly call the HTTP library, multithreaded applications instead can execute sending HTTP requests in the background within an agent (or daemon) task.
This demo shows how to use range requests to download files from AWS S3 http server. Range requests are natively supported in the coreHTTP API when you use
HTTPClient_AddRangeHeader() to create the HTTP request. For a
microcontroller environment, range requests are highly encouraged - by downloading a large file in separate ranges, instead of in a single request, each section of the file can be processed without blocking the network socket. Range requests lower the risk of having dropped packets, which require retransmissions on the TCP connection, and so they improve the power consumption of the device.
This example uses a network transport interface that uses mbedTLS to establish a mutually authenticated connection between an IoT device client running coreHTTP and AWS S3 HTTP server.
The core HTTP S3 download demo project uses the
FreeRTOS Windows port, so
you can build and evaluate it with the free Community version of Visual Studios on Windows, without the need for any
particular MCU hardware.
Source Code Organization
The demo project is called
http_s3_download_demo.sln and can be found in the
FreeRTOS-Plus/Demo/coreHTTP_Windows_Simulator/HTTP_S3_Download directory of the
main FreeRTOS download (and on Github).
Building the Demo Project
The demo project uses the
edition of Visual Studio. To build the demo:
- Open the
http_s3_download_demo.sln Visual Studio solution file from within the Visual Studio IDE.
- Select 'Build Solution' from the IDE's 'Build' menu.
Note: If you are using Microsoft Visual Studio 2017 or earlier, then you must select a 'Platform
Toolset' compatible with your version: 'Project -> RTOSDemos Properties -> Platform Toolset'.
Configuring the Demo Project
The demo uses the
FreeRTOS+TCP TCP/IP stack, so follow the instructions provided for the
TCP/IP starter project to ensure you:
- Install the pre-requisite
components (such as WinPCap).
- Optionally set a static or
dynamic IP address, gateway address and netmask.
- Optionally set a MAC address.
- Select an Ethernet
network interface on your host machine.
- ...and importantly
test your network connection before attempting to run the HTTP demo.
As delivered, the TCP/IP stack is configured to use a dynamic IP address.
Configuring the AWS S3 HTTP Server Connection
This demo uses a presigned URL to connect to the AWS S3 HTTP server and authorize access to the object to download.
The AWS S3 HTTP server's TLS connection uses server authentication only. At the application level, access to the
object is authenticated with parameters in the presigned URL query. Follow the steps below to configure your
connection to AWS.
- Set up an Amazon Web Services (AWS) account:
- If you haven't already,
create and activate an AWS account (which includes a free tier).
- Accounts and permissions are set using AWS Identity and Access Management (IAM). IAM allows you to
manage permissions for each user in your account. By default, a user doesn't have permissions until
granted by the root owner.
- To add an IAM user to your AWS account, see the IAM User Guide.
- Grant permission to your AWS account to access FreeRTOS and AWS IoT by adding these policies:
- Create a bucket in S3 by following the steps in
How do I create an S3 Bucket? in the Amazon Simple Storage Service Console
- Upload a file to S3 by following the steps in
How do I upload files and folders to an S3 bucket?.
- Generate a presigned URL using the script located at
For usage instructions, see
The demo retrieves the size of the file first. Then it requests each byte range sequentially, in a loop, with range sizes of
Source code for the demo can be found on
Connecting to the AWS S3 HTTP Server
connectToServerWithBackoffRetries() attempts to make a TCP connection to the HTTP server. If the
connection fails, it retries after a timeout. The timeout value will exponentially increase until the maximum number of
attempts are reached or the maximum timeout value is reached.
connectToServerWithBackoffRetries() returns a failure status if the TCP connection to the server cannot be established after the configured number of attempts.
The source code for
connectToServerWithBackoffRetries() can be found on
prvConnectToServer() can be found on
prvConnectToServer() demonstrates how to establish a connection to the AWS S3 HTTP
server using server authentication only. It uses the mbedTLS-based transport interface that is implemented in the file
Creating a Range Request
The API function
HTTPClient_AddRangeHeader() supports serializing a byte range into the HTTP request
headers to form a range request. Range requests are used in this demo to retrieve the file size and to request each
section of the file.
prvGetS3ObjectFileSize() retrieves the size of the file in the S3 bucket. The
"Connection: keep-alive" header is added in this first request to S3 to keep the connection open after the response
is sent. The S3 HTTP server does not currently support HEAD requests using a presigned URL, so the 0th byte is
requested. The size of the file is contained in the response's
Content-Range header field. A
206 Partial Content response is expected from the server; any other response status-code received is an
The source code for
prvGetS3ObjectFileSize() can be found on
After it retrieves the file size, this demo creates a new range request for each byte range of the file to
download. It uses
HTTPClient_AddRangeHeader() for each section of the file.
The source code can be found
Sending Range Requests and Receiving Responses
prvDownloadS3ObjectFile() sends the range requests in a loop until the entire file is
downloaded. The API function
HTTPClient_Send() sends a request and receives the response synchronously.
When the function returns, the response is received in an
xResponse. The status-code is then verified to
206 Partial Content and the number of bytes downloaded so far is incremented by the
Content-Length header value.
The source code for
prvDownloadS3ObjectFile() can be found on
Copyright (C) Amazon Web Services, Inc. or its affiliates. All rights reserved.