Skip to content

dns-smitk/amazon-kinesis-video-streams-producer-c

 
 

Repository files navigation

Amazon Kinesis Video Streams C Producer

Amazon Kinesis Video Streams | Secure Video Ingestion for Analysis & Storage

Coverage Status

Key FeaturesBuildRunDocumentationRelatedLicense

Key Features

Amazon Kinesis Video Streams Producer SDK for C/C++ makes it easy to build an on-device application that securely connects to a video stream, and reliably publishes video and other media data to Kinesis Video Streams. It takes care of all the underlying tasks required to package the frames and fragments generated by the device's media pipeline. The SDK also handles stream creation, token rotation for secure and uninterrupted streaming, processing acknowledgements returned by Kinesis Video Streams, and other tasks.

Build

Download

To download run the following command:

git clone --recursive https://github.com/awslabs/amazon-kinesis-video-streams-producer-c.git

Note: If you miss running git clone with --recursive, run git submodule update --init in the amazon-kinesis-video-streams-producer-c/open-source directory You will also need to install pkg-config, automake and CMake and a build enviroment

Configure

Create a build directory in the newly checked out repository, and execute CMake from it.

mkdir -p amazon-kinesis-video-streams-producer-c/build; cd amazon-kinesis-video-streams-producer-c/build; cmake ..

By default we download all the libraries from GitHub and build them locally, so should require nothing to be installed ahead of time. If you do wish to link to existing libraries you can use the following flags to customize your build.

Cross-Compilation

If you wish to cross-compile CC and CXX are respected when building the library and all its dependencies. See our ci.yml for an example of this. Every commit is cross compiled to ensure that it continues to work.

CMake Arguments

You can pass the following options to cmake ...

  • -DBUILD_DEPENDENCIES -- Whether or not to build depending libraries from source
  • -DBUILD_TEST=TRUE -- Build unit/integration tests, may be useful for confirm support for your device. ./tst/producer_test
  • -DCODE_COVERAGE -- Enable coverage reporting
  • -DCOMPILER_WARNINGS -- Enable all compiler warnings
  • -DADDRESS_SANITIZER -- Build with AddressSanitizer
  • -DMEMORY_SANITIZER -- Build with MemorySanitizer
  • -DTHREAD_SANITIZER -- Build with ThreadSanitizer
  • -DUNDEFINED_BEHAVIOR_SANITIZER Build with UndefinedBehaviorSanitizer
  • -DALIGNED_MEMORY_MODEL Build for aligned memory model only devices. Default is OFF.
  • -DLOCAL_OPENSSL_BUILD Whether or not to use local OpenSSL build. Default is OFF.
  • -DCONSTRAINED_DEVICE -- Change thread stack size to 0.5Mb, needed for Alpine.

DMEMORY_SANITIZER, DTHREAD_SANITIZER etc. flags works only with clang compiler and not with gcc compiler. For using clang (on Ubuntu), do the following -

sudo apt-get install clang
export CC=/usr/bin/clang
export CXX=/usr/bin/clang++
cmake .. -DMEMORY_SANITIZER=TRUE

DLOCAL_OPENSSL_BUILD requires first installing OpenSSL 1.1 (brew install openssl@1.1 for Mac) and then running export PKG_CONFIG_PATH="<YOUR-PATH>/openssl@1.1/lib/pkgconfig" (your path can be printed to terminal using which openssl on Linux/Mac).

DMEMORY_SANITIZER, DTHREAD_SANITIZER etc. flags works only with clang compiler and not with gcc compiler. For using clang (on Ubuntu), do the following -

sudo apt-get install clang
export CC=/usr/bin/clang
export CXX=/usr/bin/clang++
cmake .. -DMEMORY_SANITIZER=TRUE

Build

To build the library run make in the build directory you executed CMake.

make

Run samples

To run the samples:

export AWS_SECRET_ACCESS_KEY=<YourAWSSecretAccessKey>
export AWS_ACCESS_KEY_ID=<YourAWSAccessKey>

For audio+video, run ./kvsAudioVideoStreamingSample <stream-name> <streaming-duration-in-seconds> <sample-location> <audio-codec> <video-codec> <images-flag>

The last 5 arguments are optional. By default,

  • the streaming-duration-in-seconds is 20 seconds
  • sample-location is ../samples
  • audio-codec is aac
  • video-codec is h264
  • images-flag is 0

Valid inputs for <audio-codec> in these samples are:

  • aac
  • alaw

Valid inputs for <video-codec> in these samples are:

  • h264
  • h265

If you want to use the sample for PCM_ALAW/G.711 frames, run ./kvsAudioVideoStreamingSample <stream-name> <streaming-duration> <sample-location> alaw h264 0

This will stream the video/audio files from the samples/h264SampleFrames or samples/h265SampleFrames and samples/aacSampleFrames or samples/alawSampleFrames respectively.

If you want to enable KVS events in fragment metadata, change the <images-flag> parameter from 0 -> 1. This feature is found only in the audio/video sample, but can be written into the video only sample as well.

For video only, run ./kvsVideoOnlyRealtimeStreamingSample <stream-name> <video-codec> <streaming-duration-in-seconds> <sample-location> OR ./kvsVideoOnlyOfflineStreamingSample <stream-name> <video-codec> <streaming-duration-in-seconds> <sample-location>

Example: ./kvsVideoOnlyRealtimeStreamingSample myTest OR ./kvsVideoOnlyOfflineStreamingSample myTest

This will stream the video files from the samples/h264SampleFrames.

OR you can run ./kvsVideoOnlyRealtimeStreamingSample myTest h265 OR ./kvsVideoOnlyOfflineStreamingSample myTest h265

This will stream the video files from the samples/h265SampleFrames.

Note: The resulting sample video is the same.

For audio only, run ./kvsAudioOnlyStreamingSample <stream-name> <streaming_duration> <sample_location> <audio-codec>.

This will stream the audio files from the samples/aacSampleFrames or samples/alawSampleFrames (as per the choice of audio codec in the last argument) respectively.

Running with IoT credential provider

To run the samples with IoT credential provider:

  1. Run the IoT thing generation script available under scripts: source scripts/generate-iot-credential.sh. For more information on IoT set up, visit AWS KVS IoT Set up
  2. Uncomment #define IOT_CORE_ENABLE_CREDENTIALS 1 in the relevant sample
  3. Build the changes: make
  4. Run the sample using the instructions in previous section.

Fragment metadata

./kvsVideoOnlyRealtimeStreamingSample is the only sample that has the fragment metadata implemented out of the box.

In addition to the required arguments above, this sample has an additional argument:

./kvsVideoOnlyRealtimeStreamingSample <stream-name> <video-codec> <streaming-duration-in-seconds> <sample-location> <num-metadata>

num-metadata -- the number of sample fragment metadata key-value pairs that are added to each fragment. Min: 0, Max: 10. Default: 10.

Setting log levels

Setup logging:

Set up the desired log level. The log levels and corresponding values currently available are:

  1. LOG_LEVEL_VERBOSE ---- 1
  2. LOG_LEVEL_DEBUG ---- 2
  3. LOG_LEVEL_INFO ---- 3
  4. LOG_LEVEL_WARN ---- 4
  5. LOG_LEVEL_ERROR ---- 5
  6. LOG_LEVEL_FATAL ---- 6
  7. LOG_LEVEL_SILENT ---- 7
  8. LOG_LEVEL_PROFILE ---- 8

To set a log level, you can set it using the deviceInfo structure.

pDeviceInfo->clientInfo.loggerLogLevel = LOG_LEVEL_DEBUG;

By default, our samples set the log level to LOG_LEVEL_DEBUG.

The SDK also tracks entry and exit of functions which increases the verbosity of the logs. This will be useful when you want to track the transitions within the codebase. To do so, you need to set log level to LOG_LEVEL_VERBOSE and add the following to the cmake file: add_definitions(-DLOG_STREAMING) Note: This log level is extremely VERBOSE and could flood the files if using file based logging strategy.

Run unit tests

Since these tests exercise networking you need to have AWS credentials specified, specifically you need to:

export AWS_SECRET_ACCESS_KEY=<YourAWSSecretAccessKey>
export AWS_ACCESS_KEY_ID=<YourAWSAccessKey>

Now you can execute the unit tests from the build directory as follows: ./tst/producer_test

Offline mode

The samples run in near real time mode by default. In order to set up offline mode, the following APIs can be used in the samples instead of the realtime variant:

For video only: createOfflineVideoStreamInfoProviderWithCodecs() For video and audio: createOfflineAudioVideoStreamInfoProviderWithCodecs()

The 2 APIs are available in this header file.

DEBUG

  • When building OpenSSL during cmake .., if you encounter an architecture error such as ld: symbol(s) not found for architecture i386, building with a local OpenSSL build may help. First install OpenSSL 1.1 (for Mac: brew install openssl@1.1). Next set export PKG_CONFIG_PATH="<YOUR-PATH>/openssl@1.1/lib/pkgconfig" (your path can be printed to terminal using which openssl on Linux/Mac). Now set the following flag to ON when building: cmake .. -DLOCAL_OPENSSL_BUILD=ON. If there are still errors regarding locating the local OpenSSL library:
    • The following environment variables may need to be set to export: export LDFLAGS="-L<YOUR-PATH>/openssl@1.1/lib" and export CPPFLAGS="-I<YOUR-PATH>/openssl@1.1/include"
    • The path to OpenSSL’s root directory may need to be specified when running cmake: -DOPENSSL_ROOT_DIR="<YOUR-PATH>/openssl@1.1/include/openssl"
    • If you need to have openssl@1.1 first in your PATH, run: echo 'export PATH="<YOUR-PATH>/openssl@1.1/bin:$PATH"' >> ~/.zshrc

Development

The repository is using develop branch as the aggregation and all of the feature development is done in appropriate feature branches. The PRs (Pull Requests) are cut on a feature branch and once approved with all the checks passed they can be merged by a click of a button on the PR tool. The master branch should always be build-able and all the tests should be passing. We are welcoming any contribution to the code base. The master branch contains our most recent release cycle from develop.

Release

The repository is under active development and even with incremental unit test coverage where some of the tests are actually full integration tests, we require more rigorous internal testing in order to 'cut' release versions. The release is cut against a particular commit that gets approved. The general philosophy is to cut a release when a set of commits contribute to a self-containing feature or when we add major internal functionality improvements.

Versioning

We deploy 3 digit version strings in a form of 'Major.Minor.Revision' scheme.

  • Major version update - Major functionality changes. Might not have direct backward compatibility. For example, multiple public API parameter changes.
  • Minor version update - Additional features. Major bug fixes. Might have some minor backward compatibility issues. For example, an extra parameter on a callback function.
  • Revision version update - Minor features. Bug fixes. Full backward compatibility. For example, an extra fields added to the public structures with version bump.

Documentation

Related

License

This library is licensed under the Apache 2.0 License.

Releases

No releases published

Packages

No packages published

Languages

  • C 56.9%
  • C++ 40.1%
  • CMake 2.3%
  • Shell 0.7%