We are back. Here is what we are planning to do on the project in the coming months:
- Allow several workflows so you can create you own Activity task and you can execute them in the order you want. All will be setup in the config file, no code change. This is a requirement to allow the creation of a Split/Merge workflow.
- Rework the WorkflowTracker to look into the SWF history everytime instead of keeping workflow status in memory. Thus making the Decider fail tolerant and scalable.
- Better thumbnail generation
- Enable Smarter transcoding. Won't upscale the resolution or bitrate. Discard stupid transcoding requests.
- Finish Travis test
- Normalize code to PHP standard
Several have asked this question. The simple reason I used PHP is at the time I started this project there was only 3 AWS SDKs available: Java, Ruby and PHP.
PHP is by far the language I was the most fluent with. I needed to be fast coding this, so I went with PHP. It may not be sexy, but PHP works just fine for this. It's well supported and it is getting better and faster (it's about time). Moreover, most of the work is done by AWS and FFmpeg.
There is now an official AWS Python SDK, too bad it was not there at the time.
Cloud Transcode is a custom distributed transcoding stack using Amazon AWS services.
The goal of this project is to create an open source, scalable and cheap distributed transcoding platform where users have complete control over performance and cost.
We start with video transcoding which is the most costly, but the goal is to transcode any media (audio, documents and images). We use FFMpeg for video transcoding.
Today's commercial solutions for video transcoding are very expensive for large volumes. With this solution you can transcode large quantity of videos at the pace you want, thus controling your cost.
With Cloud Transcode, you control: scale, speed and cost. You can run everything locally if you want, no Cloud instance required. You only need an Amazon AWS account and an Internet connection to use the required Amazon services: SWF, SQS and S3.
It means that you can have a local, hybrid or full cloud setup on Amazon Ec2 instances, it's up to you.
- Video to Video transcoding: One video IN, many videos OUT. Any formats and codecs supported by your ffmpeg.
- Video to Thumbnails transcoding: Snapshot at certain time in video or intervals snapshot every N seconds.
- Watermark integration in video: Take image IN and position a watermark on top of the video. Custom position and transparency.
A Vagrant box (Virtual Machine) which provides pre-configured environment to run the stack has been created to help you test the stack and work on it. You can use Vagrant on any OS and quickly bootstrap.
See: https://sportarchive.hackpad.com/Cloud-Transcode-project-poG8vKTC16J#:h=Quick-start-with-Vagrant
Find the detailed documentation here: https://sportarchive.hackpad.com/Cloud-Transcode-project-poG8vKTC16J
Download the spreadsheet to compare the different Amazon EC2 instances cost and performances: https://github.com/sportarchive/CloudTranscode/blob/master/benchmark/benchmark-aws-ffmpeg.xlsx