Transcoding has typically been a file-in, file-out workflow, and transcoding jobs are distributed around the transcoding farm, as capacity becomes available. This architecture works well for bulk transcoding where the amount of content saturates the transcoding farm and you have 100% utilization of your resources.
For the situation where a customer has content that requires an immediate turnaround (i.e. breaking news stories, late edits to on-air content), RadiantGrid has developed a revolutionary process called grid transcoding, where source content can be transcoded in parallel across all your available transcoding resources. This process distributes virtual slices (ex: 1 minute) of the source content across all available resources, and once these slices have been transcoded, they are reassembled to generate the final transcoded file.
Grid transcoding has the ability to scale performance linearly as you add compute resources. If an MPEG-2 source file takes 30 minutes to transcode to H.264 on a single server, then with grid transcoding enabled, that same file will take less than 6 minutes to transcode on five transcoding nodes. Or, if you had ten transcoding nodes in your farm, that same file would take less than 3 minutes to transcode.
In addition, grid-based transcoding provides better utilization of your transcoding farm, when transcoding content of mixed lengths. No longer will you be waiting on a single hour-long movie to finish, while the rest of your farm sits idle. With grid transcoding, your source content – short or long form – is transcoded via virtual slices, which are load-balanced across the entire farm, and provide the best efficiency and speed possible.