Most customers start out with a limited deployment of a few transcoding nodes, but very often they find themselves needing to expand the farm to manage more internal projects, and more concurrent transcoding. When scaling compute resources, and sharing between multiple departments or internal projects, customers find themselves facing issues around resource management.
RadiantGrid provides several layers of resource management in the RadiantGrid Platform. First, we manage the number of concurrent workers on each transcoding server. The capacity of workers per server is related to the number of cores or CPUs installed on the server.
Next, we manage the number of nodes in the farm. Typically each server in the farm can be utilized as a transcoding node, but we can also segment the server architecture to use some dedicated nodes for storage management, and leave the rest solely for transcoding.
Finally, we can dynamically manage the capabilities and responsibilities of the nodes within the farm. Some customers want to allocate resources by project, by department, or even by format. Our innovative technology called job tagging allows each transcoding job to be routed to the proper segment of the farm based on customer needs, the outgoing format, or prioritization. The segmentation of the farm by tags is dynamic and can be configured on the fly by an operator, and these tags can be assigned to each job on the fly by the end-user.