Scheduled routing can also be useful in situations where different streams of message traffic have different bandwidth needs. By preallocating differing amounts of schedule time to each stream, bandwidth is divided as necessary for efficient application performance, even when all the streams are being driven full-out: the higher-bandwidth streams don't crowd out the lower-bandwidth streams, nor do a number of lower-bandwidth streams combine to reduce the required bandwidth on one higher-bandwidth stream. Message prioritizing in traditional online routing systems is similar to this approach, but does not offer as fine a degree of control over the relative bandwidth needs of a collection of streams.
As an example, consider a rendering engine that generates images in real time using inter-node communications. Periodically, the system may change to a different image by distributing a large data set to all the nodes. However, the time-critical communications between the nodes as they render the last few frames of the image should not be compromised by the high-volume data set for the next image being downloaded into the mesh.
Similarly, it is also possible to tune a stream's latency to meet application needs. A given stream may be low bandwidth, but may want to run with as little latency as possible. This need can be met by appropriate scheduling, without compromising the ability of high-bandwidth streams to run at their full designated bandwidth.