Synchronizing between a message and a tf.
I am writing an application, where I subscribe to large messages (such as images or point clouds) and I use them in conjunction with a tf transform. The message arrives, I query the tf server for the transform, and if the transform satisfies a certain criteria, I do something with the message. Otherwise, I ignore the message.
If the message publisher and message subscriber are on different computers over a network, this generates a lot of wasted traffic for all the discarded messages. Is there an elegant way to resolve this? For example, subscribe to a specific tf, process it, and if it meets the criteria, request a single message?