Currently parallel suffixes are lists of requests, because of the fact that suffixes need to be run in parallel they never become very long. A more realistic approach would be to use lists of lists of requests, e.g. batches of parallel requests. This approach is taken by the quickcheck-state-machine library.