WIP First attempt to make TransportSvc thread safe by removing mutable parameter upd…

First attempt to make TransportSvc thread safe by removing mutable parameter updates or protecting them with mutex locks.

The TransportSvc uses a lot of internal caching of previous results, to speed up subsequent ones (on the assumption the next points are likely to be close to the previous). I could not see a way to make these bits thread safe without mutex locks. I have though tried to isolate them as much as possible, to minimize overheads...

I think though this cache might need to be revisited at some point, as in the threaded environment its not clear to me its going to work as well as in a sequential case...