How one syncs data does massively depend on the scale of data and also on the architecture surrounding this data.
In most cases I've encountered so far in (near) real time systems, “syncronizing data” was just having sending data from the “input unit” to a “processing unit” and from there to the “display unit”.
This scales pretty well with the number of input and display units.
Once there are more than one computational units, then I've often seen the introduction of data-units. But those make it harder to stay inside of real time requirements and therefore I've seen even more often that consistency was dropped in favor of availability…