There are probably several valid reasons for this. One that comes to mind is that this is a streaming transfer (of potentially a lot of data) through the LWAPP tunnel and it probably has a fairly tight timeout value to avoid CPU hogging. So if the transfer can't complete in time you'd likely see a message like that.
Can you give more information like the version of ZoneFlex you are running and any insight into the latency between ZD and AP's? Is this a scripted function you are running? If so, a test would be to log directly into the APs to get the info.
MTU might also potentially be a factor