Jeff Rearick proposed this definition via email.
Payload: those data bits in a retargeted test whose source was the original PDL test, as distinguished from other bits added by the retargeter to configure the access network, control the interface(s) involved, and implement the protocol(s) being used.
Michele commented: I don't agree with this. The whole point of 1687 is to have access network control inside the payload rather then external operations such as IR selection. I would rather go with this definition:
Payload: Data bits resulting from a retargeted test whose source was the original PDL test, containing also the other bits added by the retargeter to configure the access network, control the interface(s) involved, and implement the protocol(s) being used. To be distinguished from the Control bits/information which might be added by a transaction.