by Darius Kazemi, April 23 2019
A success and some bugs
RFC-113 is titled “Network Activity Report: UCSB <—> RAND”. It's authored by Harslem and Heafner of RAND and Jim White of UCSB. It's dated April 5th, 1971.
The technical content
RFC-105 was a specification for running jobs at the UCSB Computer Center, and a month later we have this report of RAND successfully running jobs at UCSB via that protocol.
The folks at RAND ran “over 100 jobs […], each job consisting of many Network transmissions.” Over the course of those 100 jobs, they encountered five network failures. Two of these involved an error with the RFNM (Request for New Message) signal not being transmitted and its loss was not reported as an error either. The other three seem to be the same type, involving “garbling of a data message” but they seem to have no idea the cause of it or even what data was garbled in what way.
Since they estimate “800 to 2800” punch card images being sent over about 100 separate jobs, we can estimate that there were about 180,000 punch card images sent. Five serious errors is not a whole lot!