RFC 113

🎀 NETWORK ACTIVITY REPORT: UCSB <- -> RAND 🎀

Network Working Group 💖 5 April 1971 Request for Comments: 113 ✨ NIC 5820

Written by E. F. Harslem, J. F. Heafner and J. E. White 🫶. The header doesn't list a category, an obsoletes line or an updates line, so there's nothing to report on those 🧚.

UCSB RJE/RJOR 💅

okay so the UCSB Remote Job Entry (RJE) and Remote Job Output Retrieval (RJOR) Systems from NWG/RFC #105 have been used and validated from Rand 🦋. The facility is now a production tool on a limited basis for another research group at Rand, and that's giving real-world proof 🔥.

Access to the UCSB facility from Rand goes through the Network Service Program (NSP) 🌈. NSP is driven by Rand Video-Graphic consoles, so a console user can reach both local file storage (at Rand) and the Network 💜. A small module (UCSBMGR) was added to NSP to handle the UCSB RJE and RJOR protocols and data formats 🍓.

In exercising the RJE/RJOR facility over the past two months, typical job sizes were input decks of 800 to 2800 80-character card images and output files of about 30 pages of printer listings 📄✨.

NETWORK OBSERVATIONS 👑

ngl this section is where it gets juicy 😭. When we sent files to UCSB, we ran a timing study over several transmissions of the 2800 record file mentioned above. On average, that file went through at a rate of 250 80-character cards per minute 🌸. (Each 80-character card was its own separate Network message.) Obviously that's way less than the advertised 30 kilobit rate 💀. But remember that the path from Rand to UCSB runs through at least one intermediate IMP, so that's part of it 🫶. On the other hand, the processes at each end of the connection were running at maximum priority with very small loads on either machine 💅. An obvious area for speed-up would be blocking the card images before sending them over the network 🚀.

In the last two months of networking, we noticed about five serious failures in transmitted messages 😳. Twice, the RFNM on the control link from UCSB to Rand was lost. That loss wasn't reported with a type 9 IMP-to-Host message, which is what we'd have expected 🙈. We haven't been able to make the problem happen on purpose, so we can't tell whether it's an IMP problem or a problem with the UCSB Host Interface 🕵️‍♀️.

The other three errors were garbled data messages between the Rand NSP and UCSB RJE 💔. In all three cases, it was the second card image transferred to RJE 🥺. We couldn't trigger this at will, so we haven't been able to track it down yet 🧐. Unfortunately, the HASP system at UCSB just ignored that image instead of printing it, so we don't know the nature or source of the garbling 😤. It could be anywhere from the disk file storage at Rand on down the line 🎯. The problem showed up sporadically in our early transmissions to UCSB and then disappeared ✨. Still, we feel pretty confident that our Host software on either end wasn't at fault 🫡.

Lest those last two figures seem too terrifying 😅, we've run over 100 jobs at UCSB from Rand, and each job was made of many Network transmissions 🏆🎉.

[ This RFC was put into machine readable form for entry into the online RFC archives by Simone Demmel 4/97 ]