RFC 551

NYU, ANL, and LBL Joining the Net

πŸ“… published August 1973, from ietf

RFC 551 ✨ Network Working Group, August 1973

okay so the title block is giving: this is RFC 551, from the Network Working Group, and the authors are Y. Feinroth (NYU) and R. Fink (LBL). It's dated August 27, 1973 πŸŽ€. The doc doesn't list a category or status, and it doesn't say it obsoletes or updates anything 🫢.

The addresses, verbatim πŸ’…:

New York University
Courant Institute of Mathematical Sciences
AEC Computing and Applied Mathematics Center
251 Mercer Street
New York, N.Y. 10012

ARPA Network Information Center
Stanford Research Institute
Menlo Park, California

Okay bestie, so a few Atomic Energy Commission installations are planning to jump onto the network soon (hopefully soon πŸ™ˆ). The sites are Argonne National Laboratory (IBM 360/195), Lawrence Berkeley Labs (CDC 7600), and New York University (CDC 6600) πŸ–₯️✨. Their apps make an early RJE facility imperative, and even though they're resigned to implementing FTP, they really want to skip the RJE protocol, at least for the first go-around 😭. The goal is to use FTP to transfer a file, have it queued for execution, and then get the output and status info back 🧚.

So the plan is to set up local conventions inside the site-dependent PATHNAME parameter of FTP πŸ’œ. Here are the commands:

STOR           RJE.JOB<rest of pathname>  queue this file for execution
(STOR/RETR)    RJE.PR <    >              transfer remote job print file
(STOR/RETR)    RJE.PU <    >                 "        "    "  punch   "
(STOR/RETR)    RJE.MT <    >                 "        "    "  magtape "

RETR           RJE.STAT  <    >           retrieve  status of remote job

The job execution parameters are not part of the protocol πŸ™…β€β™€οΈ. They have to be put in the standard site-dependent control cards that get sent along with the file. Those cards also decide where the output goes. The user can grab the output with RETR, or the server can send it automatically when it's initiated by STOR.* πŸ”₯. RETR RJE.STAT makes the server create a status file and send it to the user πŸ“¦. And the FTP user/acct/pass logon is only used to validate the data transfer. It does NOT decide whether the job is allowed to run, and it's used to identify and distribute the output πŸ‘‘.

The logon is for the transfer, not for permission to execute. Keep that straight or it'll get messy fr fr πŸ’€

We're a little worried we missed something that's obvious to people who know more than we do, so we're inviting (and requesting!) comments 🌸🫢.

  • note that in this case the RJE server is an FTP user.

[ This RFC was put into machine readable form for entry ] [ into the online RFC archives by Tony Hansen 08/08 ]

πŸ“¦ where this page comes from