RFC 343

IMP System Change Notification 📡✨

Okay bestie, this one is a 1972 RFC 🧚 It's from A. McKenzie at BBN-NET, dated 19 May 1972, and it's filed under category B.1. It obsoletes RFC #331 💅 The whole thing is about the failed attempt to release IMPSYS 2600 on May 15 (the one RFC #331 talked about), and how they're trying again. Spoiler: it was a operational fail, not a coding bug 😭

The original also has a little machine-readable note at the bottom, from when Ronald de Leeuw put it into the online RFC archives in January 1999. Keeping that here too 🎀

Why the first try flopped 💀

Okay so the release attempt on May 15 didn't work, and the team says the reasons were all about operations, not programming bugs. Two main problems:

1. The network had gotten way too big to coordinate operators at every site. Iconic growth, but a nightmare to manage 🙈
2. Here's the messy part: while the new system was being released gradually, IMPs running the new system were getting "poisoned" by neighboring IMPs still on the old system, and vice versa. Working around that made the rollout take way longer than planned 😩

They think they see fixes for both, and they're ready to schedule the next attempt(s). So they're asking every site to help out in a few ways 🫶

What sites need to do 📋

I. Model 516 IMP switch settings 🔧

Sites with Model 516 IMPs were contacted by phone and asked to set their switches like this:

HALT INH     - off
W.D.T.       - off
AUTO RSTRT   - on
MEMORY PRTCT - off

Keep those switches in these positions until further notice 🦋

II. No Trace routine 🚫

Nobody should use the IMP's "Trace" routine during the day(s) of the attempted release 🙅‍♀️

III. Terminal IMPs may go offline ⏸️

Terminal IMPs might be turned off for big chunks of the release day(s). Sorry in advance for the inconvenience, they're asking for patience 🥹

IV. The new schedule 📅

They're not omniscient (lol, relatable 💖), so they might get thwarted again. The plan is:

• The next attempt is Wednesday, May 31
• After that, every Tuesday (June 6, 13, 20, ...) until it actually works 🌈
• The release is set up so site operators don't need to be on duty, but normal coverage is still nice in case something goes sideways 👑
• The release starts early, around 7 a.m. EDT ⭐
• Sites are again told not to schedule production work over the network during the attempt(s) 🚫📡
• Each site will get a phone call once the release is successfully done 📞

They end with a big thank-you to all the site operations staff who worked with them on May 15, and an apology in advance for all the contortions this release is putting everyone through 🫶💜

AAM/jm

[ This RFC was put into machine readable form for entry into the online RFC archives by Ronald de Leeuw 1/99 ]