RFC 613

RFC 613: Network connectivity, a response to RFC #603 💖

okay bestie, so the title block is giving 1974 vibes 🎀 RFC 613 is by Alex McKenzie from BBN-NET (NIC # 21525), it's from the Network Working Group, and it's dated January 21, 1974. It doesn't list an obsoletes or updates field, and it doesn't say what category it is. It's basically a reply to RFC #603 🦋

Network topology is a complicated political and economic question with obvious technical overtones 🌸 Alex isn't trying to cover every possible argument here. She just wants to respond directly to the points raised in RFC #603. literally the most focused little reply ever 💅

Responses to RFC #603 ✨

1. Connectivity counts, not just the failed node 🔥

okay so the big thing is whether it's good or bad for a node (AMES) to be four connected. The key question isn't how many circuits get affected when a node fails. It's how well the network is still connected after a node fails 🧚 If ALL nodes were four-connected, nobody would really argue that was bad for reliability, fr fr.

The weak spots aren't the three-connected and four-connected nodes. They're the ONE-connected nodes (Hawaii, London) and the two-connected nodes 💀 Alex agrees with Burchfiel's implied argument that it's better to have two adjacent three-connected nodes than a four-connected node next to a two-connected node. The problem is that the realities of installing interfaces and common carrier services make the Network grow in sub-optimal ways 😭

2. Loops aren't good on their own 🌈

"Loops" aren't good per se. They just look good because making loops increases connectivity, which reduces the effect of multiple failures 🙈 But adding circuits costs ARPA money, both the capital cost for IMP interfaces and the recurring cost for the circuits 💸

BBN's network group had suggested to ARPA several times that "connectivity should be increased." It wasn't until late December 1973 that they actually made specific suggestions for where to put additional circuits 🍓 Those recommendations weren't about building loops (even though they may have that effect). They were about breaking the long chains of IMPs that formed as the Network grew 🫶 ARPA and NAC are now presumably evaluating the suggestions and maybe coming up with other possibilities ✨

🎀 Archive note, kept exactly as it was:

   [ This RFC was put into machine readable form for entry ]
   [ into the online RFC archives by Alex McKenzie with    ]
   [ support from GTE, formerly BBN Corp.            10/99 ]