Response to RFC 597: Host status
A short list of questions about how well connected the ARPANET map of November 1973 was.
Burchfiel, J.D., "Response to RFC 597: Host status", RFC 603, DOI 10.17487/RFC0603, December 1973, <https://www.rfc-editor.org/info/rfc603>.
@misc{rfc603,
series = {Request for Comments},
number = 603,
howpublished = {RFC 603},
publisher = {RFC Editor},
doi = {10.17487/RFC0603},
url = {https://www.rfc-editor.org/info/rfc603},
author = {J.D. Burchfiel},
title = {{Response to RFC 597: Host status}},
pagetotal = 1,
year = 1973,
month = dec,
} RFC text
Network Working Group J.D. Burchfiel
RFC # 603 BBN-TENEX
NIC # 21022 31 December, 1973
Response to RFC # 597: Host Status
I have several questions about the November 1973 ARPANET
topographical map:
1. AMES is 4-connected, i.e. four network connections will go down
if the IMP fails. Is there some aspiration that IMPs should be
no more than three connected?
2. The seven IMPS in the Washington area are arranged into a loop.
This guarantees that local communication can take place even if
one connection fails, and is probably a worthwhile preparation
for area routing. On the other hand, for example, a break
between MIT-IPC and MIT-MAC will require them to communicate
through a 12-hop path through Washington. This can be remedied
by a short (inexpensive) connection between Harvard and Lincoln
Labs. Is there a plan to pull the Boston area, the San
Francisco area, and the Los Angeles area into loops like the
Washington area?
[ 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 ]
AI ELI5
On this page
§1The Map Under Review
RFC 603 is a response to RFC 597, which was about host status. J. D. Burchfiel of BBN-TENEX dated it 31 December, 1973 and says he has several questions about the November 1973 ARPANET topographical map.
The document is one page long and consists of two numbered questions. It proposes no protocol and sets no requirements.
Someone studied a map of the network and wrote down two questions about it.
§2Too Many Links on One Node
The first question is about AMES, which is 4-connected: four network connections will go down if its IMP fails. The author asks whether there is some aspiration that IMPs should be no more than three connected.
If one switch has four wires plugged into it, all four stop working when it breaks. The author asks if four is too many.
§3Loops Around Regions
The second question concerns the seven IMPs in the Washington area, which are arranged into a loop. That guarantees local communication can continue if one connection fails, and the author thinks it is probably a worthwhile preparation for area routing.
He contrasts this with Boston: a break between MIT-IPC and MIT-MAC would force them to communicate through a 12-hop path through Washington. He suggests a short, inexpensive connection between Harvard and Lincoln Labs as a remedy, and asks whether there is a plan to pull the Boston, San Francisco and Los Angeles areas into loops like Washington's.
A ring of friends can still pass notes if one link breaks. Some neighbourhoods had no ring, so a broken link meant a very long detour.