Message-ID numbers
BBN widens the IMP/Host link field by four bits so messages can be told apart.
Walden, D.C., "Message-ID numbers", RFC 533, DOI 10.17487/RFC0533, July 1973, <https://www.rfc-editor.org/info/rfc533>.
@misc{rfc533,
series = {Request for Comments},
number = 533,
howpublished = {RFC 533},
publisher = {RFC Editor},
doi = {10.17487/RFC0533},
url = {https://www.rfc-editor.org/info/rfc533},
author = {D.C. Walden},
title = {{Message-ID numbers}},
pagetotal = 1,
year = 1973,
month = jul,
} RFC text
Network Working Group David Walden
RFC # 533 BBN-NET
NIC # 17452 July 17, 1973
MESSAGE-ID NUMBERS
It has been noted that the IMP message format (specified in BBN
Report 1822) constrains network users to low throughputs if messages are
to be uniquely identified via the link number. We have recently
implemented a change in the IMP/Host protocol which alleviates this
difficulty. The link field in the IMP/Host and Host/IMP leaders has
been extended to the right by four bits (in other words, it now uses
bits 17 to 28 of the leader), and the name of the link field has been
changed to the _message-id field_. We expect this field to be used to
uniquely identify messages between the IMPs and the Hosts. All 12 bits
in the message-id will be returned to the Host in RFNMs, INCOMPLETE
TRANSMISSIONs, etc. Note that no ordering is implied by the use of the
message-id field.
Given that the Host/Host protocol uses the old link field to
identify connections, the additional four bits in the new message-id
field might well be used to uniquely identify messages on a given
connection (i.e., over a given link). Of course, there is no obligation
for Hosts to uniquely identify their messages (the IMP subnetwork
certainly doesn't care), so this change in the IMP/Host message format
is completely backward compatible. In fact, since the receiving Host
does not have to look at these extra four bits on arriving messages, the
change is completely backward compatible even when senders of messages
do use the extra four bits to uniquely identify messages.
Please store this RFC with your copy of BBN Report 1822 until 1822
is updated.
[ 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 Throughput Problem
The IMP message format is specified in BBN Report 1822. The note observes that this format constrains network users to low throughputs if messages are to be uniquely identified through the link number.
BBN reports that it has already implemented a change in the IMP/Host protocol to ease this difficulty.
If every message only had a small label, you had to wait before sending the next one. That made sending slow.
§2The Message-ID Field
The link field in the IMP/Host and Host/IMP leaders is extended to the right by four bits, so it now uses bits 17 to 28 of the leader. Its name changes to the message-id field, and it is expected to be used to identify messages uniquely between the IMPs and the Hosts.
All 12 bits of the message-id are returned to the Host in RFNMs, INCOMPLETE TRANSMISSIONs and similar replies. No ordering is implied by the field.
The name tag on each message got four more spaces. The tag comes back with the reply so you know which message it was about.
§3Backward Compatibility and Use
The Host/Host protocol uses the old link field to identify connections, so the four new bits might well be used to identify messages on a given connection (that is, over a given link). Hosts have no obligation to identify their messages uniquely, and the IMP subnetwork does not care.
Because a receiving Host does not have to look at the extra four bits, the change is completely backward compatible, even when senders do use them. The note asks readers to keep it with their copy of BBN Report 1822 until that report is updated.
Nobody has to use the new spaces. Old computers still work, and new ones can use them if they want.