10

I'm trying to complete some information on the history of DNS and name resolution in networking. I've found some information on ARPANET and BBN's protocols relying on the numerical machine code for the 5 bit destination portion of the packets. I then found that host files seemed to have started their use around 1974 to help resolve names to the machine number.

However, in 1971 they also sent the first email with SNDMSG and PBS said they used the email CSX@UCLA. If that's true, did they have some kind of resolving on their hosts that wasn't a host file?

Additionally, the TIP guide from the 1972 conference shows them entering 70 for host #70 and 65 for UCLA. How did they connect with the remote computers before then to teletype to it, did they use the machine number? When did they start to use name resolution?

Edit: Here is some more relevant information I found:

  1. RFC 226, 247, 273, 280, 289, and 384 all relate to what could be considered a "hosts" file. Initially started in RFC 226 by Peggy Karp as Host Mneumonics. There were a few iterations over this through the other RFCs until the Hosts.txt file maintained by Elizabeth in 1974
  2. If we combine this with PBS's claim that Ray's first email using CSX@UCLA, presuming that is correct. That would line up with RFC 226, as in 247 Peggy is already recommending that UCLA use the host name UCLA-7 and UCLA-91. I can't find an exact date yet, but most claim that Ray's email was sent in October of 1971.
  3. I think CYPNET used the same host-to-IMP protocol as any other connection they would have made even after setting up the "Host Mneumonics". It still would be nice to see how CYPNET worked to resolve the name for the terminal to start to package.

Hopefully this is cool for others to read. I'll leave the questions up in case there is more information. Specifically, on the CYPNET protocol

Improve this question
17
  • I recently watched a Youtube video produced by RetroBytes regarding the history of the internet. I can't vouch for his accuracy, but I have no reason to doubt it either. There is a section relating to DNS that you might find interesting beginning at 20:40. Here is the link: youtu.be/r03AIGRmRAM?si=CCzVCdbA4gE_8_S9&t=1241
    – Geo...
    Commented Jul 17 at 21:09
  • 2
    I remember back when we had to use bang paths. Mail addresses looked like host1!host2!...!hostN!user, where host1 was the name of a computer reachable by 1 hop from your computer as listed in your local /etc/hosts. Similarly, host2 was a computer reachable by 1 hop from host1, and so on to hostN which was your recipient's computer, with user name user. You had to know the exact route of hops (often 6 or more) from your computer to the recipient. DNS was a welcome improvement.
    – DrSheldon
    Commented Jul 18 at 0:47
  • 3
    @DrSheldon, I thought that bang paths were for hosts that were not connected to Arpanet/Internet. I thought that the "hops" were achieved by UUCP connections. en.wikipedia.org/wiki/UUCP#Bang_path I agree that DNS was a big improvement over,... um,... I guess everything that came before, but you couldn't use DNS, until you got connected to the Internet.
    – Solomon Slow
    Commented Jul 18 at 19:39
  • 2
    DEC's internal network managed to handle tens of thousands of nodes using the equivalent of host files. Essentially, every system updated its copy of the database on a nightly basis. I forget the details now but I think a hierarchical distribution was arranged - i.e., you got a copy from a topologically-nearby distribution hub.
    – dave
    Commented Jul 19 at 2:10
  • 1
    @JessFuckett it wasn't malicious but the amount of shenanigans that used to go on in the university computing labs where I was an undergraduate in the 1990's meant a bit more security would have been welcome :) One thing that I still laugh about is a PhD colleague who somehow managed to wrangle herself a dedicated Solaris sun station - but never learn that you could run any process on any other machine on the network. So we ran all our jobs on her machine and other than always complaining that it was quite slow she never worked it out.
    – David Waterworth
    Commented Jul 22 at 6:07

0

Reset to default

You must log in to answer this question.

Browse other questions tagged .