[ih] Internet: 57 years later, what does Professor Kleinrock think of the his creation?

[John Day]

You are correct, The ARPANET was what one might call a ‘weak virtual circuit’ as opposed to a pure datagram, like IP.. (The packets followed the same path unless there was the routing caused a change. However, it was the front-end functionality that reassembled packets into message, not by intervening IMPs.

The perspective that hosts were considered part of the network is consistent with that.  They are subtle differences but they made a big difference in the thinking going forward. In fact, Hosts are still considered not part of the network.

IMP-Host created a connection from Host to Host. Then the Host-to-Host protocol created a connection from process-to-process.

Whereas TCP created process-to-process connections and there was no underlying host-to-host connection, like what IMP-Host did.

> On Oct 31, 2026, at 12:23, vinton cerf <vgcerf at gmail.com> wrote:
> 
> Thanks John,
> correct about Wes contribution for sure. The IMP presented a packet interface (BBN 1822) but internally set up a two way connection between end IMPs. Between IMPs the packets flowed in store-and-forward fashion and allowed for adaptive alternative routing, so the routing was not a fixed circuit although the end-point IMPs established an end/end relationship that could be seen as a 'circuit' especially considering reservation of reassembly space as part of the "set up". 
> 
> v
> 
> 
> On Sat, Oct 31, 2026 at 10:51 AM John Day <jeanjour at comcast.net <mailto:jeanjour at comcast.net>> wrote:
>> Vint,
>> A slight correction to your note.
>> 
>> A APRA PI meeting in early 67 rejected Roberts plans for the ARPANET because their systems were too loaded as it was. But Wes Clark suggested putting minicomputers in front of the hosts as a front-end, hence the IMPs. If you look closely, the IMP is really a virtual circuit packet switch on the backend and a front-end to the host.  Host-IMP protocol is part of the front-end part that sets up connections between hosts and reassembles packets into multi-packet messages. It was the front-end function that was the source of the deadlocks we talk about.
>> 
>> Take care,
>> John
>> 
>> > On Oct 31, 2026, at 07:16, vinton cerf via Internet-history <internet-history at elists.isoc.org <mailto:internet-history at elists.isoc.org>> wrote:
>> > 
>> > The 10/29/1969 was a two node test of the Arpanet and involved connecting a
>> > terminal at UCLA to a host at SRI International that were connected via two
>> > Interface Message Processors (IMPs) which we would call packet switches
>> > today. There was no host level protocol for that test. The protocol that
>> > linked the Arpanet hosts was developed under the leadership of Steve
>> > Crocker and was eventually called either the host-to-host protocol or
>> > Network Control Protocol (or Program) in 1970 (see IETF RFC document
>> > series, also started by Steve Crocker; RFC 33 Feb 1970). Three of us worked
>> > at the Network Measurement Lab, led by Leonard Kleinrock: Steve Crocker,
>> > Vint Cerf and Jon Postel, among others. The Internet development started in
>> > the Spring of 1973 at Stanford and was initiated by Bob Kahn (then at
>> > DARPA) who invited Vint Cerf to work with him on the project. The
>> > Transmission Control Protocol (TCP) protocol was published in May 1974 in
>> > IEEE Transactions. By Dec 1974, a complete TCP spec was published in Dec
>> > 1974 (RFC 675) written by Cerf, Yogen Dalal and Carl Sunshine at Stanford.
>> > Three networks made up the cohort populating the early Internet: Arpanet,
>> > Packet Radio network and the Atlantic Packet Satellite network. The first
>> > two-network test of the Internet protocols occurred in Summer 1976 linking
>> > the Packet Radio network to the Arpanet and was conducted by SRI
>> > International. The first three network test was conducted Nov 22, 1977
>> > including Packet Radio, Packet Satellite and Arpanet. Internet went
>> > operational on January 1, 1983.
>> > 
>> > 
>> > 
>> > 
>> > 
>> > On Sat, Oct 31, 2026 at 3:19 AM Frantisek Borsik via Internet-history <
>> > internet-history at elists.isoc.org <mailto:internet-history at elists.isoc.org>> wrote:
>> > 
>> >> "Leonard Kleinrock and a team of scientists at UCLA gave life to the
>> >> Internet at precisely 10:30 PM on October 29, 1969, quite literally
>> >> changing the world.
>> >> 
>> >> 57 years later, what does Professor Kleinrock think of the his creation?"
>> >> 
>> >> https://x.com/lainaminute/status/2115898117229039925
>> >> <
>> >> https://x.com/lainaminute/status/2115898117229039925?a=bXFqldnx0o8OUK9A6xw&s=46
>> >>> 
>> >> 
>> >> https://www.facebook.com/share/v/1BU8EMAfdG/?mibextid=wwXIfr
>> >> 
>> >> https://www.instagram.com/reel/DfFw5IGSTy9/
>> >> <https://www.instagram.com/reel/DfFw5IGSTy9/?mdxt=MTMzMTRwdDliNjcyNw==>
>> >> 
>> >> 
>> >> 
>> >> 
>> >> All the best,
>> >> 
>> >> Frank
>> >> Frantisek (Frank) Borsik
>> >> 
>> >> In loving memory of Dave Täht: 1965-2025
>> >> https://libreqos.io/2025/04/01/in-loving-memory-of-dave/
>> >> 
>> >> 
>> >> https://www.linkedin.com/in/frantisekborsik
>> >> Signal, Telegram, WhatsApp: +421919416714 <tel:+421%20919%20416%20714> <+421%20919%20416%20714>
>> >> iMessage, mobile: +420775230885 <tel:+420%20775%20230%20885> <+420%20775%20230%20885>
>> >> Skype: casioa5302ca
>> >> frantisek.borsik at gmail.com <mailto:frantisek.borsik at gmail.com>
>> >> --
>> >> Internet-history mailing list
>> >> Internet-history at elists.isoc.org <mailto:Internet-history at elists.isoc.org>
>> >> https://elists.isoc.org/mailman/listinfo/internet-history
>> >> -
>> >> Unsubscribe:
>> >> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
>> >> 
>> > -- 
>> > Internet-history mailing list
>> > Internet-history at elists.isoc.org <mailto:Internet-history at elists.isoc.org>
>> > https://elists.isoc.org/mailman/listinfo/internet-history
>> > -
>> > Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
>>