[ih] NBS seminar on TCP/IP (was TCP RTT Estimator)

Jack Haverty jack at 3kitty.org
Sun Apr 27 11:48:37 PDT 2025


I'm working on it.....

On 4/27/25 06:11, vinton cerf wrote:
> thanks for a great story, Jack! Your adventures in Internet-land 
> deserve a book of its own.
>
> v
>
>
> On Sat, Apr 26, 2025 at 3:55 PM Jack Haverty via Internet-history 
> <internet-history at elists.isoc.org> wrote:
>
>     The TCP implementation I did for Unix on a PDP-11/40 was for use
>     in the
>     BCR project.  I also had lots of performance issues, some of which
>     were
>     finally traced to hardware issues such as connectors!
>
>     In the late 1960s as an undergraduate I had a student job
>     programming a
>     PDP-8 doing data collection at the MIT Instrumentation Lab (also
>     known
>     as "The ILab").  The group I worked for was designing and deploying
>     inertial navigation systems.  At the time their focus was on the
>     Apollo
>     moon missions and "PIGA" devices (Pendulous Integrating Gyroscopic
>     Accelerometers), but their technology had been in use for years in
>     older
>     systems such as the Minuteman ICBMs.   (Google "minuteman piga" if
>     you're curious).
>
>     In EE classes, we had been learning about all sorts of Engineering
>     techniques for optimizing circuit designs - things like Karnaugh
>     maps.
>     One day while sitting at my desk in the ILab, I realized that the
>     engineer sitting at the next desk was an actual "rocket scientist"
>     working on rocket stuff.  So I asked him what Engineering
>     principles and
>     techniques he found most useful in his design work.
>
>     His answer surprised me -- "Connectors are all that matters!".  All
>     designs were focussed on minimizing the number of connectors. Nothing
>     else was considered important, as long as it fit in the size, weight,
>     and power budget.
>
>     Over years of accumulated field data, they had determined that
>     failures
>     were mainly associated with connectors.  A few extra logic gates
>     didn't
>     matter.  An extra connector made the system noticeably less reliable.
>
>     That's why computer problems even today sometimes disappear if you
>     simply unplug and replug the cables.   Any sliding metal-metal
>     contact
>     (i.e., a "connector") eventually corrodes and disrupts whatever
>     signal
>     was travelling through it. The sliding action of replugging usually
>     cleans the contacts and things work again.
>
>     Plugging and replugging connectors isn't easy when the connector is
>     somewhere between the Earth and Moon, or deep inside a missile
>     buried in
>     some farmer's field.  So that engineer told me that they avoided such
>     problems by assuring that every circuit passing through a connector
>     always had some tiny current flowing through it.   Even a few
>     microamps
>     was enough to avoid corrosion.  That was another crucial design
>     principle.
>
>     They didn't teach such things in school back then.   I wonder if
>     that's
>     changed.
>
>     My desk neighbor was Oriental, and told me of an old "Confucian
>     Curse",
>     from a time far before electricity was discovered.  It applies to all
>     sorts of "interfaces" and undoubtedly loses in the translation,
>     but he
>     said today's curse would be "May You Be In Charge Of Connectors!"
>
>     TCP does a wonderful job of hiding such problems.  Such
>     experiences were
>     what motivated me later to push hard for SNMP as instrumentation --
>     including instrumentation of TCP behavior in end-users computers.
>
>     My Unix TCP had counters for things like duplicates, retransmissions,
>     and the like.  Such data was invaluable to even detect that something
>     wasn't working as it should, whether it was a bug in the software, a
>     defect in the protocol, or a flaky connector under the floor.  
>     When I
>     was later involved in operating a large intranet, such mechanisms
>     were
>     very useful in figuring out why the users were rightly complaining
>     about
>     network performance.
>
>     Connectors in networking, whether they are plugs/sockets,
>     protocols, or
>     APIs, matter.
>
>     Jack Haverty
>
>
>     On 4/26/25 03:58, Vint Cerf wrote:
>     > I worked closely with Ed Cain at DCEC. In fact, while I was
>     still at
>     > Stanford, I used to fly in on the red-eye to Dulles, go to DCEC
>     > building which is now walking distance from my Reston Office at
>     > Google, and worked on the BCR project.I would take a late afternoon
>     > flight to get back to SFO and go to my morning lectures the next
>     day.
>     > If memory serves, some of the BCRs were installed at DCEC, on
>     Wiehle
>     > Ave/Sunset Blvd. I even seem to recall that BCR TCP retransmissions
>     > hid a flaky connection until we realized the data rates we were
>     > getting were much worse than reasonably expected. Ed or someone
>     > reporting to him pulled up a floor plate to discover the flaky
>     connector.
>     >
>     > v
>     >
>     >
>     > On Fri, Apr 25, 2025 at 7:13 PM Jack Haverty via Internet-history
>     > <internet-history at elists.isoc.org> wrote:
>     >
>     >     The managers in the military hierarchy mostly had to trust their
>     >     technical staff, who were either within the military or one
>     of their
>     >     contractors, for technical issues.    DCA was responsible for
>     >     operating
>     >     communications infrastructure and changing it as new
>     technologies
>     >     became
>     >     viable.   ARPA was responsible for supplying a stream of new
>     >     technologies.   Together they ran the "pipeline" to move
>     ideas from
>     >     research to operational systems.
>     >
>     >     DCA had a "lab" arm, called the Defense Communications
>     Engineering
>     >     Center (DCEC) where technologies, such as TCP/IP, were
>     tested and
>     >     judged
>     >     to be, or not yet be, appropriate, for general deployment
>     throughout
>     >     DoD.  Ed Cain was in charge of a lab for evaluating TCP/IP.
>     He was
>     >     also a member of Vint's ICCB.  So DCA was aware of the
>     outstanding
>     >     issues.
>     >
>     >     But other issues are usually surfaced as any new technology goes
>     >     into an
>     >     operational environment. Getting "in the field" experience as
>     >     early as
>     >     possible provides a way to shake out those operational
>     issues, and
>     >     feed
>     >     them back into the technical evolution.
>     >
>     >     SATNET had been operated by ARPA for a while, but MATNET, a
>     clone of
>     >     SATNET, was also being evaluated by a Navy lab, run by Frank
>     >     Deckelman,
>     >     with its nodes on ships such as the USS Carl Vinson. Similarly,
>     >     Packet
>     >     Radio networks were in use at places like Fort Bragg and
>     elsewhere to
>     >     get similar operational experience in the Army.  All such
>     testbeds
>     >     could
>     >     generate feedback of issues important to military needs.
>     >
>     >     TCP/IP was ready enough to go "into the field", but was known to
>     >     be not
>     >     yet "done".  There was a list of outstanding issues that the
>     ICCB
>     >     kept,
>     >     of things that needed to be addressed but were not yet
>     resolved by us
>     >     "techies".  Many of the issues involved routing, as well as
>     how to
>     >     best
>     >     support a mix of traffic types with different needs for "type of
>     >     service".  These were all things that us techies simply didn't
>     >     know how
>     >     to do yet.
>     >
>     >     While the then-current version of TCP/IP (V4) was deployed and
>     >     operational experience gathered, the outstanding issues could be
>     >     addressed by the ongoing Research Community, to discuss, debate,
>     >     test,
>     >     and choose appropriate mechanisms for each of the pending
>     issues, and
>     >     incorporate them in the next release of TCP/IP Specifications.
>     >
>     >     Meanwhile, the limitations of the current release were
>     understood and
>     >     avoided.   IP Fragmentation is one example; we knew it
>     didn't work
>     >     very
>     >     well, but the "Don't Fragment" bit was added to provide a way to
>     >     avoid it.
>     >
>     >     So, important "features" weren't "omitted".  They were targetted
>     >     to be
>     >     solved and incorporated in a future release, once the technical
>     >     experts
>     >     reached "rough consensus and running code".
>     >
>     >     The Defense Research Internet (DRI) was supposed to be a new
>     >     high-speed
>     >     network to provide a foundation for continuing such
>     research, such as
>     >     the "policy routing" and "type of service" requirements that
>     were
>     >     on the
>     >     to-do list.
>     >
>     >     I don't myself know much about the DRI though, or what role
>     it might
>     >     have played in creating the "next release" of TCP/IP (V6?).
>     Or if
>     >     DRI
>     >     ever got built at all.  Perhaps someone else does.
>     >
>     >     I think the whole Internet world changed quite a bit when
>     NSF got
>     >     involved, and when the commercial world chose TCP/IP as
>     their target
>     >     architecture.  The military became just one of a global list of
>     >     customers.   NSF instigated the creation of a bunch of regional
>     >     networks, with a mandate and deadline to become
>     self-sufficient, and
>     >     public ISPs started to spread as a result - leading to The
>     >     Internet we
>     >     know today.
>     >
>     >     Jack Haverty
>     >
>     >     On 4/25/25 15:17, Greg Skinner wrote:
>     >     > Just to clarify, I have listened to the computer-freaks
>     podcasts
>     >     about
>     >     > Joe Haughney.  He (and his successors) had access to
>     information
>     >     about
>     >     > the TCP/IP implementations that were discussed at the meetings
>     >     > summarized in the IEN notes.  So in theory, there was a means
>     >     for them
>     >     > to raise concerns about retransmission algorithms, or collect
>     >     > information that could be passed on to people who had those
>     >     concerns.
>     >     >  Furthermore, they were getting feedback on the tcp-ip
>     list about
>     >     > implementation concerns, in general. [1]
>     >     >
>     >     > So far, based on what I’ve read, I don’t see any evidence
>     that the
>     >     > concerns of the military, or users of lossy networks, were
>     given
>     >     > insufficient consideration.  I see that there were
>     features left
>     >     out
>     >     > of RFC 793 that could have mitigated some retransmission
>     issues
>     >     that
>     >     > impacted performance.  But there were DCA people involved
>     who, in
>     >     > theory, could have questioned the wisdom of omitting those
>     >     features,
>     >     > at least.
>     >     >
>     >     > --gregbo
>     >     >
>     >     > [1] https://www.columbia.edu/~rh120/other/tcpdigest_paper.txt
>     >     >
>     >     >
>     >     >> On Apr 25, 2025, at 11:11 AM, Jack Haverty
>     <jack at 3kitty.org> wrote:
>     >     >>
>     >     >> Since I'm listed as one of the speakers, I was probably
>     there.
>     >     But I
>     >     >> don't remember that particular seminar at all.  OTOH,
>     during the
>     >     >> early 1980s I went from Boston to DC probably at least once a
>     >     week,
>     >     >> often to brief someone or some group on The Internet.  I
>     do recall
>     >     >> going to NBS, but can't remember exactly why. It certainly
>     >     could have
>     >     >> been to give a talk on TCP for an hour to some audience.
>     >     >>
>     >     >> That seminar was almost certainly part of DCA's efforts
>     to support
>     >     >> the standardization of TCP as a DoD Standard.  See, for
>     >     example, RFC
>     >     >> 761, published in January 1980 - with a seminar scheduled for
>     >     November.
>     >     >>
>     >     >> The adoption of TCP as a requirement for DoD procurements
>     (not
>     >     just
>     >     >> research contracts) triggered a lot of big and small
>     government
>     >     >> contractors to get interested in exactly what this thing
>     was that
>     >     >> they were going to have to implement.  I recall even
>     getting a
>     >     phone
>     >     >> call from a cousin, who worked at a big government shop,
>     to get a
>     >     >> little free education and advice.
>     >     >>
>     >     >> So the NBS seminar was probably more of an educational
>     venue for
>     >     >> people new to TCP; it was likely not a place where nuances of
>     >     >> retransmission algorithms would be of interest.
>     >     >>
>     >     >> At the time, the ARPANET had been transferred from ARPA
>     to DCA.
>     >     >> Research results were progressing towards operation, part
>     of the
>     >     >> "technology transfer" impetus.  Joe Haughney was in
>     charge of the
>     >     >> ARPANET in DCA Code 535.  IIRC, there's a lot more detail
>     in the
>     >     >> podcasts which Joe's daughter Christine put together
>     recently -
>     >     see
>     >     >> https://www.inc.com/podcasts/computer-freaks
>     >     >>
>     >     >> Jack Haverty
>     >     >
>     >
>     >     --
>     >     Internet-history mailing list
>     > Internet-history at elists.isoc.org
>     > https://elists.isoc.org/mailman/listinfo/internet-history
>     >
>     >
>     >
>     > --
>     > Please send any postal/overnight deliveries to:
>     > Vint Cerf
>     > Google, LLC
>     > 1900 Reston Metro Plaza, 16th Floor
>     > Reston, VA 20190
>     > +1 (571) 213 1346
>     >
>     >
>     > until further notice
>     >
>     >
>     >
>
>     -- 
>     Internet-history mailing list
>     Internet-history at elists.isoc.org
>     https://elists.isoc.org/mailman/listinfo/internet-history
>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 665 bytes
Desc: OpenPGP digital signature
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20250427/1213520c/attachment-0001.asc>


More information about the Internet-history mailing list