[ih] Internet at Sea

[Jack Haverty]

Some better search term on discover.dtic.mil found this - the sequel to 
the report I just mentioned, published a year later:

https://apps.dtic.mil/sti/tr/pdf/ADA239925.pdf

Chapter XVI is titled "SIMNET" which has a special, but likely as yet 
untold, history with The Internet.

Sometime in late 1982 or early 1983, ARPA asked our "Internet" group at 
BBN to get involved with a project that might be able to use the 
emerging Internet for its communications infrastructure.  That program 
was called SIMNET, or SIMulation NETwork, and the concept was to create 
a video-game type of training tool for use in combat training.  The 
initial use was training M-1 tank crews, but the concept included other 
participants, such as helicopters.  Mike Kraley and I went to a bunch of 
meetings to brainstorm and flesh out the ideas, as part of our ongoing 
work on ARPA Internet-related contracts.

It became clear that for such "gaming" applications, network latency was 
important.  It mattered a lot.  If you fired at the enemy, you should be 
able to see the results immediately and consistently.  A training system 
had to be accurate for the things that mattered, but could cut corners 
to save costs for the things that didn't.

Inside the Internet world, that need was one of the motivations for the 
introduction of the TOS field (Type Of Service) in the IP header.  Our 
conclusion was that the Internet would have to support at least two 
different types of behavior.  Possibly more since SIMNET was also 
envisioned to simulate radio traffic and "chatter" between the crews in 
the simulation, using packet voice.

Datagrams associated with things like firing weapons or vocal snippets 
could be small, but had to get delivered quickly. Datagrams associated 
with things like detailed maps could be delivered at a more leisurely 
pace.  Terrestrial routes would be good for the former, and 
geosynchronous satellites appropriate for the latter.  Of course there 
would also need to be new appropriate routing mechanisms to make it all 
work as envisioned.

At BBN, we wrote a proposal to start an actual SIMNET project. Shortly 
thereafter, in July 1983, BBN reorganized and that project was approved 
and the contract assigned to a part of BBN that had been doing various 
training systems.  So I never got to drive an M1 tank (which was an ARPA 
mandated requirement for everyone assigned to the project).

SIMNET ended up being very successful, as detailed in that report. But 
the implementors discovered that the Internet, which hadn't implemented 
any mechanisms for TOS, couldn't provide the communications services 
that SIMNET needed.   They had to build their own private communications 
system instead.

In retrospect, we probably didn't do enough to lay out that plan for 
coordinating the SIMNET and Internet evolution.  ARPA reorganized at 
about the same time, the ICCB became the IAB, the people involved 
changed, and the plan was lost.  SIMNET was successful, but TOS support 
in the Internet didn't happen.

/Jack Haverty

-------------- 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/20251004/ee8a253c/attachment.asc>