[ih] Overlay networks

[Dave Crocker]

On 8/20/2025 4:23 PM, Jack Haverty via Internet-history wrote:
> an Internet built on top of a separate Internet


In the early 1990s, as discussions were leading to an IPv6, one of the 
concerns about about doing testing at scale, since no one had a test 
network that was large enough.

At a meeting (in Colorado?) during such a discussion, I noted that a 
number of organization DID have sizeable test networks and perhaps we 
could connect them via Internet IPv4-based tunnels, and create a much 
larger, aggregate test network.  The idea was received with what I will 
politely call skepticism, but apparently it later caught on.

Prior to this, while I was at Ungermann-Bass, we adapted their existing 
NETBIOS over UB's XNS to run over UB's nascent TCP/IP. Almost as an 
afterthought, we added a configuration option to allow sending NetBios 
packets to a remote (enterprise) network. There were several other 
proprietary NetBios over TCP implementations but we were the only ones 
to have done this.  It became a point of contention during the 
multi-vendor NetBios standards effort that followed.

When there was contention in the group about differences between vendor 
choices, discussions often broke down.  Eventually, Vint had to be 
brought in to facilitate.  This entire process was where I finally 
learned the importance of the difference between being a skilled 
designer and being a skilled implementer.  Prior to that, I'd mostly 
been around people with both skills...

d/

-- 
Dave Crocker

Brandenburg InternetWorking
bbiw.net
bluesky: @dcrocker.bsky.social
mast: @dcrocker at mastodon.social