[246] in Commercialization & Privatization of the Internet
Re: Other Researchers
daemon@ATHENA.MIT.EDU (pgross@NRI.Reston.VA.US)
Thu Feb 28 18:27:29 1991
To: Sean McLinden <sean@dsl.pitt.edu>
Cc: com-priv@psi.com
Date: Thu, 28 Feb 91 18:02:17 -0500
From: pgross@NRI.Reston.VA.US
> The first is GOSIP. In theory, ...
There is much misunderstanding about what GOSIP actually means. It does
*not* dictate the use of OSI, it does *not* explicitly rule against the
use of TCP/IP (or any other protocol family), and it does *not* mandate a
transition to OSI from any current technology.
GOSIP is the "Govt OSI Profile". OSI has so many parameters and options
that it is common to write a "Profile", so that different OSI implementations
can interoperate by using the same parameter/option settings.
Govt network purchasers can buy and use whatever network technology they
want -- they just *also* have to purchase OSI. One intent of GOSIP is to
mandate that all govt purchasers have one standard interoperable method
(ie, OSI) by which they could all communicate. However, GOSIP does not
say that they must use that method as their only or primary means of
communication. I know this sounds confusing, but the bottom line is that
GOSIP says govt network purchasers have to buy OSI, but it does not say
that they have to use it.
Please see RFC 1169, "Explaining the Role of GOSIP". One of the co-authors
of RFC 1169 was Kevin Mills of NIST, who was also one of the key players in
developing GOSIP.
Phill Gross
