[Info-vax] Rethinking DECNET ?

[Shark8]

On 02-Sep-14 07:45, bill at server3.cs.scranton.edu (Bill Gunshannon) wrote:
> In article <Xm5Nv.107728$Fo3.79127 at fx09.iad>,
> 	Shark8 <OneWingedShark at gmail.com> writes:
>> On 01-Sep-14 06:48, Simon Clubley wrote:
>>> That still doesn't change the fact you would be running a TCP/IP
>>> stack alongside your new experimental stack if you wanted your
>>> machine to interact with the rest of the world.
>>
>> That's true and also something that you may, in fact, *not* want: take,
>> for instance, something like the Army's secure network, SIPRnet.
>>
>> IMO, it would be a mistake to make TCP/IP networking mandatory -- sure
>> having it set-up and usable by default is all well and good, but let's
>> /not/ start assuming that everyone has/needs an internet connection.
>> (One of my disappointments with certain office-style programs is the
>> assumption and sometimes requirement for network connecting.)
>>
>>
>> SIPRnet -- http://en.wikipedia.org/wiki/SIPRNet
>
> SIPRnet uses TCP/IP.  It is not a bunch of secure machines connected to
> the INTERNET using an some incompatable protocol, it is a completely
> separate and paralel INTERNET for secure communications.  And, yes,
> there actually are gateways between it and the NIPRnet (the unsecure
> one that connects directly to the INTERNET) that allow unclas traffic
> to pass back and forth while protecting against "leakage".
>
> bill
>

Yes, I realize it's an internet (TCP/IP-based), my point was that just 
throwing up a generalized gateway to different network (even using the 
same protocols) is not always wanted. {The SIPRnet/NIPRnet gateways are 
decidedly *not* generalized in functionality else they could not "allow 
unclassified traffic to pass back and forth while protecting against 
'leakage'".}