[479] in Commercialization & Privatization of the Internet
Re: Why commercial sites aren't on the net
daemon@ATHENA.MIT.EDU (Morten Reistad)
Fri Mar 29 06:51:35 1991
Date: 28 Mar 91 20:15 +0100
From: Morten Reistad <MRR@boers.uu.no>
To: com-priv@uu.psi.com
Cc: mrr@boers.uu.no
In <9103241957.AA02674@volitans.morningstar.com> Bob Sutterfield
<bob@morningstar.com> writes (in a reply)
> [...] Just as in the UUCP and Fido worlds, IP
> links can now be established, used, and decomissioned as needed; all
> without prior approval of central authorities. Routing between those
> two sites who want to exchange packets becomes a local issue. Dialup
> IP can be "networking for the rest of us".
Unfortunately, this kind of IP networking will be A LOT like uucp or
Fido, that is, with a lot of builtin chaos due to the total lack of
a coordinated infrastructure. (no critisism of the few people trying
to coordinate uucp or fido intended.)
To prevent this from becoming a total chaos we should make some
attempt at navigating the existing carrier networks before we end
up in the uucp mess. It could be as simple as address-resolution
servers for each of the (telephone/X25/etc) networks.
> This still doesn't address the issue of whether the central management
> at NIC.DDN.MIL is going to authorize the propagation of *.su routes
> through the NSFnet's core gateways. But at least Mr Volodin and his
> comrades can begin enjoying the benefits of internal IP connectivity
> while the political dust settles. All they need is a PC or a UNIX box
> or a Netblazer at each end, modems to connect them, and someone
> willing to pay the phone bills.
The whole idea of a central management deciding who should connect to
whom is an athoritarian one, and one that is not in keeping with the western
ideas. The central management should regulate the namespace, period.
Then NSFnet, or anyone else, should decide for themselves what kind of
traffic they want to carry as long as they pay for it. But do not let
this interfere with the current excellent efforts to keep the namespaces
of network numbers and domains so clean and uncluttered as they are.
(If the Moscow Stock Exchange should take off, I would VERY much like
connectivity in the .su direction. ;-)
I.E: We should stick to the normal principles laid down in international
transportation law, and keep the regulators, carriers and customers well
apart.
In <9103241314.aa08097@looking.on.ca> Brad Templeton <brad@looking.on.ca>
comments:
> I would venture that stock tickers are one of the special few
> applications for which satellites are the right answer, since
> they are a one to many broadcast sort of thing.
Well, in the outset it sounds OK, but then we want a link layer, and a
network layer, and by then (simplex, delayed) sattelite paths are not
so fun, any longer.
In <9103242046.AA13828> Brian Lloyd <brian@napa.telebit.COM> comments
>> "I do, however, still sense a void in the just-below-ip strata in
>> the protocol stack. Isn't there a need for a simple connection-
>> less packet-mover that can accomodate moving IP packets on a pure
>> link level, arp-driven low-to-medium speed net, based on standard
>> leased/switched circuits ? The popularity of non-protocols like
>> SLIP should attest to this. Just something to carry IP frames
>> without costing a bundle."
>
> There is significant effort in this area. The protocol is called the
> Point-to-Point Protocol (PPP). It provides the features of SLIP but
> will run on either synchronous or asynchronous links. It provides
> options for negotiating IP addresses, compression, authentication, and
> encryption. I recommend that you pick up RFC's 1171 and 1172.
I am well aware of PPP and have tried hard to find out just what is
going on in this field. Unfortunately everything I have seen assumes
you want to let IP do all the routing, and just use point-to-point links
for interconnecting IP networks. I see nothing especially new about PPP,
except that is is a real protocol instead of the mostly empty protocol
layer SLIP provides. Some gear that multiplexes PPP would perhaps
be an answer, or we could use X.25 (shudder) or Frame Relay as the
network to put under IP for the local loop to the customer.
In <9103232308.AA12998@napa.telebit.COM> he also comments :
> The bottom line is that policy is likely to go out the window when
> people want badly enough to participate in the "internet." Formal
> rules and regulations tend to go out the window when enough people
> wish to operate counter to said rules and regulations (at least they
> do in this country -- no slight on any other country; it is just that
> I have the greatest experience with actions and reactions in the
> United States). Actually, dial-up IP is just a special case in the
> growth and proliferation of internetworking. The presence today of
> commercial carriers such as PSI and Alternet is just as significant.
> Given that this is going to happen, how does one structure a mechanism
> to adapt and coexist rather than partition, limit, and control.
>
> Hey folks, we need SIMPLER ways to manage internetworks, not more
> complex rulesets. THAT is the challenge: to simplify.
>
As I said above, there IS actually a set of functioning international
law on the subject on transportation that is quite easy to adapt to.
Requiring Internet (I.E: NSF) connectivity to obtain a domain would be
downright silly if it were not for the consequences; what does the
namespace registry have to do with your carrier aggreements ?
The frequency assignment of a radio station has absolutly nothing to
do with the equipment it purchases, nor should domains or internet
numbers have anything to do with the carriers you choose.
In <9103270633.AA09930@sloth.mlb.semi.harris.com> Jim Ray
<jdr@mlb.semi.harris.com> comments
> Definitly. I can think of no one that wants to pay an additional
> fee to be free of the commercial use restriction. I do believe that
> this is the beginning of a great movement though ( CIX ).
Well, my employer is about as commercial as you can get, so for us
it is either a net free of restrictions, or no net at all.
That should be a count of one. ;-)
Morten Reistad
System Designer
Oslo Stock Exchange Information <mrr@boers.uu.no> +47 2 341 700
