Slimelite BBS

How a new nodelist would work.....

Back to FREE25

From Jason Steck · 1996-04-07 18:09 · #73008 in FREE25

Saturday April 06 1996 02:20, Paul Edwards wrote to Jason Steck:

 PE> I think I should be able to "declare" somewhere, that I am happy to
 PE> route, at no cost, data from any incoming fidonet node to any Sydney
 PE> downlink.

What I am contemplating is a way to integrate any/all routing information into the nodelist at the node level.  Each node record would have a definition for what systems they can route TO and what upstream system(s) can provide a route to them.

By examining the chains provided by this type of information, an application would be able to determine the least-cost way to send the message which would place it in a chain by which it would reach the destination.

Basically, this scheme would provide routing tables in the nodelist and thus remove the reliance on user-defined routing tables which has proven to be the achilles heel of the existing routing system.

 PE> that link, whenever they see an opportunity.  Maybe one of my
 PE> downlinks, say 3:711/934.13, is happy to route at no cost, any mail
 PE> for 13:100/*@othernet.  Put them together, and basically a Sydney
 PE> fidonet person can probably get their mail to a Sydney othernet node.

Even moreso, this type of system would remove artificial and counterproductive barriers between networks altogether.  Basically, anyone and everyone who subscribed to the Node Contact System (NCS) would be an instant gateway all by themselves with access to any/all information about other gateways and routing arrangements.

 PE> Do you actually have a problem you are trying to solve here?  You need
 PE> to state what the problem is before you start coming up with elaborate
 PE> solutions.

Several problems.  One of the greatest strengths that the internet has over current ftn technology is the way you can send a message to almost any internet-style address and you don't have to CARE if your local system knows the location or route to that end system.

The NCS would give ftn technology a comparable capability which would be maintainable through the nodelist service itself (thus removing reliance on user-defined tables which are usually obsolete or incorrect).

The net result of this would very possibly be an explosion of VIABLE so-called "othernets".  Fidonet proper would become a kind of ARPANET -- still useful, but not nearly as useful as the internet creature that it spawned.  It and its political structures would become one among many -- the technology would be freed from the shackles imposed by a single network.

I wish I had resources to draw a picture of what I am talking about in greater detail.  This whole idea makes alot more sense in a dataflow format.  Maybe I'll create a graphic and make it available for file request and FTP.

Basically, the two major problems I target are:

1)  The utter unreliability and inefficiency of existing routed netmail.  Even many /0 nodes cannot/will not reliably route inbound mail and direct use of /0 inbound routing STILL requires a long-distance call.  The routing system which follows echomail paths is much more efficient, but suffers from massive areas where the service is not available at all (due to many sites refusing to provide the service) or where it has become completely unreliable due to the presence of outdated software and obsolete user-neglected routing tables.
    A large part of this problem is caused by large scale lack of knowledge of and attention paid to the question of netmail routing.  This problem would be solved by putting the routing information in the nodelist itself, thus removing both problems of individual sysop's need for extensive routing education and the maintainence of user-defined routing tables on each system in EVERY potential routing chain.

2)  The problems in reaching other networks.  Right now, you need to have either a physical copy of that other network's nodelist or you need to have manually configured a route to someone else who does have that kind of access. In the NCS, the nodelist itself would contain the information provided directly by the gateway system(s) themselves.  Furthermore, the nodelist would contain submissions from many of those "othernets" so even more direct routing might be possible other than single-gateway routing.

Jason
Internet: tutakai@oasis.novia.net



--- JMail-G 2.80d
 * Origin: PROZ Software -- JMail/JMQWK (FIDONET 1:285/424)