Subj : finaly for me
To   : all
From : Markus Reschke
Date : Sat Mar 17 2018 16:21:52

Hi!

Mar 17 10:37 2018, mark lewis wrote to Alexey Vissarionov:

 BF>>>> Thanks to the FidoWeb yes, but still not the most efficient path
 BF>>>> from within R20.

When you deploy dupe rings the redundancy matters. Anyway, it's still FIFO, i.e. the fastest path wins even if it's longer than another one. The echomail traveling the slower path ends up in the dupe area.

 ml>>> there is no provision in anything FTN for figuring out and following
 ml>>> the fastest path from one system to another

 AV>> Most of the Fidonet doesn't need that.

Something like dynamic routing for netmail would be interesting.

 ml>>> so there's no way to figure any sort of ""efficient path"" from
 ml>>> anywhere to anywhere else...

 AV>> There is at least one quite similar to OSPF, but documenting it would
 AV>> take a whole eternity - first translate the "brief" description from
 AV>> Russian, after that write an FSP... just for 10...20 more nodes to
 AV>> have a bit faster netmail delivery?

Could you please elaborate on that?

 ml> we're talking about echomail, though... but yes, it isn't worth it to 
 ml> try to do anything to figure it out... documenting not withstanding... 
 ml> it would require non-mail traffic and a lot of probing of mailers and 
 ml> tossers to somehow pull routing information that can be consolidated 
 ml> into some sort of a map... that was painful back in the day when it 
 ml> was done and many refused to provide such mappings because it is none 
 ml> of anyone else's business where one's mailer(s) connect for mail 
 ml> transfers...

The internet wouldn't work with that attitude. But there are filters for hiding your secrets ;)

I think we should move this discussion to another echo.

ciao,
Markus

--- 
 * Origin: *** theca tabellaria *** (2:240/1661)

.