Subj : Rhenium Reports
To : All
From : Ozz Nixon
Date : Tue Jul 10 2018 10:29 am
Those are 3 of the 7 reports Rhenium mailer generated. Every couple thousand
seconds Rhenium randomly polls differnet systems in the nodelist. Mainly to
help me verify my nodelist indexer is working, and that Rhenium can negotiate
with each node. Then every monday, it resamples every system that it was able
to connect to and generates these reports:
NOT5D.TXT - list of systems that are presenting a 3D or 4D address without the
domain (if I recall its just two nodes).
NETLIST.TXT - list of all domains in the 5D addresses, no counts just a list of
every unique FTN 5D domain.
NETHOG.TXT - list of the top 10 nodes that have the most networks in their
M_ADR header.
CONNSPEED.TXT - Top 30 fastest connections and Top 30 slowest connections
BOSSOS.TXT - Top Operating Systems (per Mailer M_NUL VER string).
BOSSAPP.TXT - Top Mailers based upon M_NUL VER string. (currently BinkD),
modified this morning to also break out systems running binkp/1.0, binkp1.1 and
unspecified. (which is not compliant!)
BINKPEXT.TXT - Which M_NUL OPT extensions are presented network wide. No one
else is supporting SHA1. Only one system is running CRAM with a 1024bit public
key.
Is there anything else that would be useful to collect and report? I am
debating about adding a compare of M_NUL NDL versus nodelist FLAGs - reporting
variances. I also noticed a couple systems report the M_NUL TIME in non-RFC
compliant format. I know authors get their feathers ruffled, but, we could use
the data collected to help enforce standards (ducking from the flying tomato(s)
on that topic)... List Argus/Radius/Taurus are not compliant in the fact they
present every Command wih ^@ (nul terminator), but the specifications do not
mention that. Plus the A/R/T systems put ^@^@ after the last bytes of a file
transfer... not in the specifications (or I overlooked that), in the specs I
recall it mentioning that some systems *may* send an empty command to denote
eof, but that would be <80><0><0> in hex, not <0><0>. Of course we all worked
around this and tested against Argus. Then another topic of non-compliance is
jNode and Internet Rex have space in their M_NUL VER string, and per
specification the only space is supposed to be between product/version/info
binkp/1.0 to desinate product versus binkp core capabilities. (okay, bring the
flying tomatoes...)
