[Cluster-tech] Bad data in the system

[Lee Sawkins]

Hi Dirk.

The PC92 protocol goes a long ways towards eliminating bad data from the
cluster system.  However, there is still some bad data being sent
around.  Some of this is caused by Spider not verifying new connections
between clusters.  

As a minimum I would suggest that an exchange of PC20/22s should be done
and also a verification of callsigns before a connection is used.  As it
is now a cluster can change its call, say from 4N1EXY to YU1EXY-7.  If
another cluster can connect to the old IP address, it will report a
connection to 4N1EXY even though the other end said "*** Connected to
YU1EXY-7".  

Another related problem happens if a Spider cluster connects to say
VE7CC-1 and VE7CC-1 does not have the connecting call in its node
table.  The other cluster will report VE7CC-1 as a connected cluster
with no users.  This will override earlier broadcasts from the VE7CC-1
cluster which gives a user list.  VE7CC-1 will sometimes show with no
users and then with many users on other clusters.

Lee