[CQ-Contest] Cluster network update - was: Reverse Beacon Network - too much success?

Fri Mar 5 04:23:23 PST 2010

> Yes, it's time to get control over DXCluster. This beast is out of
> control.
> Authentication system and registration license based.

> Why couldn't a DX operator just request that no spots be posted for them
> if
> they don't wish to be spotted.  Seems like it shouldn't be to hard to do.

> I would dare to say that more than 95% of dxcluster accesses come from
> internet,
> so, why not to have user/passwords? There are a couple of "open id"
> systems
> around. We could develop our own, or maybe elect an existing one. That
> would not
> be that hard to implement.

Same answer to all 3... the existing software and network implementation
won't handle it.  Remember, the existing network was derived from the rf
packetcluster network where by law encryption and security were prohibited.
So all the existing software has an open text backbone structure that allows
easy connection of net nodes to the network.  Because of this it is
extremely easy for new software developers to write software that is
compatible with the existing network.  This has resulted in a wide variety
of telnet, rf, and web based methods of accessing the network, some with
good filtering capabilities, some with authentication capabilities, but most
with poor filtering and no authentication.  

There is also a large world wide infrastructure, many independent nodes, and
a VERY large user base.  At this point I don't think anyone can even
estimate how many users may exist, let alone how many node computers may be
on the existing network running how many different software packages.
needless to say, this network will not be easy to modify, I would say
virtually impossible.  There is just too much investment in too widely
varied configurations by too many independent operators to switch it to
something else.  And all it takes is ONE open web cluster, or one unsecured
rf connection for someone to get in to disrupt the network.

The only way I can see the existing network going away is to design a new
system from the ground up that includes strong backbone security, not too
complicated but secure user identification, and all the other features that
everyone wants to enable opting in/out of spotting, filters, larger volumes
of spots, elimination of bottlenecks and loops, must run on various versions
of windows, linux, mac's, must include telnet, rf, web, and some kind of new
secure user access for future expansion, must include localized language
capabilities and built in translation of talk/announce/comments, use the new
non-ascii url system, accommodate skimmer spots automatically, allow Unicode
character sets, email and bulletin distribution, emergency disaster
overrides, etc, etc, etc.  And then of course all the user logging and other
access programs will have to update to handle whatever the authentication
system is.  This is obviously not a simple project... and it won't make the
old system go away until enough nodes are set up and enough users on the old
system see it as being more useful than the old network to abandon it.  If
the development started today collecting requirements I would estimate no
less than 5 years to the shutdown of the existing network... and that
assumes a fairly dedicated open development group.

Build a better network, and they will come.

David Robbins K1TTT
e-mail: mailto:k1ttt at arrl.net
web: http://www.k1ttt.net
AR-Cluster node: 145.69MHz or telnet://dxc.k1ttt.net

More information about the CQ-Contest mailing list