[Dxspider-support] Curiosities about network topology

[Dirk Koopman]

What I would be interested in is the maximum and average time a PC92 
type message traverses the network. One is never going to get a 
granularity of less than one second, 'cos the time (as in 
PC92^RW3XA-8^*135*^...) is in seconds. And there are some issues with 
people not syncing their time, not to mention to remember that the time 
rolls over back to 0 at midnight UTC. Oh and, sometimes, the time value 
could be 135.1 or 135.7 etc which only indicates that more than one 
message was originated in that second from that node.

An extra little project for you :-) Although, arguably it could be 
something that is maintained by each node and can be inspected (by the 
sysop?) to (maybe) quash this annoying habit (to me at least) of users 
connecting to as many nodes as possible simply to "get an edge".

73 Dirk G1TLH


On 02/12/2024 17:59, Kin via Dxspider-support wrote:
> Hi,
>
> I was curious to know what would be the largest distance (hops) that could
> occur in our network with the current topology.
> I started with a graph that you can see here:
> http://clustermap.cronux.net/route_finder.html , and I ended up with a few
> lines of code :(
>
>  From the data obtained from the PC92C sentences, the following data can be
> inferred:
>
> Hops   Quantity    Nodes
> ----	--------	 -----
>     1	1782	  563
>     2	20003	  561
>     3	57289	  561
>     4	54012	  561
>     5	19279	  561
>     6	4001	  539
>     7	639	  256
>     8	67	    49
>     9	9	      7
>
> The 'Quantity' column counts pairs of nodes that have a route with N hops.
> The 'Nodes' column counts unique nodes involved in any path with N hops.
>
> It represents the minimum hops between nodes (all to all).
> It is based on a study of 563 nodes.
> For each node, the distance to all other nodes has been calculated using BFS
> (Breadth First Search).
>
> It has to be said, that it is possible (for sure) that there are other
> routes and nodes that could not be used in this study. The reason for this
> is the lack of information (existence) about them or their neighbouring
> nodes. Among them are ARCluster, DXNet, CLX, CCCluster, especially those
> nodes that are not directly linked to nodes using PC9x.
>
> Not even 10 hops are necessary to reach a node.
> We could infer that with 4 links on average, the 563 nodes would be at a
> maximum distance of 5 hops. [SUM(i=1 to 5) f(i)]
> Makes you think, doesn't it?
>
>
> 73 de Kin EA3CV
>
>
> EA3CV-2 dxcluster.cronux.net 7300
> http://clustermap.cronux.net/map.html
>
>
>
> _______________________________________________
> Dxspider-support mailing list
> Dxspider-support at tobit.co.uk
> https://mailman.tobit.co.uk/mailman/listinfo/dxspider-support
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.tobit.co.uk/pipermail/dxspider-support/attachments/20241203/c7fe1cdd/attachment-0001.htm>