[Dxspider-support] A new version...

Dirk Koopman djk at tobit.co.uk
Thu Jun 21 22:32:29 CEST 2007


Brendan Minish wrote:
> The new 1.54 Version please.
> 
> Both my clusters run on linux under vmware server so if something goes
> pearshaped it's a dead simple job to restore a snapshot and go back to
> how things were.
> 
> Am I right in thinking that with the new protocol there are some real
> advantages having more than one or two peers?

Dunno about advantages. More no disadvantages other than the extra 
bandwidth, and the lack of routing loops.

> 
> How does the new routing work in comparison to the likes of OSPF, BGP or
> OLSR protocols used on lan and and networks?

It is vaguely OSPF like in that that only "local" configurations and 
changes are broadcast. Loops/spanning tree broadcasts are dealt with by 
deduping on a (PC9x, callsign, sequence no) tuple. There is also the 
concept that after three update periods and no routing broadcast, a node 
will be expunged from the routing table.

However, there are no HELLO packets as such and no explicit keepalive 
messaging.

> 
> I have quite a bit of experience dealing with dynamic routing on large
> networks, What tools are there for understanding and debugging the new
> routing?

watchdbg + mk 1 eyeball (and sh/newc)

> 
> Thanks for all the work that has gone into this.
> 

A pleasure.

Dirk G1TLH



More information about the Dxspider-support mailing list