[Dxspider-support] Curiosities about network topology

[Kin]

Hi Rudy,
The network is quite stable. There are many nodes that only reboot when there’s an update, and others never do :(
With the advent of fiber optics and the high bandwidths we now have, even in the worst-case scenario, a spot will reach us "instantaneously." I don’t think a few milliseconds will cause us to miss a contest or a DXpedition.
Regarding the data provided by N1MM about the nodes, I don’t know where it retrieves that information, but for the ones I manage, I can confirm the data is incorrect.
A real cluster (3 nodes) in HA, with load balancing, hosted at a site with UPS, several gigabits of bandwidth, and so on, shows availability values ranging from 90% to 94%, despite being monitored 24/7/365. Simply impossible.
 
73 de Kin
 
 
EA3CV-2 dxcluster.cronux.net 7300
http://clustermap.cronux.net/map.html
 
 
De: Dxspider-support <dxspider-support-bounces at tobit.co.uk> En nombre de Rudy Bakalov via Dxspider-support
Enviado el: martes, 3 de diciembre de 2024 11:59
Para: The DXSpider Support list <dxspider-support at dxcluster.org>
CC: Rudy Bakalov <r_bakalov at yahoo.com>
Asunto: Re: [Dxspider-support] Curiosities about network topology
 
Quantity vs quality of partners- when nodes run on beat effort basis, they could go down at any time.  On average, about 10% of my partners are down at any given time. It is also true that about 20% of them are rarely down, if ever.
 
I find the dynamic nodes list in N1MM quite useful as it shows 1/how often users connect to each cluster and 2/cluster availability as experienced by users. It confirms that there are too many clusters.
 
Rudy N2WQ
 
Sent using a tiny keyboard.  Please excuse brevity, typos, or inappropriate autocorrect.
 



On Dec 3, 2024, at 2:27 AM, IZ2LSC via Dxspider-support <dxspider-support at tobit.co.uk> wrote:

This topic is very close to what happened with BGP (Border Gateway Protocol) that is the routing protocol that manages all the internet routes (so very solid, well designed, scalable).

If we want to do something more structured we can take inspiration from BGP introducing the concept of Autonomous System, Route Reflector and Confederations.
 
73
 
Andrea
iz2lsc
 
 
-->
 
 
Il giorno lun 2 dic 2024 alle ore 22:13 Keith Maton via Dxspider-support <dxspider-support at tobit.co.uk <mailto:dxspider-support at tobit.co.uk> > ha scritto:
Yeah, we have too many nodes and the nodes we have, have many more partners than they need.  I’m as guilty of this as the next person.
 
If someone asks me to be a partner node, I find it very difficult to turn them down, in fact, I don’t think I’ve rejected anyones request.  If they’ve gone out of their way to ask me, they must have a reason for it so I think it’d be rude to reject it.  I do now insist that they connect to me rather than vice versa.
 
On the positive side it means we have a very robust system and huge chunks of it could fail but we’d all still get our spots, and we’d still get them quickly.  
 
73 Keith
 



On 2 Dec 2024, at 17:59, Kin via Dxspider-support <dxspider-support at tobit.co.uk <mailto:dxspider-support at tobit.co.uk> > wrote:
 
*snip*
 
Makes you think, doesn't it?
 
_______________________________________________
Dxspider-support mailing list
Dxspider-support at tobit.co.uk <mailto:Dxspider-support at tobit.co.uk> 
https://mailman.tobit.co.uk/mailman/listinfo/dxspider-support
_______________________________________________
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/20241203/76879989/attachment-0003.htm>