Next to AMS-IX, or are you actually peering on AMS-IX? A valid consideration for future
POP locations should be internet exchange connectivity. A 100gig connection could be great
(if used) but if you want lower latency an IX will be much better than some large transit
provider, especially if you are connecting to a geographically local POP. The whole POP
proposal reads likes a CDN build but without mentioning the criticality of being peered on
exchange points. IXs serve many purposes but one of them is keeping local traffic local,
and thus latency "low". I only ask to bring it up as part of the discussion and
so I know at least someone else who is dealing with 44net prefixes and an IX 😉
-C
-----Original Message-----
From: 44Net <44net-bounces+colin.bodor=imperium.ca(a)mailman.ampr.org> On Behalf Of
Rob PE1CHL via 44Net
Sent: Tuesday, August 3, 2021 1:21
To: 44net(a)mailman.ampr.org
Cc: Rob PE1CHL <44net(a)pe1chl.nl>
Subject: Re: [44net] On Allocations, PoPs, and Proposals
Pierre,
You should not read too much into that remark that bandwidth is not the first concern.
I does not mean "we may as well have bad bandwidth and poor performance", but
rather that we are not using that much bandwidth that we need to have the fastest
connections and the fastest routers.
At the moment our 44.137.0.0/16 gateway has a 1Gbit internet connection but it on average
uses only about 30 Mbit/s with peaks of up to ~100 Mbit/s (in each direction).  At this
moment we are busy moving the servers to a different datacenter because the old one is
going to close down, and there we will get 10 Gbit/s but the new datacenter offers us up
to 100 Gbit/s. We have chosen not to go that way for now because our equipment has only
up to 10 Gbit/s interfaces and going higher would mean more investment. I think that is
not warranted at this time. Maybe if we would host a PoP in a worldwide backbone and
would get connections from more countries. But at that time it can always be changed.
This of course has no influence on latency, that is determined by physical location. We
will be next to the AMS-IX.
Rob
On 8/3/21 12:48 AM, pete M via 44Net wrote:
I think that performance IS of prime concerne as much
as reliability
and redundency (wich is the 2 face of a same problem) Reliability with lack of
performance will not fix our situation. adding hundreds of milisecond to a network
conection , even if stable is in no way an improvement.
Pierre
VE2PF
________________________________________
De : 44Net <44net-bounces+petem001=hotmail.com(a)mailman.ampr.org> de la
part de Rob PE1CHL via 44Net <44net(a)mailman.ampr.org> Envoyé : 2 août
2021 17:10 À : 44net(a)mailman.ampr.org Cc : Rob PE1CHL Objet : Re:
[44net] On Allocations, PoPs, and Proposals
Yes I think the design and deployment of the new backbone network
which would replace IPIP tunnels and would offload the mandatory task of (complex)
routing from the beginning user should be the first priority.
It would lower the bar of entry and it would also solve the issues that the recent TAC
proposal appears to target.
When budget is a problem, I think some of the existing gateways (certainly the one here
in the Netherlands) would be able
to offer the function or the hosting of that function for their region. We currently
have 1Gbit internet connection, soon
to be upgraded to 10Gbit, and a powerful system. Furthermore I think that bandwidth and
performance is not really
the first concern at the current load. Focus should be on
reliability, possibly through having enough redundancy in the network.
Rob
_________________________________________
44Net mailing list
44Net(a)mailman.ampr.org
https://mailman.ampr.org/mailman/listinfo/44net