29 Dec
2020
29 Dec
'20
10:49 p.m.
Marius, I dont know where you got the idea you need to pay hundreds of $ a month for a vps
that do the BGP annonce of a /24 block.
My VPS is costing me 3.50$US a month and it is in New Jersey, I have a latency of 4-6 ms
when connected to it.
Pierre VE2PF
De : Marius Petrescu via 44Net<mailto:44net@mailman.ampr.org>
Envoyé le :29 décembre 2020 17:55
À : AMPRNet working group<mailto:44net@mailman.ampr.org>
Cc : Marius Petrescu<mailto:marius@yo2loj.ro>
Objet :Re: [44net] ipencap routing question -> What about the future ?
I think your xlx example is not the best for the use case of 44net
addresses.
For efficient routing and high speed and low latency as required for
voip systems, it is much more efficient to do a direct connection to the
server on a public interface instead of doing it the 44net way unless
the server sits on a BGP announced subnet.
In our current configuration, why would someone want to go client (let's
say me in YO) -> ampr-gw (in the US) -> xlx server(again YO), with a
350msec turnaround, if I can do it directly, with a 4msec latency?
Yes, a 44 BGP network would do the trick, but I am certainly not willing
to pay hundreds of USD per month for such an endeavor. BGP peering is
not cheap and not readily available in the whole wide world unless it is
not piggy backed on another preexisting AS for a select few working in
the network business.
And even in this case, it is of no use for the client to have a 44net
address, since it will again need to go the ampr-gw route to take
advantage of that IP, while masquerading to the local gateway IP would
yield better results.
We should look into other things, not available or not possible on the
regular internet, and voip and regular services are not it.
Marius, YO2LOJ
On 29.12.2020 22:21, Toussaint OTTAVI via 44Net wrote:
Le 29/12/2020 à 20:48, Rob Janssen via 44Net a écrit :
_________________________________________
44Net mailing list
44Net(a)mailman.ampr.org
https://mailman.ampr.org/mailman/listinfo/44net
Such images can be made as a by-product of a new
network design, but
please
understand that the main objective of the network changes should be that
such special images should become unnecessary to get a working
network connection.
I agree with you. But when I talk with some friends that are "basic"
hams (ie, they are interested in radio topics, but they are not IT or
network experts), I often hear the same thing :
"Why on earth should I bother with AMPR addressing ? It just works
fine over Internet".
In a previous discussion, I looked at the XLX hosts database, and
extracted the 44.x addresses among all Internet addresses. I don't
remember the exact percentage, but it was very low. Nobody is using
AMPR addressing for VoIP / digital modes. And those are not users,
those are sysops of repeaters and gateways (ie, people with skills
above average). Percentage for users would be even lower.
--
Providing RPi images is a convenient way to make complex tools
available for the masses. And it works ! There are dozens of
distributions available, for various RPi clones, and for various
applications. None of them do need an AMPR address to work. You just
plug it on your Internet box.
Of course, the best way of doing things would be to add a router with
AMPR logic inside (such as your Mikrotik, or our OpenWRT "TKBox")
between the "application" and the "network". That's what a
network
engineer would do. But why would a basic user do that ? What
additional feature will it bring ?
If we have a world-wide standardized description of an AMPRNet
tunneling protocol, developers such as MW0MWZ (dev of Pi-Star) could
add the tunnel into the distro, and provide a field in the GUI where
the user would enter the FDQN of its preferred AMPR POP (in the same
way he enters the IP of his BM Master). Of course, this would not be
the best way of doing things (= a tiny router). But this would allow
real plug-and play, and this would allow a lot of people to discover
AMPR addressing. Doing so, I think we would gain a lot of users. And
they could later move to a router-based configuration once they
understand the advantages of doing it.
If we keep only the "network engineer" approach, with a dedicated
router, my fear is we won't be attractive enough for the masses...
73 de TK1BI
_________________________________________
44Net mailing list
44Net(a)mailman.ampr.org
https://mailman.ampr.org/mailman/listinfo/44net