From Enric Lleal Serra@2:343/107.1 to Todos on Wed Aug 5 07:26:36 2009
* Originally in ESP.FUNDAM_FIDO
* Copiado en ESP.INTERNET
Un interesant°simo escrito de Michiel Van Der Vlist, el actual Administrador de
FTSC, aparecido en la £ltima FidoNews.
FidoNet and IPv6
By Michiel van der Vlist 2:280/5555
It seems to be a human trait, almost as if it is in the genes: to
grossly underestimate the effects of growth. Tom Jennings version 1 of
the BBS/mailer programme Fido had room for 1024 nodes. This was
quickly corrected by first adding nets, then zones and last but not
least: points, The fionet address space is 60 bits wide. There is room
for 1.15*10^18 or just over one trillion nodes. (In the European
number naming system.) As Fidonet is no longer growing, there is no
danger of Fidonet ever running out of address space. This contrary to
the InterNet. In the classic internet numbering system (IPv4) there is
room for 2^32 or just over four milliard addresses. Again using the
European number naming system. Note that FidoNet has a much larger
addressing space than the InterNet.
Not all of the mathematically available numbers in a 32 bit wide
number are available as public IP adresses. Some adressem have been
reserved for special purposes. Some 3.6 milliard addresses are
available for unique public addresses, Of these only 11% are now
unallocated and it is expected that the IANA will run our of their
pool of unassigned addresses within two years. That does not mean that
the last IPv4 address will be issued then, there is still the pools
administered by the five Regional Internet Registries, Their pools are
expected to last at least another nine month. These projections may be
inaccurate, but one thing is certain: the IPv4 numbers will run out in
a couple of years. This article is nopt abpout when exactly it will
happen, this is about the consequences for Fidonet when it happens.
The remarkable thing is that the InterNet community is so slow to take
action. The next generation standard, IPv6 was already drafted in 1990
or so but only now do we see some attempts to get ISP's and end users
interested. Well, better late than never, IPv6 is coming. In The
Netherlands major ISP Xs4ALL started offering native IPv6 to end
So how will this effect FidoNet? Maybe not at all. That the IPv4
addresses run out, does not mean that IPv4 will go out like a lamp. It
is not like the oil running out, and all cars coming to a halt. The IP
numbers that were assigned can still be used and The IPv4 network can
and will co-exist with the IPv6 network for years if not for decades
to come. It is just that no new IPv4 numbers will be issued. Maybe
FidoNet will die before any problem emerges. Maybe the depletion of
IPv4 numbers will only affect new sysops. Or maybe not...
I see two scenarios:
1) ISP's will go IPv6 and new users will only get IPv6 addresses. In
Europe there is some pressure from "above" to get moving with the
transition to IPv6. The European Commission has decreed that by 2010,
25% of all users must be IPv6 capable. I do not know if that target
will be reached, but at least there is some pressure. So ISP's will
probably migrate to IPv6 when their pool of IPv4 addresses for end
users runs out.
2) In the US and other countries where there is less pressure, ISP's
may choose a "solution" that requires less long term investment. They
may choose to no longer issue public addresses to their end users, but
put them behind a NAT and issue them IPv4 numbers in a private range
(10.x.x.x or 192.168.x.x). If that happens it will affect FidoNet big
time, because it will effectively stop end users from running servers.
Many ISP's do not like end users running servers and this might just
be the excuse to curb the practise. They may- or may not - offer a
public IP at extra cost.
The answer in both case is to go IPv6. It is a common misunderstanding
that users must wait for their ISP to go IPv6. But one does not have
to wait for the ISP. One can use an IPv6 over IPv4 tunnel. There are
dozens of so called "tunnel brokers" where one can get a tunnel to a
so called Point of Presence. Some do it for free, some ask a fee. All
"new" operating systems have IPv6 already build in. Windows from XP
on, Linux, and MAC-Oos have it, it just needs to be turned on. For
Window 2000 there is an add-on.
Getting access to the IPv6 infrastructure is not all that hard. But of
course that is not enough. The applications must also be IPv6 capable.
For the "popular" stuff like browsers and e-mail agents, that is no
problem. But what about the Fidonet stuff? Well, there is some work to
do. Of the known IP mailers none support IPv6 at present. We can
probably forget about the none open source stuff like Irex. Most of it
is abandonware, so there is no way to add IPv6 capability, The best
hope is probably for binkd as that is open source and so adding IPv6
capability should not be all that hard. Argus/Taurus/Radius may also
be an option. although the developers have abandoned it, the source is
available, so there is hope.
But if we can not uograde our FidoNet software to support IPv6 and
ISP's force us to use IPv6 on;y, there us still another trick: IPv4
over IPv6 tunneling. If every node instaal IPv4 over IPv6 tunnel
software, we can build our own virtual private IPv4 network and run
Fido over IP over that. It is a bit of a kludge, but it can be done if
the need arises.
Most important is that we regain the pioneer spirit that once was so
prevalent in FidoNet . There was a time when FidoNet was a leader in
network communication, instead of bunch of old farts in wheel chairs
thet just keeps rolling on inertia. BBSing and FidoNet were a driving
force behind the modem industry. We - the sysops - were pioneers. Let
us do it again. Let us see the conversion to IPv6 as a challenge.
Instaed of just waiting for the inevitable to happen, let us make an
active contribution. Let us be a the louse in the hide of the ISP's as
we once were for the telcos. Let us nag them to get moving with IPv6.
Let us make our systems available through IPv6 and so make a
contribution towards better global communication.