• FSP-1038.001

    From Michiel van der Vlist@2:280/5555 to All on Mon Nov 4 22:24:32 2013
    * Originally in FTSC_PUBLIC
    * Crossposted in IPV6 **********************************************************************
    FTSC FIDONET TECHNICAL STANDARDS COMMITTEE **********************************************************************

    Publication: FSP-1038
    Revision: 1
    Title: The INO4 flag.
    Author(s): Michiel van der Vlist
    Date: 3 November 2013 ---------------------------------------------------------------------- Contents:
    1. Introduction.
    2. The problem.
    3. The solution. ----------------------------------------------------------------------

    1. Introduction
    ---------------

    The InterNet is running out of addresses. To solve this problem
    it is slowly migrating from 32 bit addresses (IPv4) to 128 bit
    addresses (IPv6).

    The conversion is largely transparent for the user. When querying
    the DNS system to resolve a host name, an IPv6 capable system
    initiating a connection will look if an AAAA record is present
    and use the IPv6 address if that is the case. If not, it will use
    the IPv4 address. There is no need for a nodelist flag to signal
    IPv6 capability, the lower layers will handle it transparently.


    2. The Problem
    --------------

    Presently all IPv6 capable nodes in the nodelist also support
    connectivity over IPv4. So there is downward compatibility. At
    present IPv4 only nodes have no problems connecting to IPv6 capable
    nodes, they just connect via IPv4.

    We can however not expect that situation to last forever. While the
    number of IPv6 capable nodes grows, IPv4 connectivity will eventu-
    ally degrade.

    Some day there will be nodes that no longer accept incoming connec-
    tions over IPv4. At best they will be behind a GGNAT or Carrier
    Grade NAT. That is a NAT operated by their ISP, so they will have no
    control over the port forwarding tables. At worst they will have no
    IPv4 at all. Either way, these nodes will not be reachable by nodes
    that do not have IPv6 capability.

    Attempts to connect will fail and most mailers will just repeat
    the attempt at regular intervals until the sysop intervenes. At
    present there is no way to predict such a failure beforehand.

    To avoid this undesirable situation I propose a flag that indicates
    that an otherwise IP capable node does not support incoming connec-
    tions over IPv4.


    3. The INO4 flag
    ----------------

    INO4 Indicates that an otherwise IP capable node is unable to
    accept incoming connections over IPv4.


    4. Considerations
    -----------------

    Although NOIPV4 would be more clear, it goes against the custom that
    all IP related flags start wit the letter 'I'. It is also needlessly
    long.
    Someone suggested to use IPV6 with the definition of "IPv6 only".
    This is bound to be misunderstood as other capability flags have
    always meant "in addition of". V34 means an additional V34 capabi-
    lity, not "V34 only". X75 does not mean "X75 only". Using IPV6 with
    a definition of "IPv6 only" will raise confusion and hence improper
    use .

    Note that in the very long run as IPv4 gradually fades out, we may
    reach a situation where most of the IP capable nodes in the nodelist
    will carry this flag. If and when this happens, the Fidonet communi-
    ty may consider dropping this flag and marking the minority that
    still accepts incoming connections over IPv4 with an IPV4 flag.


    References
    ----------

    [FTS-5] "The distribution nodelist", Ben Baker, Rick Moore.
    February 1989. Obsoleted by FTS-5000.

    [FTS-5000] "The distribution nodelist".
    FTSC Members, Administrator and Honoured Guests
    Dec 2012.

    [FTS-5001] "Nodelist flags and userflags".
    FTSC Members, Administrator and Honoured Guests
    March 2013.


    Contact Data
    ------------

    Michiel van der Vlist
    Fidonet: 2:280/5555
    E-mail: pa0mmv at vrza dot org
    E-mail: administrator at ftsc dot org

    History
    -------

    Rev.1, 20131103: Initial Release.
    Principal author Michiel van der Vlist.

    **********************************************************************


    --- GoldED+/W32-MINGW 1.1.5-b20110320
    * Origin: http://www.vlist.org (2:280/5555)
  • From Alexey Vissarionov@2:5020/545 to Michiel van der Vlist on Tue Nov 5 10:44:44 2013
    Good ${greeting_time}, Michiel!

    04 Nov 2013 22:24:32, you wrote to All:

    Publication: FSP-1038
    Revision: 1
    Title: The INO4 flag.

    As I already told, it's useless.

    When a system is IPv6-only, it may be listed using IP literal:

    ,9999,SysName,City,Sysop_Name,00-00-000000,300,CM,INA:[2001:0DB8::2:2:9999],IBN

    So, as the IPv6 addresses are unlikely to be allocated dynamically (and thus changed frequently), this may work as desired.

    One single not-really-a-problem raises only when the nodelist line contains INA:f.q.d.n flag, as the old mailers may be unable to detect EAI_ADDRFAMILY returned by getaddrinfo() or NO_ADDRESS returned by archaic gethostbyname()

    Once faced, that may be solved by either:
    1. Updating the mailer.
    2. Checking that manually.

    So leaving it up to sysops seems to be the best solution. And, of course, we should recommend checking for EAI_ADDRFAMILY to the developers. At least here:

    * Originally in FTSC_PUBLIC
    * Crossposted in IPV6

    --
    Alexey V. Vissarionov aka Gremlin from Kremlin
    gremlin.ru!gremlin; +vii-cmiii-ccxxix-lxxix-xlii

    ... :wq!
    --- /bin/vi
    * Origin: http://openwall.com/Owl (2:5020/545)
  • From Ward Dossche@2:292/854 to Alexey Vissarionov on Tue Nov 5 09:28:02 2013
    Alexey,

    Publication: FSP-1038
    Revision: 1
    Title: The INO4 flag.

    As I already told, it's useless.

    I concur.

    Further to that, I think that if/when IPv6 really becomes an issue, a sufficient number of people will abandon not to kill Fidonet but to make it grind to a de-facto standstill. Avian IP may be faster at that time.

    Let's all move to smartphones ...

    \%/@rd

    --- D'Bridge 3.96
    * Origin: Many Glacier / Protect - Preserve - Conserve (2:292/854)
  • From Kees van Eeten@2:280/5003.4 to Ward Dossche on Tue Nov 5 18:27:00 2013
    Hello Ward!

    05 Nov 13 09:28, you wrote to Alexey Vissarionov:

    Further to that, I think that if/when IPv6 really becomes an issue, a sufficient number of people will abandon not to kill Fidonet but to make it grind to a de-facto standstill. Avian IP may be faster at that time.

    Let's all move to smartphones ...

    Yes, with fidonet packets packed in sms messages ;)

    Kees

    --- FPD v2.9.040207 GoldED+/LNX 1.1.5
    * Origin: As for me, all I know is that, I know nothing. (2:280/5003.4)
  • From Markus Reschke@2:244/1661 to Alexey Vissarionov on Tue Nov 5 18:59:36 2013
    Hello Alexey!

    Nov 05 10:44 2013, Alexey Vissarionov wrote to Michiel van der Vlist:

    Title: The INO4 flag.

    As I already told, it's useless.

    When a system is IPv6-only, it may be listed using IP literal:

    ,9999,SysName,City,Sysop_Name,00-00-000000,300,CM,INA:[2001:0DB8::2:2 :9999],IBN

    It could be useful for IPv4-only nodes if the INO4 flag is used by a ruleset in the mailer configuration, for example:

    if IBN & !INO4 : "call using binkp"

    The node might running dual stack inside the LAN but only got IPv4 Internet access.

    Regards,
    Markus

    ---
    * Origin: *** theca tabellaria *** (2:244/1661)
  • From Stas Degteff@2:5080/102.1 to Markus Reschke on Wed Nov 6 03:33:08 2013
    Hello Markus!

    05 Nov 13 18:59, you wrote to Alexey Vissarionov:

    Title: The INO4 flag.

    As I already told, it's useless.

    When a system is IPv6-only, it may be listed using IP literal:

    ,9999,SysName,City,Sysop_Name,00-00-000000,300,CM,INA:[2001:0DB8:
    :2:2
    :9999],IBN

    It could be useful for IPv4-only nodes if the INO4 flag is used by a ruleset in the mailer configuration, for example:

    if IBN & !INO4 : "call using binkp"

    The node might running dual stack inside the LAN but only got IPv4 Internet access.

    FYI, any node with ipv4 MAY connect to IPv6 world using "6to4" tunnel or teredo or tunnel brocker. IPv6 on my node connected via 6to4 (address in the subnet 2002::)


    Stas
    Jabber-ID: grumbler@grumbler.org
    GPG key 0x72186DB9 (keyserver: hkp://wwwkeys.eu.pgp.net)

    --- GoldED+/LNX 1.1.5-b20130515
    * Origin: Grumbler at home (2:5080/102.1)
  • From Markus Reschke@2:244/1661 to Stas Degteff on Wed Nov 6 17:48:12 2013
    Hello Stas!

    Nov 06 03:33 2013, Stas Degteff wrote to Markus Reschke:

    The node might running dual stack inside the LAN but only got IPv4
    Internet access.

    FYI, any node with ipv4 MAY connect to IPv6 world using "6to4" tunnel
    or teredo or tunnel brocker. IPv6 on my node connected via 6to4
    (address in the subnet 2002::)

    Emphasis on "may" ;-) How much nodes are able to manage IPv6 transition mechanisms? And we shouldn't rely too much on Teredo since it's future is quite uncertain.

    Regards,
    Markus

    ---
    * Origin: *** theca tabellaria *** (2:244/1661)
  • From Michiel van der Vlist@2:280/5555 to Stas Degteff on Thu Nov 7 00:21:20 2013
    Hello Stas,

    On Wednesday November 06 2013 03:33, you wrote to Markus Reschke:

    It could be useful for IPv4-only nodes if the INO4 flag is used
    by a ruleset in the mailer configuration, for example:

    if IBN & !INO4 : "call using binkp"

    The node might running dual stack inside the LAN but only got
    IPv4 Internet access.

    FYI, any node with ipv4 MAY connect to IPv6 world using "6to4" tunnel

    In the case of a dual stack LAN, as Markus mentioned, the 6to4 tunnel endpoint must be in the router. Not all routers support that. well, you can change the router. But also.. A 6to4 tunnel requires a (quasi)static IPv4 address for the tunnel end point. That is something not everyone has and can change.

    or teredo or tunnel brocker.

    Teredo sucks...

    IPv6 on my node connected via 6to4 (address in the subnet 2002::)

    So does my tunnel (2001:: subnet)

    But what works for you and me, may not work for all.


    Cheers, Michiel

    --- GoldED+/W32-MINGW 1.1.5-b20110320
    * Origin: 2001:470:1f15:1117::1 (2:280/5555)
  • From Michiel van der Vlist@2:280/5555 to Ward Dossche on Thu Nov 7 00:29:11 2013
    Hello Ward,

    On Tuesday November 05 2013 09:28, you wrote to Alexey Vissarionov:

    As I already told, it's useless.

    I concur.

    Further to that, I think that if/when IPv6 really becomes an issue,

    IPv6 already IS an issue. APNIC end RIPE ran out of IPv4 addresses in 2011 and 2012. There already are places in the world where end users no longer get a globally routable IPv4 address. They get a RFC-1918 or RFC-6598 address. And it is not far away in the east, it happens right here. Today my laptop got 100.106.202.243 on its 3G dongle. That is not a globally routable address, I can not run a server on that address.

    It has not affected Fidonet yet, but I expect Fidonet to live long enough to see it happen.

    a sufficient number of people will abandon not to kill Fidonet but to
    make it grind to a de-facto standstill.

    Predicting the future is notoriously difficult and error prone. I once predicted Fidonet would be dead on 1 January 2001. Obviously I was wrong. You say IPv6 will not need be an issue for Fidonet. Let me remind you of this:

    "We do not need a telephone system, we have an excellent system of messenger boys".

    "640K ought to be enough for anybody"

    "There will be a need world wide for at most five computers".

    Let's all move to smartphones ...

    It won't be Fidonet...


    Cheers, Michiel

    --- GoldED+/W32-MINGW 1.1.5-b20110320
    * Origin: 2001:470:1f15:1117::1 (2:280/5555)
  • From Michiel van der Vlist@2:280/5555 to Kees van Eeten on Thu Nov 7 01:27:29 2013
    Hello Kees,

    On Tuesday November 05 2013 18:27, you wrote to Ward Dossche:

    Let's all move to smartphones ...

    Yes, with fidonet packets packed in sms messages ;)

    Hineously expensive on a cost per bit basis.


    Cheers, Michiel

    --- GoldED+/W32-MINGW 1.1.5-b20110320
    * Origin: 2001:470:1f15:1117::1 (2:280/5555)
  • From Ward Dossche@2:292/854 to Michiel van der Vlist on Thu Nov 7 02:35:01 2013
    Michiel,

    Let's all move to smartphones ...

    It won't be Fidonet...

    Within a decade, the PC as we know it today will have gone.

    As a scientist you know that technology renewals come faster and faster, well the next-one is overdue.

    \%/@rd

    --- D'Bridge 3.96
    * Origin: Many Glacier / Protect - Preserve - Conserve (2:292/854)
  • From Michiel van der Vlist@2:280/5555 to Markus Reschke on Thu Nov 7 11:40:11 2013
    Hello Markus,

    On Wednesday November 06 2013 17:48, you wrote to Stas Degteff:

    FYI, any node with ipv4 MAY connect to IPv6 world using "6to4"
    tunnel or teredo or tunnel brocker. IPv6 on my node connected via
    6to4 (address in the subnet 2002::)

    Emphasis on "may" ;-) How much nodes are able to manage IPv6
    transition mechanisms?

    I expect my mother and my neighbour across the street to not be able to set up a tunnel, but then they are not Fidonet sysops. Setting up a Fidonet node requires over above average ability to deal with computers and so I expect that quit a lot of Fdionet sysops will be able to manage.

    And we shouldn't rely too much on Teredo since it's future is quite uncertain.

    Indeed. It relies on third parties to provide the servers and relays. Plus that - correct me if I am wrong - it is my understanding that Teredo does not work from behind double NAT.

    But wait.... We do not need third parties to provide gateways. In Fidonet we already have a mechanism to get messages from one system to another if they can't connect directly. It is called routed mail. Any dual capability node can act as a gateway between IPv4 only and IPv6 only nodes.


    Cheers, Michiel

    --- GoldED+/W32-MINGW 1.1.5-b20110320
    * Origin: 2001:470:1f15:1117::1 (2:280/5555)
  • From Markus Reschke@2:244/1661 to Michiel van der Vlist on Thu Nov 7 14:10:04 2013
    Hello Michiel!

    Nov 07 11:40 2013, Michiel van der Vlist wrote to Markus Reschke:

    MvdV> relays. Plus that - correct me if I am wrong - it is my
    MvdV> understanding that Teredo does not work from behind double NAT.

    AFAIK it doesn't support symmetric NAT. So double NAT might work based on the NAT methods used by the firewalls/routers.

    MvdV> But wait.... We do not need third parties to provide gateways. In
    MvdV> Fidonet we already have a mechanism to get messages from one system
    MvdV> to another if they can't connect directly. It is called routed
    MvdV> mail. Any dual capability node can act as a gateway between IPv4
    MvdV> only and IPv6 only nodes.

    I see another nodelist flag coming ... IGW :-)

    Regards,
    Markus

    ---
    * Origin: *** theca tabellaria *** (2:244/1661)
  • From Benny Pedersen@1:261/38.20 to Kees van Eeten on Fri Nov 8 03:36:46 2013
    Hello Kees!

    05 Nov 2013 18:27, Kees van Eeten wrote to Ward Dossche:

    Let's all move to smartphones ...
    Yes, with fidonet packets packed in sms messages ;)

    no mms so we see who sender is :)

    quite nice if aftershork/hotdoged takes this up to support mms file areas, lol


    Regards Benny

    ... there can only be one way of life, and it works :)

    --- Msged/LNX 6.2.0 (Linux/3.11.6-gentoo (i686))
    * Origin: duggi.junc.org where qico is waiting (1:261/38.20)
  • From Benny Pedersen@1:261/38.20 to Michiel van der Vlist on Fri Nov 8 03:42:20 2013
    Hello Michiel!

    07 Nov 2013 00:29, Michiel van der Vlist wrote to Ward Dossche:

    MvdV> "640K ought to be enough for anybody"

    commodore 128d with reu1750 and microsoft basic 7.x who needs more ?

    MvdV> "There will be a need world wide for at most five computers".

    there was one million commodore 64 sold in ussr

    Let's all move to smartphones ...
    MvdV> It won't be Fidonet...

    maybe, maybe not, now that firms think about make smartphones brikede, put a 3com modem in it, and hook up your old msdos software in it :)

    motoroling always :=)


    Regards Benny

    ... there can only be one way of life, and it works :)

    --- Msged/LNX 6.2.0 (Linux/3.11.6-gentoo (i686))
    * Origin: duggi.junc.org where qico is waiting (1:261/38.20)
  • From Stas Degteff@2:5080/102.1 to Markus Reschke on Sat Nov 9 17:58:36 2013

    Hello Markus!

    06 Nov 13 17:48, you wrote to me:

    * Forwarded from area 'ipv6'
    Hello Stas!

    Nov 06 03:33 2013, Stas Degteff wrote to Markus Reschke:

    The node might running dual stack inside the LAN but only got
    IPv4 Internet access.

    FYI, any node with ipv4 MAY connect to IPv6 world using "6to4"
    tunnel or teredo or tunnel brocker. IPv6 on my node connected via
    6to4 (address in the subnet 2002::)

    Emphasis on "may" ;-) How much nodes are able to manage IPv6
    transition mechanisms? And we shouldn't rely too much on Teredo since
    it's future is quite uncertain.

    1. MAY is the term of FTA-1006.
    2. *Any* node, *ANY* (in this line "ANY" in emphasis and you may look this if your FTN reader tuned).
    I listed "teredo" and more 2 variants: "6to4" (required direct ipv4 connection or tuning of router), "Tunnel brocker".
    If somebody can't setup any one of three on your opinion - I don't doctor for it :-D

    Stas
    Jabber-ID: grumbler@grumbler.org
    GPG key 0x72186DB9 (keyserver: hkp://wwwkeys.eu.pgp.net)

    --- GoldED+/LNX 1.1.5-b20130515
    * Origin: Grumbler at home (2:5080/102.1)