Improving Relay IPv6 - RIPE Grant
Dear relay operators, I just wanted to let you know that RIPE has announced funding for The Tor Project to improve IPv6 support on relays. (RIPE is the European internet infrastructure organisation.) https://www.ripe.net/support/cpf/funding-recipients-2019 We'll have more details early in 2020, when we've worked out an implementation plan and a start time. Thanks for your patience with our current IPv6 support. And thanks to all those volunteer coders who have worked hard to get us this far. T -- teor ----------------------------------------------------------------------
* teor@riseup.net:
I just wanted to let you know that RIPE has announced funding for The Tor Project to improve IPv6 support on relays.
That's great news, congratulations. My Tor nodes already use IPv6, so if you need help with testing updated Tor versions, please let me know. -Ralph
Great, than lets drop all IPv4 only relays from consensus 2020 finally. P.S. whitelist me so it doesn't take days till my emails appear on the mailing list. Thank you On 11.12.2019 03:20, teor wrote:
Dear relay operators,
I just wanted to let you know that RIPE has announced funding for The Tor Project to improve IPv6 support on relays. (RIPE is the European internet infrastructure organisation.)
https://www.ripe.net/support/cpf/funding-recipients-2019
We'll have more details early in 2020, when we've worked out an implementation plan and a start time.
Thanks for your patience with our current IPv6 support. And thanks to all those volunteer coders who have worked hard to get us this far.
T
-- teor ----------------------------------------------------------------------
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
On 2019-12-12 17:49:22, "NOC" <tor@afo-tm.org> wrote:
than lets drop all IPv4 only relays from consensus 2020 finally.
I would be sad to no longer be able to contribute to the Tor network. My ISP, Telenor Sweden, does not provide IPv6 and have no (public) roadmap for supporting IPv6. I can't switch ISP since they provide the fiber connection for the apartment building.
On 12/12/19, Logforme <m7527@abc.se> wrote:
My ISP ... does not provide ... and has no roadmap I can't switch ISP since they provide the fiber connection for the apartment building.
Seems you should be building out your own P2P fiber mesh guerrilla network house-to-house owner-to-owner, each node independantly peering up with others... don't rent the internet monthly... be the internet. Take less than 2 years of ISP fee, invest upfront into independant P2P network node, easily gets 10+ years P2P operations in return. Did you know your central ISP spies on you, allows other centrals to spy on you, gives all your data away to big centrals. Distributed private P2P network is much harder to do that, especially when every link is independantly privately owned and fulltime link encrypted 1 or 10 GigE, and everyone is running overlays within. Go do it.
Hi,
On 13 Dec 2019, at 08:45, Logforme <m7527@abc.se> wrote:
On 2019-12-12 17:49:22, "NOC" <tor@afo-tm.org> wrote: than lets drop all IPv4 only relays from consensus 2020 finally.
I would be sad to no longer be able to contribute to the Tor network. My ISP, Telenor Sweden, does not provide IPv6 and have no (public) roadmap for supporting IPv6. I can't switch ISP since they provide the fiber connection for the apartment building.
We won't be disabling IPv4 on relays any time soon. The RIPE grant covers IPv6 address autodetection and self-testing. If the feature is reliable enough, we may turn on IPv6 on dual-stack relays by default. (When autodetection and self-testing both pass.) We don't have any plans to disable IPv4 on relays. We'd need most relays to be dual-stack first. (Or we'd need research about privacy in non-clique networks.) And we'd need to write code that allows relays to turn off IPv4. When that's all deployed, we would have to make an engineering decision about the capacity of current IPv4-only relays, and the potential capacity of IPv6-only relays. One possible transition strategy is to allow IPv6-only bridges and exits. But to do that, we need more dual-stack guards and middles. That's why we are improving support for dual-stack relays with this funding. T
than lets drop all IPv4 only relays from consensus 2020 finally.
Someone may have mentioned already... Given many places relays do and could run are still IPv4 only, that would probably impact diversity quite a bit regarding AS, regions, jurisdiction, ISPs, datacenter vs network edge type of operators, etc.
* NOC:
than lets drop all IPv4 only relays from consensus 2020 finally.
Let's not. Availabiliy of IPv6 varies with country/territory/ISP. While I personally know people who simply don't want to use IPv6 (and I keep prodding them for it), there are others who simply don't have the option. -Ralph
I am unable to see why this is a good idea actually. As someone said already, I too would be unable to provide bridges and relays thanks to my ISP (Telia) does not provide an IPv6. But removing IPv4 only nodes from the network - you are basically removing a large chunk of relays from the network without any real reason. On 12/12/2019 5:49 PM, NOC wrote:
Great,
than lets drop all IPv4 only relays from consensus 2020 finally.
P.S. whitelist me so it doesn't take days till my emails appear on the mailing list.
Thank you
On 11.12.2019 03:20, teor wrote:
Dear relay operators,
I just wanted to let you know that RIPE has announced funding for The Tor Project to improve IPv6 support on relays. (RIPE is the European internet infrastructure organisation.)
https://www.ripe.net/support/cpf/funding-recipients-2019
We'll have more details early in 2020, when we've worked out an implementation plan and a start time.
Thanks for your patience with our current IPv6 support. And thanks to all those volunteer coders who have worked hard to get us this far.
T
-- teor ----------------------------------------------------------------------
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
I see a great benefit here, you could default to IPv6 for everything and enable IPv4 only as fallback that would allow to run Relays behind carrier grade NAT. I can get here 10 GBit/s symmetric but because the ISP did not get it's IP space 30 years ago (like some other ISPs which see no need for IPv6) it does have much more customers than IPv4 addresses, so running a Relay is not possible. And this trend will not shift back, you will see more and more ISPs which won't offer a public IPv4 address, simply because they can't get more IPv4 space. You see that even on the VServer market today where you can get very cheap Servers with a /64 but every IPv4 costs extra. So loosing some relays which really can't enable IPv6 is for me worth the trade off for many high speed relays which could finally join the network. On 16.12.2019 12:25, Jonathan Sélea wrote:
I am unable to see why this is a good idea actually. As someone said already, I too would be unable to provide bridges and relays thanks to my ISP (Telia) does not provide an IPv6. But removing IPv4 only nodes from the network - you are basically removing a large chunk of relays from the network without any real reason.
On 12/12/2019 5:49 PM, NOC wrote:
Great,
than lets drop all IPv4 only relays from consensus 2020 finally.
P.S. whitelist me so it doesn't take days till my emails appear on the mailing list.
Thank you
On 11.12.2019 03:20, teor wrote:
Dear relay operators,
I just wanted to let you know that RIPE has announced funding for The Tor Project to improve IPv6 support on relays. (RIPE is the European internet infrastructure organisation.)
https://www.ripe.net/support/cpf/funding-recipients-2019
We'll have more details early in 2020, when we've worked out an implementation plan and a start time.
Thanks for your patience with our current IPv6 support. And thanks to all those volunteer coders who have worked hard to get us this far.
T
-- teor ----------------------------------------------------------------------
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
* NOC:
I see a great benefit here, you could default to IPv6 for everything and enable IPv4 only as fallback [...]
Preferring IPv6 over IPv4 is not even remotely the same as your original call to "lets drop all IPv4 only relays from consensus 2020 finally", as you wrote in message <08fee42f-f45d-a8c4-d4e6-c83c05b4fb08@afo-tm.org>. Dropping support for IPv4 nodes would be counterproductive and is not going to happen in the forseeable future, as was clarified here before by teor. -Ralph
On 16.12.2019 20:37, Ralph Seichter wrote:
* NOC:
I see a great benefit here, you could default to IPv6 for everything and enable IPv4 only as fallback [...] Preferring IPv6 over IPv4 is not even remotely the same as your original call to "lets drop all IPv4 only relays from consensus 2020 finally", as you wrote in message <08fee42f-f45d-a8c4-d4e6-c83c05b4fb08@afo-tm.org>.
Read the full sentence, dropping all IPv4 only relays is unavoidable because they can't reach relays behind CG-NAT, preferring IPv6 over IPv4 is only a handy side effect of this.
Hi,
On 18 Dec 2019, at 02:17, NOC <tor@afo-tm.org> wrote:
On 16.12.2019 20:37, Ralph Seichter wrote:
* NOC:
I see a great benefit here, you could default to IPv6 for everything and enable IPv4 only as fallback [...] Preferring IPv6 over IPv4 is not even remotely the same as your original call to "lets drop all IPv4 only relays from consensus 2020 finally", as you wrote in message <08fee42f-f45d-a8c4-d4e6-c83c05b4fb08@afo-tm.org>.
Read the full sentence, dropping all IPv4 only relays is unavoidable because they can't reach relays behind CG-NAT, preferring IPv6 over IPv4 is only a handy side effect of this.
There hasn't been any new information in this thread for some time, and it appears to be escalating, so I've asked the moderators to consider closing it. I've already shared some details of the grant here:
On 13 Dec 2019, at 17:26, teor <teor@riseup.net> wrote:
The RIPE grant covers IPv6 address autodetection and self-testing. If the feature is reliable enough, we may turn on IPv6 on dual-stack relays by default. (When autodetection and self-testing both pass.)
We don't have any plans to disable IPv4 on relays.
We'll get back to you next year when we have a timeline, and some more specific information. T
hi Johathan, this question was already addressed last week by Teor. email on this mailing list doesn't always seem to arrive in a logical sequence.... possible due to spam filters and so on. gr. Paul teor teor@riseup.net via <https://support.google.com/mail/answer/1311182?hl=en> lists.torproject.org Fri, Dec 13, 8:26 AM (3 days ago) We won't be disabling IPv4 on relays any time soon. The RIPE grant covers IPv6 address autodetection and self-testing. If the feature is reliable enough, we may turn on IPv6 on dual-stack relays by default. (When autodetection and self-testing both pass.) We don't have any plans to disable IPv4 on relays. We'd need most relays to be dual-stack first. (Or we'd need research about privacy in non-clique networks.) And we'd need to write code that allows relays to turn off IPv4. When that's all deployed, we would have to make an engineering decision about the capacity of current IPv4-only relays, and the potential capacity of IPv6-only relays. One possible transition strategy is to allow IPv6-only bridges and exits. But to do that, we need more dual-stack guards and middles. That's why we are improving support for dual-stack relays with this funding. T On Mon, Dec 16, 2019 at 12:33 PM Jonathan Sélea <jonathan@selea.se> wrote:
I am unable to see why this is a good idea actually. As someone said already, I too would be unable to provide bridges and relays thanks to my ISP (Telia) does not provide an IPv6. But removing IPv4 only nodes from the network - you are basically removing a large chunk of relays from the network without any real reason.
On 12/12/2019 5:49 PM, NOC wrote:
Great,
than lets drop all IPv4 only relays from consensus 2020 finally.
P.S. whitelist me so it doesn't take days till my emails appear on the mailing list.
Thank you On 11.12.2019 03:20, teor wrote:
Dear relay operators,
I just wanted to let you know that RIPE has announced funding for The Tor Project to improve IPv6 support on relays. (RIPE is the European internet infrastructure organisation.) https://www.ripe.net/support/cpf/funding-recipients-2019
We'll have more details early in 2020, when we've worked out an implementation plan and a start time.
Thanks for your patience with our current IPv6 support. And thanks to all those volunteer coders who have worked hard to get us this far.
T
-- teor ----------------------------------------------------------------------
_______________________________________________ tor-relays mailing listtor-relays@lists.torproject.orghttps://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
_______________________________________________ tor-relays mailing listtor-relays@lists.torproject.orghttps://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
participants (7)
-
grarpamp -
Jonathan Sélea -
Logforme -
NOC -
Paul Geurts -
Ralph Seichter -
teor