Dear DNS folks, As many of you are aware, a group of researchers and other interested people and groups has been building and operating a root server testbed over the past year. This is the Yeti DNS project: http://yeti-dns.org/ We have just finished our first experiment, and have started our second one. The network has 25 root servers from 14 different organizations, along with a number of recursive resolvers configured to use them. We have approached the RIPE NCC to encourage them to run a Yeti root server, and they feel that this is something that we should engage with the RIPE community about. We are still looking for more Yeti root operators, and we think that the RIPE NCC would be awesome in this capacity, given their unique role in the DNS world (they are the only RIR who is also a root operator, plus they do a lot of work supporting various ccTLD). Personally I think that the goals of the Yeti project fit in very well with the mission of the RIPE NCC. Please let us know what you think about this idea. I'm happy to answer any questions about the project or what it would mean for the RIPE NCC to run a Yeti root server. Cheers, -- Shane
I think this fits the RIPE NCC very well given their role as “more than just an RIR”, their past and ongoing research work as well as their expertise in running of the Internet’s DNS root servers and their ability to both inform future experiments and translate outcomes into root server operations as applicable. Joao
On 24 May 2016, at 14:41, Shane Kerr <shane@time-travellers.org> wrote:
Dear DNS folks,
As many of you are aware, a group of researchers and other interested people and groups has been building and operating a root server testbed over the past year. This is the Yeti DNS project:
We have just finished our first experiment, and have started our second one. The network has 25 root servers from 14 different organizations, along with a number of recursive resolvers configured to use them.
We have approached the RIPE NCC to encourage them to run a Yeti root server, and they feel that this is something that we should engage with the RIPE community about.
We are still looking for more Yeti root operators, and we think that the RIPE NCC would be awesome in this capacity, given their unique role in the DNS world (they are the only RIR who is also a root operator, plus they do a lot of work supporting various ccTLD). Personally I think that the goals of the Yeti project fit in very well with the mission of the RIPE NCC.
Please let us know what you think about this idea. I'm happy to answer any questions about the project or what it would mean for the RIPE NCC to run a Yeti root server.
Cheers,
-- Shane
Hi, Am 24.05.2016 um 14:57 schrieb João Damas:
I think this fits the RIPE NCC very well given their role as “more than just an RIR”, their past and ongoing research work as well as their expertise in running of the Internet’s DNS root servers and their ability to both inform future experiments and translate outcomes into root server operations as applicable.
Joao
Full ACK. Maybe, RIPE would like to add some built-in tests into the ATLAS probes for testing the Yeti DNS infrastructure? Greetings, Max
On Tue, May 24, 2016 at 02:41:45PM +0200, Shane Kerr wrote:
the RIPE NCC would be awesome in this capacity, given their unique role in the DNS world (they are the only RIR who is also a root operator, plus they do a lot of work supporting various ccTLD). Personally I
indeed the NCC's participation would add certain amounts of weight, however the more important "currency" would be research contribution. Could you please elaborate on your expectations in this regard? -Peter
On 24 May 2016, at 13:41, Shane Kerr <shane@time-travellers.org> wrote:
Please let us know what you think about this idea.
<No hats> Before we take this idea forward, could someone please suggest metrics and milestones that could be used to assess the success or failure of this activity? For bonus points, it would be good to know what sort of resource commitments will be needed and the level of reporting that can be expected from the NCC being involved in Yeti DNS. We should be careful to not have the NCC get into open-ended, vague commitments where there's uncertainty about the on-going value to the community. Experiments are all very well but both the WG and the NCC's DNS team need to know how/when to decide to kill the experiment or turn it into a formally supported resource or do something else with it.
Actually Jim, first comes the poll of the community to see if this fits, later comes what you suggest, if the community decides to ask the RIPE NCC to look into it, or have we become the Ministry of IP? Joao
On 25 May 2016, at 09:55, Jim Reid <jim@rfc1035.com> wrote:
On 24 May 2016, at 13:41, Shane Kerr <shane@time-travellers.org> wrote:
Please let us know what you think about this idea.
<No hats> Before we take this idea forward, could someone please suggest metrics and milestones that could be used to assess the success or failure of this activity? For bonus points, it would be good to know what sort of resource commitments will be needed and the level of reporting that can be expected from the NCC being involved in Yeti DNS.
We should be careful to not have the NCC get into open-ended, vague commitments where there's uncertainty about the on-going value to the community. Experiments are all very well but both the WG and the NCC's DNS team need to know how/when to decide to kill the experiment or turn it into a formally supported resource or do something else with it.
This. On 25 May 2016 11:50:15 CEST, "João Damas" <joao@bondis.org> wrote:
Actually Jim, first comes the poll of the community to see if this fits, later comes what you suggest, -- Sent from Kaiten Mail. Please excuse my brevity.
On 25 May 2016, at 10:50, João Damas <joao@bondis.org> wrote:
Actually Jim, first comes the poll of the community to see if this fits,
Well Joao the community already seems to be heading in that direction. YMMV. That said, it would be helpful for the WG to have a better understanding of the requirements before asking the NCC to investigate.
later comes what you suggest, if the community decides to ask the RIPE NCC to look into it,
Up to a point. Though before the NCC does look into this (or anything else for that matter), I think the NCC should have a sense of what the WG's expectations and objectives are. Unclear requirements help nobody.
or have we become the Ministry of IP?
I have no idea what this Ministry does or why they should interfere in WG matters. :-)
On Wed, May 25, 2016 at 11:50:15AM +0200, João Damas wrote:
Actually Jim, first comes the poll of the community to see if this fits, later comes what you suggest, if the community decides to ask the RIPE NCC to look into it, or have we become the Ministry of IP?
this member of the community considers the request underspecified to offer an informed opinion. I do not consider running alternative root name servers a "good fit" - everything else I do not know. Clarification appreciated. -Peter
On 5/24/16 2:41 PM, Shane Kerr wrote:
Please let us know what you think about this idea. I'm happy to answer any questions about the project or what it would mean for the RIPE NCC to run a Yeti root server.
hi, as a participant to YETI project I warmly encourage RIPE NCC (as well as the RIPE community) to support this project. it aims to conduct experiments not safely feasible in the current production environment. here you can read the problem statement: https://github.com/BII-Lab/Yeti-Project/blob/master/doc/Yeti_PS.md thank you -- antonio
Hi all, Just want to know, how far or how close YETI is as compared to the Open Root [1] project which I know for a fact that the RIR community quite vehemently voiced against it. -- Amreesh [1] http://www.open-root.eu/ On Tue, May 24, 2016 at 2:41 PM, Shane Kerr <shane@time-travellers.org> wrote:
Dear DNS folks,
As many of you are aware, a group of researchers and other interested people and groups has been building and operating a root server testbed over the past year. This is the Yeti DNS project:
We have just finished our first experiment, and have started our second one. The network has 25 root servers from 14 different organizations, along with a number of recursive resolvers configured to use them.
We have approached the RIPE NCC to encourage them to run a Yeti root server, and they feel that this is something that we should engage with the RIPE community about.
We are still looking for more Yeti root operators, and we think that the RIPE NCC would be awesome in this capacity, given their unique role in the DNS world (they are the only RIR who is also a root operator, plus they do a lot of work supporting various ccTLD). Personally I think that the goals of the Yeti project fit in very well with the mission of the RIPE NCC.
Please let us know what you think about this idea. I'm happy to answer any questions about the project or what it would mean for the RIPE NCC to run a Yeti root server.
Cheers,
-- Shane
-- Amreesh Phokeer
On 5/25/16 12:32 PM, Amreesh Phokeer wrote:
Hi all,
Just want to know, how far or how close YETI is as compared to the Open Root [1] project
from what I read on their website, I think they cannot be compared. think of yeti as a testbed for research and experiments: the project should terminate at the end of 2018 and it does not provide an alternate name space. -- antonio
On Wed, May 25, 2016 at 12:32:06PM +0200, Amreesh Phokeer <amreesh.phokeer@gmail.com> wrote a message of 110 lines which said:
Just want to know, how far or how close YETI is as compared to the Open Root [1] project
No comparison possible. 1) Yeti is an actual service, running. You can use it now if you want, you can test it, the IP addresses of the root name servers are published <http://yeti-dns.org/rootzone.html> OpenRoot is a joke, which does not have even one server. See <http://www.open-root.eu/about-open-root/how-to-install-an-open-root-website-69/> "For testing Open-Root DNS, ask us" They don't even publish the IP addresses of their root name servers (something that every alternative root, responsible or not, always does prominently). At the JDDL meeting in Lyon, one of the Open Root founders, asked to indicate these IP addresses, said they were not published "for security reasons". 2) Yeti is managed by people who know one or two things about the DNS. Open Root is managed by complete ignorants. See <http://www.open-root.eu/about-open-root/open-roots/> "In the absence of user’s choice, the root activated by default on all PCs is the Icann’s one." or "However Google (and Chrome) use their own root (copy of the ICANN root); its IP address is 8.8.8.8 or 8.8.6.6." (and this is not the worst claim on their site). 3) Yeti is a not-for-profit project, run by volunteers. Open Root is a commercial company selling TLDs <http://www.open-root.eu/our-offer/towards-new-manners/>
participants (10)
-
Amreesh Phokeer -
Antonio Prado -
Jim Reid -
João Damas -
Max Grobecker -
Niall O'Reilly -
Peter Koch -
Randy Bush -
Shane Kerr -
Stephane Bortzmeyer