> > which roots are anycast? c f i j k? > b m

g

Not according to http://root-servers.org. Hiding the places? -Hank

root-servers.org is not definative. why do you think it is?

because the community expects the root server ops to be helpful, open, honest, and transparent. you know, all that service to the community stuff. damned shame not all seem to be. randy

On Sun, 14 Nov 2004, Hank Nussbacher wrote:

I think it is time to update rfc2870 and add a few new MUST paragraphs :-)

Do you want a refund on your root-server bill?

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2004-11-12, at 03.03, Randy Bush wrote:

...

...

which roots are anycast? c f i j k? b m

thanks.

which are widely anycast, i.e. at more than three or four locations OR on three or more continents?

I think that http://root-servers.org is up to date for _most_ root-servers. - - kurtis - -----BEGIN PGP SIGNATURE----- Version: PGP 8.1 iQA/AwUBQZQcLaarNKXTPFCVEQKdTgCfSZ/sYJXC1FiGQ7HYH0bJgFEVHzIAoIlE 3cQ7TTN81E5ee+P83FCcd9Z6 =ComI -----END PGP SIGNATURE-----

and the good folks on nanog would know this why?

dunno, bill. maybe because it has to do with network operations? but we did get the answers we needed, thanks to some of those good folk. and non-answers from others. all as expected. welcome to the internet. randy

...

which roots are anycast? c f i j k?

b m

- - kurtis -

According to http://root-servers.org only m is. -Hank

kurtis@kurtis.pp.se (Kurt Erik Lindqvist) wrote:

On 2004-11-12, at 02.53, Randy Bush wrote:

...

which roots are anycast? c f i j k? b m

k (London, Amsterdam, Frankfurt) Elmar. -- "Begehe nur nicht den Fehler, Meinung durch Sachverstand zu substituieren." (PLemken, <bu6o7e$e6v0p$2@ID-31.news.uni-berlin.de>) --------------------------------------------------------------[ ELMI-RIPE ]---

Helsinki (FI); Geneva (CH). www.root-servers.org is up to date.

for those nodes that choose to use root-servers.org as a publication method, that might be true. I take your note to mean that www.root-servers.org is up to date wrt the publication of deployed sites for the K server. Don't presume to speak for the other operators please. -- bill

...

Elmar.

Andrei Robachevsky RIPE NCC

i'm saying that there is no place that is public that has connectivity information for all instances of the "B" servers. i am unaware of a single persistant place and no organization that speaks for all server operators in their capacity as root operators. On Fri, Nov 12, 2004 at 07:26:26AM -0500, Ray Plzak wrote:

I just want to be sure that I understand what you are saying, Bill.

There is no single place where one can obtain authoritative information about the root server system and all of its individual servers. There is no single organization that speaks in any collective manner for the root server operators.

Ray

...

-----Original Message----- From: owner-nanog@merit.edu [mailto:owner-nanog@merit.edu] On Behalf Of bmanning@vacation.karoshi.com Sent: Friday, November 12, 2004 7:19 AM To: Andrei Robachevsky Cc: Elmar K. Bins; nanog@nanog.org Subject: Re: anycast roots

...

Helsinki (FI); Geneva (CH). www.root-servers.org is up to date.

for those nodes that choose to use root-servers.org as a publication method, that might be true. I take your note to mean that www.root- servers.org is up to date wrt the publication of deployed sites for the K server.

Don't presume to speak for the other operators please.

-- bill

...

...

Elmar.

Andrei Robachevsky RIPE NCC

Bill, bmanning@vacation.karoshi.com wrote:

...

Helsinki (FI); Geneva (CH). www.root-servers.org is up to date.

for those nodes that choose to use root-servers.org as a publication method, that might be true. I take your note to mean that www.root-servers.org is up to date wrt the publication of deployed sites for the K server.

Yes, that's what I meant. Andrei

Don't presume to speak for the other operators please.

-- bill

...

...

Elmar.

Andrei Robachevsky RIPE NCC

On Fri, Nov 12, 2004 at 11:00:54PM +0000, Paul Vixie wrote:

bmanning@vacation.karoshi.com wrote:

...

for those nodes that choose to use root-servers.org as a publication method, ...

as far as i know, the root-servers.org web site is 100% accurate, and has 100% participation amongst root server operators wrt publishing accurate information. if bill knows otherwise, i'd like him to say so, and offer specifics both as to what information is not present, and where it can be found today, and where it can be found in the future.

if bill is not actually aware of any missing information, i'd like him to say so, since it's possible to infer from his comments here that the web site is inaccurate or out-of-date in some way he's directly aware of, and i just don't think that's true. -- Paul Vixie

please remove all public informtion about "B" from http://www.root-servers.org. The public data about "B" on that site is not 100% accurate and does not have 100% participation amoungst root-server operators wrt publishing accurate information. plesae be apprised that your thinking here is inaccurate. --bill

Paul Vixie wrote (responding to Bill Manning):

...

... The public data about "B" on that site is not 100% accurate and does not have 100% participation amoungst root-server operators wrt publishing accurate information.

you've got root on that web server. if you want something removed, you can remove it yourself.

Alright guys, we've known Bill for many years, going back to Rice. Since it has now been revealed that Bill has root there, then we should hold Bill's feet to the fire until his information is accurate. Sometimes we opine about abstractions -- "the operators did it" -- when it's PEOPLE that are the operators. We always have our best results when we get the people involved directly with each other, instead of hiding behind those impersonal corporate abstractions. Bill, you've done good work in the past. Let us know when things are fixed there.... Even where it's really somebody else's job. -- William Allen Simpson Key fingerprint = 17 40 5E 67 15 6F 31 26 DD 0D B9 9B 6A 15 2C 32

On Fri, Nov 12, 2004 at 11:00:54PM +0000, Paul Vixie <vixie@vix.com> wrote a message of 18 lines which said:

as far as i know, the root-servers.org web site is 100% accurate,

Following the recent discussion about "anycast jitter" with j.root-servers.net, I believe one information is missing: wether the node is global or local (BGP NO_EXPORT). It is not easy to find by itself (you have to do a lot of traceroutes) so, if you have access to this information, it would be quite useful. (I'm one of the persons who see a lot of jitter for j.root-servers.net with Randy Bush's experiment.)

On Wed, Nov 17, 2004 at 02:37:25PM +0100, Elmar K. Bins <elmi@4ever.de> wrote a message of 34 lines which said:

in alternating fashion, but I would assume "jns1" through "jns6" are just the individual servers of a setup called "hgtld".

That's a reasonable guess. Someone from Verisign to confirm/infirm?

On Wed, Nov 17, 2004 at 10:05:20AM -0500, Joe Abley <jabley@isc.org> wrote a message of 36 lines which said:

I have no idea about Verisign's scheme, but in case anybody notices similar distribution of queries across F root servers, it may help to know that:

xxxNa.f.root-servers.org xxxNb.f.root-servers.org xxxNc.f.root-servers.org etc

are hosts all located at the same site "xxxN".

OK, I understand. So, like Elmar Bins, I was seeing "intra-site" jitter, which is normal (it is only seen with UDP queries, probably because the Verisign load balancer is stateful and remembers the binding for TCP) and no "inter-site" jitter, which would be more serious. But I'm quite at this edge of the Internet, so let's wait for more reports with Peter Boothe's tool. And just be sure to "sanitize" the results before jumping to the wrong conclusion, like I did.

...

as far as i know, the root-servers.org web site is 100% accurate,

Following the recent discussion about "anycast jitter" with j.root-servers.net, I believe one information is missing: wether the node is global or local (BGP NO_EXPORT).

note that BGP NO_EXPORT is but one of several ways to make a "local node". for many of our peers, "keep this to one's own network and one's customers" is the default, and they have to strip off our NO_EXPORT community in order to achieve this. but your question is understandable on that modified basis.

It is not easy to find by itself (you have to do a lot of traceroutes) so, if you have access to this information, it would be quite useful.

the www.root-servers.org main page is pretty crowded already. adding an indication of global-vs-local for each city could be pretty distracting; especially since those rootops who anycast usually have their own separate web site describing their efforts, which are more easily kept up to date and which, being root-specific, are generally less crowded. some rootops even prefer to keep this level of detail out of the public eye, possibly because it's the kind of thing that makes ddos attacks easier to plan. speaking for f-root only, we don't mind that the world knows which of our anycast nodes are global and which are local. but the other rootops will have to speak for themselves, and they will probably not choose to use the www.root-servers.org web page to signal this level of detail. (remember, these are the kinds of details that can change every week or every day.) -- Paul Vixie