[routing-wg]Four byte ASN notation
Folks, One of the items that was discussed during the working group meeting in Amsterdam was the relative benefits of using asplain, asdot or asdot+ to represent four byte autonomous system numbers. The chairs feel that the discussion tended towards favouring asdot/asdot+, which is already used relatively widely, but with the note that people do need to review their in-house tools and scripts to ensure they will work with numbers expressed in this notation. Cheers, Rob
Not sure if you aware but there is an IETF Internet Draft winging its way through the system on this topic, namely http://www.ietf.org/internet-drafts/draft-michaelson-4byte-as-representation.... txt It did get discussed on the IETF idr list in October 2006, and met significant resistance. There were also comments then about NANOG taking a position on this. I haven't seen any discussion since. Current status is Application Director Watching Tom Petch ----- Original Message ----- From: "Rob Evans" <rhe@nosc.ja.net> To: <routing-wg@ripe.net> Sent: Tuesday, December 04, 2007 11:59 PM Subject: [routing-wg]Four byte ASN notation
Folks,
One of the items that was discussed during the working group meeting in Amsterdam was the relative benefits of using asplain, asdot or asdot+ to represent four byte autonomous system numbers. The chairs feel that the discussion tended towards favouring asdot/asdot+, which is already used relatively widely, but with the note that people do need to review their in-house tools and scripts to ensure they will work with numbers expressed in this notation.
Cheers, Rob
Tom Petch wrote:
Not sure if you aware but there is an IETF Internet Draft winging its way through the system on this topic, namely
http://www.ietf.org/internet-drafts/draft-michaelson-4byte-as-representation.... txt
speaking as a co-author here: "winging its way" is not a phrase I'd use for almost any draft in the IETF system, but in this case its an entirely inappropriate characterization of the pace of this document. The tracked state of this draft is "AD is Watching", but if you look a little closer you see that the IETF Area Director listed is not a current Routing Area director. So this characterization is just a little exaggerated! Like the 4byte AS draft itself I'd tend towards an adjective to describe thepace of this document through the IETF as "glacial" but maybe others would see "geological" as being closer to reality ;-). Even so, I encourage those who have some interest in this topic to read the draft and comment, either to the authors, to this mailing list, to the idr mailing list where the 4 byte ASN work was undertaken in the context of the IETF, or wherever else that might take your fancy.
It did get discussed on the IETF idr list in October 2006, and met significant resistance.
Again I have to say that this characterization is just a little exaggerated!
There were also comments then about NANOG taking a position on this.
err? NANOG "taking a position" ? How? Though some subliminal collective subconscious alignment? Or via some alignment of the planets?
I haven't seen any discussion since.
Notation is such a strange thing - all kinds of folk have flash opinions about notation but in the end notation is like pronounciation - informal use accretes social weight through continued usage and then the informaal use becomes a convention which then becomes usage rules. But when we try to apply this social process to technology we run into the issue of formal notation and rigid grammars because of the issues of have a notation that can be used conveniently by both human and machine parsers. So it makes some sense to try to define a notation convention early on in the process. The draft notes some alternatives for notation that have been observed so far and makes a recommendation to adopt one such notation ("asdot" in this case, using the terminology as described in the draft)
Current status is Application Director Watching
Actual status is "previous AD might have been watching" Rob Evan's advice to the folk on this list that: "people do need to review their in-house tools and scripts to ensure they will work with numbers expressed in this notation." is still appropriate and extends far beyond mere notation. The issue is one of looking at your operating support system and provisioning tools and even if you are not going to upgrade your routers' BGP anytime soon, what happens when your customers or peers front up with a 4 byte ASN and your systems start to see AS23456 popping up everywhere? (see slides 38 and 39 of http://www.iepg.org/2007-12-ietf70/asns.pdf for some additional pointers here as to what to review and why) regards, Geoff
Geoff Interesting. I was unsure whether or not this WG was familiar with the Internet Draft I referred to and thought that in case it was not, I would mention it. Before doing so, I checked the status on the Internet Drafts and was puzzled by what I saw - but posted here anyway. The discussion on the IETF idr list a year ago was in response to a Last Call of an individual submission (as opposed to a chartered item of a Working Group)after which, I would expect it either to be approved as an RFC, or revised and resubmitted. A year later it is not an RFC - hence my reference to meeting resistance - nor have I seen any further discussion about it on the idr list. So when I checked the Internet Drafts database, I wouldn't have been surprised to see that the ID had expired. Instead, it has advanced from version -01 to version -05; five versions in a year is the progress of a swift, rather than that of a snail, hence my reference to 'winging it'. And you have joined in as an editor of it. So I remain puzzled; is it being discussed somewhere else? (NANOG?:-) what has triggered all revised versions? where has it been in the past year and where is it going? Tom Petch ----- Original Message ----- From: "Geoff Huston" <gih@apnic.net> To: "tp" <ripe@dial.pipex.com> Cc: "Rob Evans" <rhe@nosc.ja.net>; <routing-wg@ripe.net> Sent: Friday, December 07, 2007 3:48 PM Subject: Re: [routing-wg]Four byte ASN notation
Tom Petch wrote:
Not sure if you aware but there is an IETF Internet Draft winging its way through the system on this topic, namely
http://www.ietf.org/internet-drafts/draft-michaelson-4byte-as-representation....
txt
speaking as a co-author here: "winging its way" is not a phrase I'd use for almost any draft in the IETF system, but in this case its an entirely inappropriate characterization of the pace of this document.
The tracked state of this draft is "AD is Watching", but if you look a little closer you see that the IETF Area Director listed is not a current Routing Area director.
So this characterization is just a little exaggerated!
Like the 4byte AS draft itself I'd tend towards an adjective to describe thepace of this document through the IETF as "glacial" but maybe others would see "geological" as being closer to reality ;-).
Even so, I encourage those who have some interest in this topic to read the draft and comment, either to the authors, to this mailing list, to the idr mailing list where the 4 byte ASN work was undertaken in the context of the IETF, or wherever else that might take your fancy.
It did get discussed on the IETF idr list in October 2006, and met
significant
resistance.
Again I have to say that this characterization is just a little exaggerated!
There were also comments then about NANOG taking a position on this.
err? NANOG "taking a position" ? How? Though some subliminal collective subconscious alignment? Or via some alignment of the planets?
I haven't seen any discussion since.
Notation is such a strange thing - all kinds of folk have flash opinions about notation but in the end notation is like pronounciation - informal use accretes social weight through continued usage and then the informaal use becomes a convention which then becomes usage rules. But when we try to apply this social process to technology we run into the issue of formal notation and rigid grammars because of the issues of have a notation that can be used conveniently by both human and machine parsers. So it makes some sense to try to define a notation convention early on in the process.
The draft notes some alternatives for notation that have been observed so far and makes a recommendation to adopt one such notation ("asdot" in this case, using the terminology as described in the draft)
Current status is Application Director Watching
Actual status is "previous AD might have been watching"
Rob Evan's advice to the folk on this list that:
"people do need to review their in-house tools and scripts to ensure they will work with numbers expressed in this notation."
is still appropriate and extends far beyond mere notation. The issue is one of looking at your operating support system and provisioning tools and even if you are not going to upgrade your routers' BGP anytime soon, what happens when your customers or peers front up with a 4 byte ASN and your systems start to see AS23456 popping up everywhere?
(see slides 38 and 39 of http://www.iepg.org/2007-12-ietf70/asns.pdf for some additional pointers here as to what to review and why)
regards,
Geoff
The revised version history has been a process of a) revising the document in the light of comments received or b) resubmitting the document to circumvent the draft dieing http://smakd.potaroo.net/ietf/idref/draft-michaelson-4byte-as-representation... has the full history. The most substantive discussion has been int he RIPE Routing WG, and there was earlier discussion at the ARIN meeting itself. The emerging convention appears to be "asdot" as far as I can tell, although there are some folk who insist that the added complexity in using regular expressions involving AS numebrs should rule this out. Others see this notation as being no more or less of a problem that IP address dotted quad and cite the number space of two smaller numbers as being less prone to human transcription and use errors. Its a style thing Geoff tp wrote:
Geoff
Interesting.
I was unsure whether or not this WG was familiar with the Internet Draft I referred to and thought that in case it was not, I would mention it.
Before doing so, I checked the status on the Internet Drafts and was puzzled by what I saw - but posted here anyway.
The discussion on the IETF idr list a year ago was in response to a Last Call of an individual submission (as opposed to a chartered item of a Working Group)after which, I would expect it either to be approved as an RFC, or revised and resubmitted. A year later it is not an RFC - hence my reference to meeting resistance - nor have I seen any further discussion about it on the idr list.
So when I checked the Internet Drafts database, I wouldn't have been surprised to see that the ID had expired. Instead, it has advanced from version -01 to version -05; five versions in a year is the progress of a swift, rather than that of a snail, hence my reference to 'winging it'. And you have joined in as an editor of it.
So I remain puzzled; is it being discussed somewhere else? (NANOG?:-) what has triggered all revised versions? where has it been in the past year and where is it going?
Tom Petch
----- Original Message ----- From: "Geoff Huston" <gih@apnic.net> To: "tp" <ripe@dial.pipex.com> Cc: "Rob Evans" <rhe@nosc.ja.net>; <routing-wg@ripe.net> Sent: Friday, December 07, 2007 3:48 PM Subject: Re: [routing-wg]Four byte ASN notation
Tom Petch wrote:
Not sure if you aware but there is an IETF Internet Draft winging its way through the system on this topic, namely
http://www.ietf.org/internet-drafts/draft-michaelson-4byte-as-representation....
txt
speaking as a co-author here: "winging its way" is not a phrase I'd use for almost any draft in the IETF system, but in this case its an entirely inappropriate characterization of the pace of this document.
The tracked state of this draft is "AD is Watching", but if you look a little closer you see that the IETF Area Director listed is not a current Routing Area director.
So this characterization is just a little exaggerated!
Like the 4byte AS draft itself I'd tend towards an adjective to describe thepace of this document through the IETF as "glacial" but maybe others would see "geological" as being closer to reality ;-).
Even so, I encourage those who have some interest in this topic to read the draft and comment, either to the authors, to this mailing list, to the idr mailing list where the 4 byte ASN work was undertaken in the context of the IETF, or wherever else that might take your fancy.
It did get discussed on the IETF idr list in October 2006, and met significant resistance. Again I have to say that this characterization is just a little exaggerated!
There were also comments then about NANOG taking a position on this.
err? NANOG "taking a position" ? How? Though some subliminal collective subconscious alignment? Or via some alignment of the planets?
I haven't seen any discussion since.
Notation is such a strange thing - all kinds of folk have flash opinions about notation but in the end notation is like pronounciation - informal use accretes social weight through continued usage and then the informaal use becomes a convention which then becomes usage rules. But when we try to apply this social process to technology we run into the issue of formal notation and rigid grammars because of the issues of have a notation that can be used conveniently by both human and machine parsers. So it makes some sense to try to define a notation convention early on in the process.
The draft notes some alternatives for notation that have been observed so far and makes a recommendation to adopt one such notation ("asdot" in this case, using the terminology as described in the draft)
Current status is Application Director Watching Actual status is "previous AD might have been watching"
Rob Evan's advice to the folk on this list that:
"people do need to review their in-house tools and scripts to ensure they will work with numbers expressed in this notation."
is still appropriate and extends far beyond mere notation. The issue is one of looking at your operating support system and provisioning tools and even if you are not going to upgrade your routers' BGP anytime soon, what happens when your customers or peers front up with a 4 byte ASN and your systems start to see AS23456 popping up everywhere?
(see slides 38 and 39 of http://www.iepg.org/2007-12-ietf70/asns.pdf for some additional pointers here as to what to review and why)
regards,
Geoff
Hi all,
Not sure if you aware but there is an IETF Internet Draft winging its way through the system on this topic, namely
http://www.ietf.org/internet-drafts/draft-michaelson-4byte-as-representation.... txt
The -05 version is a major rewrite w.r.t. the -01 version discussed about a year ago.
It did get discussed on the IETF idr list in October 2006, and met significant resistance. There were also comments then about NANOG taking a position on this. I haven't seen any discussion since.
Current status is Application Director Watching
That is about right: the AD is watching developments but there aren't any. This is one of those cases where there are 3 solutions which all have their pro's and con's, and where neither one is better than any of the others. The only way out is that somebody makes a decision to pick one and accept that people in favor of the other solutions will be unhappy for a while, before they realize that they'll have to live with it. Right now, the time doesn't seem right yet to make this decision, so I guess that we'll have to live with this situation for a while. Henk -- ------------------------------------------------------------------------------ Henk Uijterwaal Email: henk.uijterwaal(at)ripe.net RIPE Network Coordination Centre http://www.amsterdamned.org/~henk P.O.Box 10096 Singel 258 Phone: +31.20.5354414 1001 EB Amsterdam 1016 AB Amsterdam Fax: +31.20.5354445 The Netherlands The Netherlands Mobile: +31.6.55861746 ------------------------------------------------------------------------------ Is one of the choices leaving the office open? Alan Greenspan on the next elections
Hi, On Fri, Dec 07, 2007 at 04:23:12PM +0100, Henk Uijterwaal wrote:
Right now, the time doesn't seem right yet to make this decision, so I guess that we'll have to live with this situation for a while.
I don't understand what other time would be better. *Now* folks need to get their tools and software adapted - and they need to know what format to code for. (Actually "two years ago" would have been adequate) Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 110584 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279
*Now* folks need to get their tools and software adapted - and they need to know what format to code for.
For what it is worth, I agree with Gert. I can see a strong argument for storing the ASNs as an unsigned 32 bit number internally, but the user-interface and education needs to start happening now before the larger ASNs become routine. We only have just over a year until 32 bit ASNs are the default for RIR assignments. "asdot" is becoming the default, and one of the reasons for having the discussion in October was that if there was a strong desire for another representation, operators would have to express that before asdot became a fait accomplis. I was, and am, aware of George Michaelson's draft, which is why I used the terminology from it when I summed up what I felt the feeling in the room was. Cheers, Rob
Gert Doering wrote:
Hi,
On Fri, Dec 07, 2007 at 04:23:12PM +0100, Henk Uijterwaal wrote:
Right now, the time doesn't seem right yet to make this decision, so I guess that we'll have to live with this situation for a while.
I don't understand what other time would be better.
*Now* folks need to get their tools and software adapted - and they need to know what format to code for.
I fully agree for the real world. Here it seems that RIRs and some of the vendors use asdot(+), so a safe bet would be to use that. In the IETF world, things are a bit different: if this proposal would be put back on the table, my guess is that it would result in an endless discussion with no conclusion reached. Which is why people just wait hoping that some de-facto standard will evolve and the discussion can be avoided. Henk -- ------------------------------------------------------------------------------ Henk Uijterwaal Email: henk.uijterwaal(at)ripe.net RIPE Network Coordination Centre http://www.amsterdamned.org/~henk P.O.Box 10096 Singel 258 Phone: +31.20.5354414 1001 EB Amsterdam 1016 AB Amsterdam Fax: +31.20.5354445 The Netherlands The Netherlands Mobile: +31.6.55861746 ------------------------------------------------------------------------------ Is one of the choices leaving the office open? Alan Greenspan on the next elections
On Fri, 7 Dec 2007, Henk Uijterwaal wrote:
I fully agree for the real world. Here it seems that RIRs and some of the vendors use asdot(+), so a safe bet would be to use that.
You've left out the operators, *all* of whom are going to audit all their all their configuration state and update any string-form filters of ASNs. There's no point revisting the arguments though, so let's try avoid classifications like the above. A decision would indeed be nice. Given "asdot" is a fait accompli, by way of unmandated implementation, and there's little will to turn this barge back to port, why not write up and approve whatever policy documents are required? regards, -- Paul Jakma paul@clubi.ie paul@jakma.org Key ID: 64A2FF6A Fortune: I don't know who my grandfather was; I am much more concerned to know what his grandson will be. -- Abraham Lincoln
On Fri, 7 Dec 2007, Henk Uijterwaal wrote:
This is one of those cases where there are 3 solutions which all have their pro's and con's, and where neither one is better than any of the others. The only way out is that somebody makes a decision to pick one and accept that people in favor of the other solutions will be unhappy for a while, before they realize that they'll have to live with it.
One of the "solutions" has got "The folks who run the DBs are enamoured with it" in the "Pro" list. That seems to have beeen the only meaningful pro/con, given that both operator and vendor opinion in other forums has largely been against, but yet this unmandated proposal still was implemented.
Right now, the time doesn't seem right yet to make this decision, so I guess that we'll have to live with this situation for a while.
I think what you mean the decision has been made as a fait accompli, and the time you refer to is the time needed to cement it as so. Standard disclaimer: If any of the above language sounds harsh, please understand we're only talking about AS numbers. ;) regards, -- Paul Jakma paul@clubi.ie paul@jakma.org Key ID: 64A2FF6A Fortune: QOTD: "This is a one line proof... if we start sufficiently far to the left."
Hi, On Fri, Dec 07, 2007 at 04:17:54PM +0000, Paul Jakma wrote:
That seems to have beeen the only meaningful pro/con, given that both operator and vendor opinion in other forums has largely been against, but yet this unmandated proposal still was implemented.
I can't see any evidence for "operator .. opinion ... has been largely against" from the discussion in the RIPE routing WG. Some operators had concerns, others have been very much in favour of asdot. Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 110584 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279
Tom Petch wrote:
Not sure if you aware but there is an IETF Internet Draft winging its way through the system on this topic, namely
http://www.ietf.org/internet-drafts/draft-michaelson-4byte-as-representation.... txt
speaking as a co-author here: "winging its way" is not a phrase I'd use for almost any draft in the IETF system, but in this case its an entirely inappropriate characterization of the pace of this document. The tracked state of this draft is "AD is Watching", but if you look a little closer you see that the IETF Area Director listed is not a current Routing Area director. Like the 4byte AS draft itself I'd tend towards an adjective to describe the pace of this document through the IETF as "slow" or even "glacial" but maybe others would see "geological" as being appropriate ;-). Even so, I encourage those who have some interest in this topic to read the draft and comment, either to the authors, to this mailing list, to the idr mailing list where the 4 byte ASN work was undertaken in the context of the IETF, or wherever else that might take your fancy. Rob Evan's advice to the folk on this list that: "people do need to review their in-house tools and scripts to ensure they will work with numbers expressed in this notation." is still appropriate and extends far beyond mere notation. The issue is one of looking at your operating support system and provisioning tools and even if you are not going to upgrade your routers' BGP anytime soon, what happens when your customers or peers front up with a 4 byte ASN and your systems start to see AS23456 popping up everywhere? (see slides 38 and 39 of http://www.iepg.org/2007-12-ietf70/asns.pdf for some additional pointers here as to what to review and why) regards, Geoff
participants (6)
-
Geoff Huston -
Gert Doering -
Henk Uijterwaal -
Paul Jakma -
Rob Evans -
tp