- db-wg - mailman.ripe.net

New format for notifications from RIPE Database.
by RIPE Database Administration 14 Jul '97

14 Jul '97

Dear Colleagues, To give more time to users who wish to modify scripts, the new format for notification messages from the RIPE Database will be implemented on Monday, 21 July. If you have anymore questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration.

1 0

Cross Notification in the Routing Registry
by Carol Orange 11 Jul '97

11 Jul '97

Dear all, The following describes a mechanism for "Cross Notification in the Routing Registry" to be implemented shortly at the RIPE NCC. The mechanism has been modified to incorporate the feedback received during RIPE 27. Appended is a short history explaining how the proposal reached its current form. If this mechanism is of interest to you, please review the document and send any comments or questions you may have to myself. Unless we hear show stoppers, we will start the implementation in the week of July 21, 1997. Greetings, Carol Orange, RIPE NCC -------------------------------------------------------------------- Cross Notification in the Routing Registry Carol Orange, July 1997 Suppose you register the announcement of a route covering some /24. When you make the announcement, it may be useful to know that the /24 has already been registered as being announced in another AS. Alternatively, if a user of a /24 which is announced by you decides to become multi-homed, it may be of interest to you. Overlapping routes are often unintentionally announced in a single AS, and a mechanism for detecting such errors would certainly be handy. In the following, we propose a simple cross notification mechanism which will allow routing registry users to be notified of overlapping route announcements in the Routing Registry (RR). Consider the route object: route: [mandatory] [single] [primary/look-up key] descr: [mandatory] [multiple] [ ] origin: [mandatory] [single] [primary key] hole: [optional] [multiple] [ ] withdrawn: [optional] [single] [ ] comm-list: [optional] [multiple] [ ] advisory: [optional] [multiple] [ ] remarks: [optional] [multiple] [ ] notify: [optional] [multiple] [inverse key] mnt-by: [mandatory] [multiple] [inverse key] changed: [mandatory] [multiple] [ ] source: [mandatory] [single] [ ] To address the considerations above, we propose the following notification mechanism be implemented in the Routing Registry: 1. Whenever a route object is added to the routing registry, a check is done to see if the address space overlaps with that in any other route object in the RR. If so, a notification will be sent to the submitter of the route object. This will always happen and is not dependent on any notification attributes described below. 2. A new field called "cross-nfy:" will be added to the route and aut-num objects. This field: a) contains a contact reference, specified as either + a NIC-handle pointing to a person or role object, the email address of which will get the notification; or + a mntner ID referencing the maintainer to be notified; b) is optional; and c) is multiple (so that more than one contact can be notified). This attribute is used to request notifications of the addition or removal of route objects which overlap the address space of the already registered route objects in which this attribute is specified. If the "cross-nfy:" attribute is used in: + a route object, then a notification will be sent for any overlaps with the address space specified in the route object; + an aut-num object, then a notification will be sent for overlaps with the address space specified in each route object in which the given AS number is specified in the "origin:" attribute. The contact in "cross-nfy:" will receive an e-mail whenever: + a route is registered which overlaps this address space; or + a route is removed which overlapped this address space. If all is well, the former will indicate the introduction of multi-homing, and the latter will indicate the cancellation of the multi-homing. It is expected, however, that the former will often indicate an error, and the latter its removal. Either way, it will be useful to know about the registration and elimination of overlapping announcements. Some Important Notes -------------------- 1. At the time of implementation, the notification may take up to 24 hours, due to the current mirroring scheme employed in the RR. This time will be reduced when the mirroring scheme is improved. 2. Exact address space matches will be distinguished from non-exact overlaps in the notification messages. 3. After the implementation of RPSL, it will be useful to allow one or more route-sets to be specified in the "cross-nfy:" attribute in addition to the contact reference. For example, using: cross-nfy: <ContactA> <route-set1> <route-set3> cross-nfy: <ContactB> <route-set2> in an aut-num object would allow ContactA to be notified upon registration and deletion of routes overlapping the address space covered in <route-set1> and <route-set3> and ContactB to be notified with regards to route registrations with address space overlapping that covered in <route-set2>. In this way, a more selective notification scheme can be specified without adding "cross-nfy:" to every route object. Acknowledgements ---------------- Thanks to the RIPE Routing and Database working groups for lively discussions resulting in constructive input on this notification mechanism. Daniel Karrenberg was the first to mention the need for a notification mechanism in the RR. Chris Fletcher, Joachim Schmitz and Wilfried Woeber have all made essential contributions. ------------- ------------- ------------- ------------ ------------- Appendix - Some History At the January meeting of the Routing WG in Amsterdam, various possible hierarchies for authorization in the Routing Registry (RR) were considered. During the discussion, a suggestion was made that "hierarchical notification" would be useful even in cases where authority is not well defined. This suggestion enjoyed strong support of both the Routing WG and the Database WG. At the start of May 1997, a proposal was submitted to the mailing lists of the RIPE Routing and Database WG's describing a notification mechanism designed to meet the stated requirements. During the RIPE meeting in Dublin, it was agreed to that the notification mechanism should be implemented, with the following modifications: 1. Whenever a route object is submitted in the RR, a notification should be generated to the submitter if the address space overlaps with any other route object in the RR. [RIPE27 - Routing WG meeting] 2. The attribute should be named "cross-nfy:" rather than "x-notify:" to avoid any misunderstandings with respect to other conventions. In particular, mailers often interpret "X-field:" to mean "ignore this field". [private discussion Chris Fletcher] 3. If the address space covered in the route object matched the address space in another route object, be sure this is clear in the notification message. [RIPE27 - Routing WG meeting] 4. - Notifications should be sent to the appropriate contact not only if an overlap is introduce, but also when an overlap is eliminated. [RIPE27 - Routing WG meeting] 5. The "cross-nfy:" field should be available for the "aut-num:" object as well as for the routing object. If "cross-nfy:" is in the "aut-num:" object, it applies to all routes with that aut-num as origin. [RIPE27 - Database WG meeting] It has since been suggested that it would be useful if the mechanism would be extensible for the "route-set:" objects defined in RPSL (see draft-ietf-rps-rpsl-02). To make use of the route grouping functionality supplied by the "route-set:" object, we suggest the following extension be considered: 6. In the "aut-num:" object, allow "cross-nfy:" to be a multiple attribute, and in addition to specifying who should be notified, specify for which route-sets. If no route-sets are specified, then the notification will take place for all route objects with the given "aut-num:" as origin. [private discussion Joachim Schmitz]

1 0

RIPE Database: New format for reply messages
by Chris Fletcher 08 Jul '97

08 Jul '97

As announced at the RIPE 27 database working group in Dublin we have a new format for the automatic replies to RIPE database udpates. The new format is more consise and clearly states in the body and the subject line of the reply whether the update was completely sucessful or whether part of it failed. The new code is running on the test database for you to try out. Please forward comments to ripe-dbm(a)ripe.net. Updates to the test database should be sent to <test-dbm(a)ripe.net>. You should use a source of 'TEST' rather than the usual 'RIPE'. eg person: Test Person address: TEST phone: +31 20 535 4444 fax-no: +31 20 535 4445 e-mail: bit-bucket(a)ripe.net nic-hdl: AUTO-1 changed: chris(a)ripe.net 970704 source: TEST To query the test database invoke whois with the -h argument ... % whois -h test-whois.ripe.net TP1-RIPE person: Test Person address: TEST phone: +31 20 535 4444 fax-no: +31 20 535 4445 e-mail: bit-bucket(a)ripe.net nic-hdl: TP1-RIPE changed: chris(a)ripe.net 970704 source: TEST We would like to put this software into service on the RIPE database in a week's time on Monday 14th July. If you have automatic scripts that parse the old format and you can't update them in time please let us know (ripe-dbm(a)ripe.net). Regards, Chris Fletcher RIPE NCC

2 1

slow indexdb
by Gabor Kiss 26 Jun '97

26 Jun '97

'indexdb -s RIPE' runs 8 hours on a 4 processor SparCenter 2000. Is this normal? What are your experiences? Gabor

3 2

questions
by HJ Hwang 20 Jun '97

20 Jun '97

Hi, I am trying to use ripe-dbase-2.0.3. Document says that db.1.85.tar.gz is used for this database software. So I got this software and compile and then perl4.036 version. I installed db.1.85 successfully and as the result I copied dbm.h and ndbm.h to /usr/include. And then, I tried to compile perl4.036 but, "dbm_clearerr ld symbol not found " error occured. After looking at dbm.h, I found there wasn't dbm_clearerr there. How do I do ? So I am in deadlock status. Please Help me !! Isn't any other method to run ripe database software using perl5.004 ? Thanks in advance ! -- * _ _ ||_|| ||_| Hwang, Hyojung / jung(a)jean.kotel.co.kr / 82-2-526-5235 |_| _|_ Internet Team / Multimedia Laboratory / KT R&D Group |_| 137-792, 17 Umyun-Dong, Seocho-Gu, Seoul, Korea

2 1

None
by Remy Longueve 20 Jun '97

20 Jun '97

help

1 0

Some problem !!
by HJ Hwang 20 Jun '97

20 Jun '97

Hi, My name is Hwang, HyoJung. I am working for Korea Telecom. I am trying to install ripe-dbase software on Solaris 2.5. I did install perl5.004 and I did install db.1.85. But when I did make of ripe-dbase software, "version missmatch" error occured. So I did try to install perl4.036 on Solaris 2.5. This time, I also ran into the problem. there are no dbm_clearerr function in db.1.85. How do I you install perl4.036 and ripe-dbase software ? Please if you know why this occurs and how to do to solve the problem, let me know it. Thanks in advance !! -- * _ _ ||_|| ||_| Hwang, Hyojung / jung(a)jean.kotel.co.kr / 82-2-526-5235 |_| _|_ Internet Team / Multimedia Laboratory / KT R&D Group |_| 137-792, 17 Umyun-Dong, Seocho-Gu, Seoul, Korea

1 0

(no subject)
by N Gnanamurugan 07 Jun '97

07 Jun '97

subscribe db-wg. __________________________________________________ N Gnanamurugan, Engineer, Gulfsat Communications Company, P.O Box 2400, | email:murugan@gulfsat.com Safat, | Ph:+965-328-0770 Kuwait. | Fax:+965-328-1311 _________________________________________________

1 0

New Document available: RIPE-157
by RIPE NCC Document Annoucement Service 30 May '97

30 May '97

New/Revised RIPE Document Announcement -------------------------------------- A new document is available from the RIPE document store. Ref: ripe-157 Title: RIPE NCC Database Documentation Author: A. M. R. Magee Date: 29 May 1997 Format: PS=1145210 TXT=187252 Obsoletes: ripe-153 Ref: ripe-153 Title: RIPE NCC Database Documentation Author: A. M. R. Magee Date: 28 Jan 1997 Format: PS=788674 TXT=137793 Obsoletes: Version 1.2 of the document. Obsoleted by: ripe-157 Short content description ------------------------- This is an updated version of ripe-153. It includes information on the new whois -v feature, the new output from whois -t, hierarchical authorisation of the creation of inetnum objects, mirroring the RIPE Database, appendixes containing descriptions of the attributes and the objects. Comments to <ripe-dbm(a)ripe.net>. Accessing the RIPE document store --------------------------------- You can access the RIPE document store via the World Wide Web; RIPE-157 is available from the WWW at the following URL: http://www.ripe.net/docs/ripe-157.html The RIPE document store is also available via anonymous FTP to ftp.ripe.net, in the directory ripe/docs. The full filenames of the new documents on the FTP-server are: ripe/docs/ripe-157.ps for the PostScript version ripe/docs/ripe-157.txt for the plain text version Documents can also be retrieved from the RIPE document store using a mail server program. For more information on how to use the program, send email to <mail-server(a)ripe.net> with "send HELP" in the body text. If you have anymore questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration.

1 0

RIPE 27 Database Workshop Slides.
by RIPE Database Administration 29 May '97

29 May '97

Dear Colleagues, The slides of the Database Workshop which was run at RIPE 27 are available at: <http://www.ripe.net/meetings/ripe/ripe-27/pres/db> <ftp://ftp.ripe.net/ripe/presentations/ripe-m27-orange-magee-db-workshop.ps.…> If you have anymore questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration.

1 0

New Database Document
by RIPE Database Administration 22 May '97

22 May '97

Dear Colleagues, The Database Document, ripe-153, is obsolete. It is superceded by ripe-157, available at: <http://www.ripe.net/db/doc.html> <ftp://ftp.ripe.net/ripe/docs/ripe-157.ps> <ftp://ftp.ripe.net/ripe/docs/ripe-157.txt> Help on using whois is available by doing whois HELP Help on creating, updating and deleting objects is available by sending mail to <auto-dbm(a)ripe.net> with HELP in the Subject line. If you have anymore questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration.

1 0

subscribe db-wg
by Dieterle 21 May '97

21 May '97

subscribe db-wg

1 0

(rough) draft agenda for DB-WG mtg. during RIPE 27
by Wilfried Woeber, UniVie/ACOnet 17 May '97

17 May '97

Dear all! This is the initial, and still very rough, draft of a proposed agenda for the RIPE DB-WG Meeting, currently scheduled for Tue, May 22. Due to the fact that the workload here at home is extremely high at the moment, and that we only got a 60min slot assigned, I have to be considerably less ambitious this time. Please bear with me. Here you go... See you in Dublin, Wilfried. ~~~~~~~~~~~~~~~~~~~~ Proposed Agenda for the Database Working Group Meeting RIPE-27, May 1997, Dublin 1st Draft, May. 17, 1997 A. Administrative stuff (W.Woeber, chair) 10min - volunteering of the scribe - WG-agenda bashing - handling of minutes and actions B. New functionality, proposals, reports 30min - Database Referral mechanism (Q: rhois compatibility ?) - Database consistency - Security and Authentication (if time permits) B. DB-SW status report by the RIPE NCC (DB-SW maintainers, RIPE-NCC) 10min - operational changes (handles mandatory) - new functionality implemented and bugs fixed - documentation (current status, progress) - feedback from the users (DB WS usefulness?) G. General input from other WGs 10min - Hier. Authorisation and Notification in the RR (Routing WG) Z. AOB WW, 17-MAY-1997 21:22:05 -------------------------------------------------------------------------- Wilfried Woeber : e-mail: Woeber(a)CC.UniVie.ac.at Computer Center - ACOnet : Vienna University : Tel: +43 1 4065822 355 Universitaetsstrasse 7 : Fax: +43 1 4065822 170 A-1010 Vienna, Austria, Europe : NIC: WW144 --------------------------------------------------------------------------

1 0

whois -v option.
by RIPE Database Administration 16 May '97

16 May '97

Dear Colleagues, A new feature is available in the RIPE Database; "whois -v". This option displays detailed information of any type of object that is accepted by the RIPE database. Each attribute of the object is briefly described, as is the object itself. A more detailed template of the object is also given. A full list of objects is available by typing "all" after the "-v" option. The new, detailed template only is available using "whois -t". A full list of objects may be obtained by typing "all" after the "-t" option. The new template lists each of the attributes of an object together with a list of flags. Flags are enclosed in square brackets and separated within the list by white space. The flags are ... [mandatory] - An object MUST NOT contain zero instances of this attribute. [optional] - An object MAY contain zero instances of this attribute. [single] - An object MUST NOT contain more than one instance of this attribute. [multiple] - An object MAY contain more than one instance of this attribute. [look-up key] - Attribute is indexed. [inverse key] - Attribute is in the 'reverse' index. (use -i) [primary key] - Attribute is (part of) the primary key. [primary/look-up key] - Attribute is both indexed and is (part of) the primary key. The "-t" option does not display a list of obsolete attributes, as before. The "-v" option must be used. Example: [an-lar]$ whois -h whois.ripe.net -t person person: [mandatory] [single] [primary/look-up key] address: [mandatory] [multiple] [ ] phone: [mandatory] [multiple] [ ] fax-no: [optional] [multiple] [ ] e-mail: [optional] [multiple] [look-up key] nic-hdl: [mandatory] [single] [primary/look-up key] remarks: [optional] [multiple] [ ] notify: [optional] [multiple] [inverse key] mnt-by: [optional] [multiple] [inverse key] changed: [mandatory] [multiple] [ ] source: [mandatory] [single] [ ] To use "whois -v", you will need a new whois client. The new version of the whois client is available at <ftp://ftp.ripe.net/tools/ripe-whois-tools-2.1.tar.gz> If you have anymore questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration.

3 2

whois -v option.
by RIPE Database Administration 16 May '97

16 May '97

Dear Colleagues, A new version of the C client for whois is available at ftp://ftp.ripe.net/tools/ripe-whois-tools-2.2.tar.gz Thanks to Steven Bakker for the patch. However, the problem appears to be system dependent as I never had a core dump with "whois -t xyz" or "whois -v abc". If you have any questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration.

2 1

Hierarchical Authorisation in the RR
by Carol Orange 15 May '97

15 May '97

Hi again, Here is another brief proposal for your perusal. This time for an implementation of hierarchical authorization in the Routing Registry. The mechanism described below will implement a hierarchy in the RR for which I believe there is consensus. The ideas in the proposal below were sifted from discussions which took place on the <routing-wg(a)ripe.net> mailing list and in the Routing WG meetings at RIPE-25 and RIPE-26. Please review this if you have time. We are hoping to have a go ahead on this from the Routing and Database WG's at or shortly after RIPE-27. Greetings, Carol Orange RIPE NCC -------------------------------------------------------------------- Hierarchical Authorisation in the RR Proposal for an Implementation Carol Orange, May 1997 At the January meeting of the Routing WG in Amsterdam, various possible hierarchies for authorization in the Routing Registry (RR) were considered. Whereas extensive discussion took place on the extent to which authority can be established in the RR, there was clear agreement that the maintainer of an AS should have authority over what routes are announced with a given aut-num in the "origin:" attribute. In the following, we specify an implementation to support the authority of "aut-num:" maintainers to determine who can announce routes under their AS. The mechanism can be extended as the need arises and consensus on other forms of authorization is achieved. For more information on the discussions leading up to this proposal, see: http://www.ripe.net/wg/routing/haro-d.html. Implementation -------------- If you (or your organization) manages an AS, then you should have authority over the routes announced in your AS. This can be implemented if we: a) add a "mnt-lower:" attribute to the aut-num object b) allow routes to be announced with a given "origin:" by those given authority as defined in the mntner object specified in the "mnt-lower:" attribute of the aut-num object. Example ------- If we add a "mnt-lower:" attribute to the aut-num object of the RIPE NCC, then only those who know what peEw8Gb4xBNqI encrypts can add and remove routes originating in AS3333. ---- aut-num: AS3333 ... mnt-lower: AS3333-MNT ... ---- mntner: AS3333-MNT descr: RIPE-NCC Maintainer ... auth: CRYPT-PW peEw8Gb4xBNqI ... ---- route: 193.0.0.0/23 descr: RIPE-NCC origin: AS3333 ... Summary ------- Other forms of hierarchical authorization and notification can be implemented in the future if a well defined hierarchy can achieve consensus. To provide some initial functionality which may meet the needs of many RR users, we propose to implement the above in the short term.

2 2

Re: whois -v option.
by Steven Bakker 15 May '97

15 May '97

==> From: RIPE Database Administration <ripe-dbm(a)ripe.net> ==> Thu, 15 May 1997 13:02:17 +0200 > To use "whois -v", you will need a new whois client. The new > version of the whois client is available at > > <ftp://ftp.ripe.net/tools/ripe-whois-tools-2.1.tar.gz> And it still goes "boom" when you try to use "whois -v" or "whois -t": kappa-~ & whois -v netnum zsh: 5855 segmentation fault (core dumped) whois -v netnum I have a patch for it (incidentally identical to the one I sent in last September ;-), which I'll attach below. It's basically a tiny error in the argc/argv handling. Cheers, Steven

2 1

X-Notification in the RR
by Carol Orange 15 May '97

15 May '97

Hi all, The following is a proposal for a cross notification mechanism in the Routing Registry. I post it for your perusal and hope to discuss it during the Routing WG meeting in Dublin. The initial suggestion was made by Daniel Karrenberg at the Routing WG meeting in January, and discussions with Joachim Schmitz and Wilfried Woeber have helped me work it into the following proposal. Such a mechanism was given a high priority by those who participated in the recent RIPE Database priority survey, so we are hoping to reach consensus on this at or shortly after RIPE-27. Greetings, Carol Orange RIPE NCC ---------------------------------------------------------------------- Cross Notification in the Routing Registry A Proposal Carol Orange, May 1997 At the January meeting of the Routing WG in Amsterdam, various possible hierarchies for authorization in the Routing Registry (RR) were considered. During the discussion, a suggestion was made that "hierarchical notification" would be useful even in cases where authority is not well defined. This suggestion enjoys strong support of both the Routing WG and the Database WG. In the following, we propose a simple cross notification mechanism which will allow routing registry users to be notified of overlapping route announcements in the RR. Route Announcement Notifications in the RR ------------------------------------------ Suppose you announce a route in your address space covering some /24. It may be useful when you make the announcement, to know that the /24 has already been announced in another AS. Likewise, should the user of that address space later decide to become multi-homed, it may be interesting to know that as well. Moreover, overlapping routes are often unintentionally announced in a single AS. In all of these cases, it would be useful to get a notification from the RR. Consider the route object: route: [mandatory] [single] [primary/look-up key] descr: [mandatory] [multiple] [ ] origin: [mandatory] [single] [primary key] hole: [optional] [multiple] [ ] withdrawn: [optional] [single] [ ] comm-list: [optional] [multiple] [ ] advisory: [optional] [multiple] [ ] remarks: [optional] [multiple] [ ] notify: [optional] [multiple] [inverse key] mnt-by: [mandatory] [multiple] [inverse key] changed: [mandatory] [multiple] [ ] source: [mandatory] [single] [ ] Suppose we add a new field called "x-notify:" (cross-notify) which a) can contain either + a NIC-handle pointing to a person or role object, the email address of which will get the notification; or + a mntner ID pointing to the maintainer who will be notified b) is optional The contact in "x-notify:" will receive an e-mail whenever + a route is announced with overlapping address space + the address range in the route being added (with the "x-notify:" attribute) overlaps any address space currently announced in the RR. Issues ------ Does this cross notification satisfy the needs of network operators? Due to the global nature of the RR, some effort will be involved to provide notifications in a timely manner. What is an acceptable upper bound for cross notifications to be sent out? Using a NIC-handle or mntner for notification is different than the traditional e-mail address. However, it simplifies data maintenance for users significantly.

1 0

nic-hdl mandatory
by RIPE Database Administration 15 May '97

15 May '97

Dear Colleagues, The nic-hdl attribute is now mandatory in the person object. If you have any questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration.

1 0

A DNS Database Referral Mechanism
by Carol Orange 14 May '97

14 May '97

Hi all, Below is a proposal put together by Wilfried Woeber and myself to address the need for a referral mechanism for domain name queries in the RIPE database. The mechanism is designed to address the immediate needs of those trying to get domain related information, and of DNS administrators trying to provide it. Certainly the mechanism may be adapted with time (see the three options below), however we choose to start with something that can provide the required functionality immediately. We hope to discuss this proposal on the mailing lists and in the database and dns working group meetings in Dublin. Carol Orange RIPE NCC --------------------------------------------------------------------- RIPE Database Referral Mechanism for Domain Queries - A Proposal Carol Orange & Wilfried Woeber, April 1997 ------------------------------------------ In the following, we propose a mechanism for forwarding whois queries to the appropriate database if the RIPE database does not contain the authoritative data for a given tree in the DNS name space or subset thereof. It is thus applicable for the "domain" object only: domain: [mandatory] [single] descr: [mandatory] [multiple] admin-c: [mandatory] [multiple] tech-c: [mandatory] [multiple] zone-c: [mandatory] [multiple] nserver: [optional] [multiple] sub-dom: [optional] [multiple] dom-net: [optional] [multiple] remarks: [optional] [multiple] notify: [optional] [multiple] mnt-by: [optional] [multiple] mnt-lower: [optional] [multiple] changed: [mandatory] [multiple] source: [mandatory] [single] The goal of the mechanism is to provide a means for TLD domain administrators (and other "high-level" domains) to enable users to obtain an authoritative response for a domain object query sent to the RIPE database, regardless of whether the data for the domain is maintained in the RIPE database itself. Algorithm --------- If the following query is submitted to the RIPE database: whois [flags] xxx.yyy.zz the algorithm will work as follows: ------- NAME="xxx.yyy.zz" until (NAME = "") { If object with domain = NAME found, If object contains referral (see "Referral" below) forward query (see "Forward" below) Else return object Else Strip(NAME) (xxx.yyy.zz -> yyy.zz, etc) } NOTES ----- 1. We move up the tree here, and return the next level answer if present. If the query requests "tuintje.cwi.nl" (Piet, do you mind?), they currently get "No entries found ...". In the new mechanism, they would get: domain: cwi.nl . . . mnt-by: NL-DOMREG changed: hostmaster(a)domain-registry.nl 19950227 source: RIPE Moreover, if the object for cwi.nl contains a referral, the query would be passed on to the specified server as explained below. 2. The algorithm will be set off by any query which causes a search in the domain object index. This means any query with "-T domain", or any general query (no -T flag) with something that "looks like a domain". Referral -------- A whois referral can be entered in a "refer" attribute by specifying the type of database in which the data is maintained, and the domain name of the server that should be queried: refer: <type> <host> <port> where <type> is one of RIPE, InterNIC or SIMPLE, indicating which style of whois service is provided. <host> is the DNS name of the whois service. <port> is the TCP port number (optional: 43 is the default). Examples: (CWI again): refer: RIPE domain-registry.nl 43 (InterNIC): refer: InterNIC whois.internic.net (Generic): refer: SIMPLE my.dns.tld Initially queries for the three types of databases shown here would be supported. If RIPE is specified, then the initial query together with all arguments would be passed to the specified whois server. If an InterNIC database is specified, then a query expected by the InterNIC database software would be generated for the specified domain name. The same would be true of "SIMPLE". In that case a simple query for the domain name would be passed on. If in the future, another domain name database is implemented with a given query language, it can be added. Forward ------- If the TLD object contains a whois referral, we can a) query the server, and pass the response to the requester, preceded by a comment of the form: "The following data has been obtained from domain-registry.nl". b) pass the referral to the requester c) send the query to the server with the address of the requester a) has the advantage of giving the user an immediate answer, and requires that only the RIPE database software be modified, not that of the TLD admins. Nothing has to be changed there. a) has the disadvantage that the RIPE server can become a bottleneck. However, local mirrors of referred to databases can be set up on a RIPE NCC server to alleviate this problem. Summary ------- We propose the above mechanism with (option (a) request forwarding) be used for TLD/domain referrals. We believe it will make the domain part of the database more transparent to users and easier to manage for TLD administrators. Acknowledgements ---------------- We would like to thank Chris Fletcher, Daniel Karrenberg, and David Kessens for useful comments.

3 6

Re: A DNS Database Referral Mechanism
by Wilfried Woeber, UniVie/ACOnet 14 May '97

14 May '97

Hi Carol, Blasco, Piet et.al.! >We actually discussed this point a bit in a slightly different light. >An important question is: should the forward type be up to the object >maintainer, to the whois server providing the referral (in this case >RIPE), or the whois client, or end user. > >As proposed here, it is the whois server, but could be modified in the >future to also be the whois client. > I think this touches on a line of discussion that we already began a while ago, but became forgotten later: What's the mechanism and who does maintain (e.g. supply, control, register) the necessary information? >The object maintainer can in fact always prevent automatic request >forwarding by putting the referral information in a remarks field. > >However, if such a mechanism should become popular, then it may be >suitable that whois clients be developed that can parse the refer >field and resend the request to the appropriate server. If we allow >the object maintainer to determine this in the <forward-type> field, >then it actually limits flexibility in the future. When and if we decide to eventually implement that flexibility (and we already had some private exchange on this aspect), then my first reaction would be to move that to some sort of maintainer object!? Initially, given the small set of affected objects, this can probably be sorted out infomally, but as soon as we open the flood gates... >Or is my thinking twisted? I don't think so, but mine might be :-) Wilfried.

1 0

A DNS Database Referral Mechanism
by Wilfried Woeber, UniVie/ACOnet 13 May '97

13 May '97

>Subject: A DNS Database Referral Mechanism >Date: Tue, 13 May 1997 10:44:31 +0200 >From: Carol Orange <Carol.Orange(a)ripe.net> Dear Carol! >Hi all, > >Below is a proposal put together by Wilfried Woeber and myself to >address the need for a referral mechanism for domain name queries in >the RIPE database. Thank you very much for putting that into proper wording and having it distributed to the groups. I feel it is appropriate to add that, while Carol and me did some finishing and sanity checking, we already received a considerable amount of very valuable comments and ideas from quite a few others. A big thank you to all of them! I thus urge the community to have a focussed discussion of the proposal, leading to a conclusion and implementation decision at the DNS & DB-WG meeting(s) is Dublin. Thanks, Wilfried. -------------------------------------------------------------------------- Wilfried Woeber : e-mail: Woeber(a)CC.UniVie.ac.at Computer Center - ACOnet : Vienna University : Tel: +43 1 4065822 355 Universitaetsstrasse 7 : Fax: +43 1 4065822 170 A-1010 Vienna, Austria, Europe : NIC: WW144 --------------------------------------------------------------------------

1 0

NIC-Handle Mandatory.
by RIPE Database Administration 06 May '97

06 May '97

Dear Colleagues, >From 15th May, new person objects must have a NIC-handle. Normally, RIPE nic-hdl's are used. RIPE format NIC-handle's may be obtained using the "AUTO-1" process (see below). If a person being registered in the RIPE database currently holds a registered NIC-handle assigned by the the InterNIC or the APNIC, it can be used as the NIC-handle in the RIPE database person object. If, in conjunction with the work on registry database exchange taking place in the RIDE working group of the IETF, the list of databases, and in particular the list of acceptable NIC-handles should be expanded or otherwise modified, the above statement will of course be modified accordingly. If you have anymore questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration. How to get a RIPE NIC-Handle ------------------------- You can get a RIPE NIC-handle by putting the following in the NIC handle field of the person object update: nic-hdl: AUTO-1 OR nic-hdl: AUTO-1YourInitials The second case advises the database software to use YourInitials (no more then 4 characters) for build- ing the NIC handle while the first case asks the database software to find the initials itself. You can use the same identifiers (AUTO-1 or AUTO-1YourInitials) in the same update message in other objects as a reference. The database software will then fill in the freshly assigned NIC handles in the objects. Note that you can also use other numbers (example: AUTO-2) so that you can update more person objects and objects that reference the persons in one E-mail message. Example: domain: over.ripe.net admin-c: AUTO-1 tech-c: AUTO-2 [ ... stuff deleted ...] person: Ambrose Magee nic-hdl: AUTO-1 [ ... stuff deleted ...] role: RIPE DBM nic-hdl: AUTO-2 [ ... stuff deleted ...] N.B. To obtain a NIC-handle for a role object, use the same procedure as for a person object. Further information is available from: <http://www.ripe.net/docs/ripe-153.html> <ftp://ftp.ripe.net/ripe/docs/ripe-153.ps> <ftp://ftp.ripe.net/ripe/docs/ripe-153.txt>

1 0

Inaddr Robot <auto-inaddr@ripe.net>
by Carol Orange 22 Apr '97

22 Apr '97

Dear All, First off, my apologies to all who get more than one copy of this message. I'm hoping not to miss anyone who may be effected. As many of you may know, we have long been planning to more fully automate our inaddr robot <auto-inaddr(a)ripe.net>. Well, we've gotten around to it and are happy to announce we will soon be doing almost all processing automatically, with just a touch of human intervention at the end to make certain our robot behaves as expected. The result will be a substantial reduction in rather tedius work for the NCC staff, and consistent short response times (one working day) for inaddr requests. The new robot, programmed by Mal Morris, will be installed on *** Tuesday, May 6, 1997 *** It is essential that users take the following changes into account: 1. The domain or inetnum database object must be in the RIPE database before the request is submitted to <auto-inaddr(a)ripe.net>. 2. As is required in ripe-140 (Section 3.4.2), the "status" field must be filled in properly for any inetnum associated with the request. The robot will complain and fail if either of these does not hold. We hope the new robot serves you well, Carol

1 0

Re: Proposal for db change
by Wilfried Woeber, UniVie/ACOnet 18 Apr '97

18 Apr '97

David, => Honestly, how about offering that for as long as there is no => authentication interaction? => => As soon as we hit the wall we might have a solution ready or we might => have to *not* offer that comfort for signed or checksummed updates? = =Please, don't get me wrong: I don't have a problem with John's proposed =feature but I wanted to make clear which trade-offs are made and that =there are certainly a few disadvatages involved that might not be obvious =at first sight. that was perfectly understood here. Cheers, Wilfried. -------------------------------------------------------------------------- Wilfried Woeber : e-mail: Woeber(a)CC.UniVie.ac.at Computer Center - ACOnet : Vienna University : Tel: +43 1 4065822 355 Universitaetsstrasse 7 : Fax: +43 1 4065822 170 A-1010 Vienna, Austria, Europe : NIC: WW144 --------------------------------------------------------------------------

1 0