- db-wg - mailman.ripe.net

Re: Changing rrdb objects in the future
by Wilfried Woeber, UniVie/ACOnet 14 Apr '97

14 Apr '97

=At 02:46 PM 4/08/97 -0400, you wrote: =>At the last ripe meeting it was agreed to track object history =>including UTC time zone, although I'm not entirely clear what this =>means. Presumebly it means that the changed field will be stamped =>automatically by the db software with utc time and the from =>line of the email. = =That would certainly sort it out. The RIPE-DB WG proposed to have the db update mechanism *automatically* tag objects being updated with a UTC timestamp (and some other data). This data would *not* be under the control of the submitting entity. It is not meant to replace or modify any data in the changed: field as supplied by the submitting entity. This change is already on the RIPE-NCC's to-do-list for implementation. => I think another holdup is the issue of the millenium approaching. =>The current date format (ie, YYMMDD) will be insufficient for those =>future dates. And so this problem must be fixed too. => I would think that YYYYMMDD is an obvious solution. The RIPE DB production software already accepts objects with the extended 8 character format. =That will take care of the y2k problem, but won't help those of us in =unusual time zones... For that we would need something like YYYYMMDDHHMMZZZ =(ZZZ=time zone). This is a bit more of a mouthful, of course. I think local time details (time zones, DST, whatever) should become a non-issue as soon as we get the changed: data decoupled from the UTC time stamps as applied by the db update process? =>The db software =>will be converted from ripe181 to rpsl over the coming 6 months. =>Maybe when the db's are officially changed to rpsl we can update =>the objects from YYMMDD --> YYYYMMDD format. => =>ps Yes, this is exactly the correct forum to bring this up. = =Good :) Although I wonder whether it would be useful and more consistent to convert all the exiting objects from the 6 character to the 8 character format and to expand the 6 character format on the fly during the update processing... 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

Tools based on whois -t output.
by RIPE Database Administration 09 Apr '97

09 Apr '97

Hi, all ! We are contemplating changing the output of the "-t" option of "whois" to give more information. Please inform us if you write or use tools based on this output. Example: The current output for "-t person" is person: [mandatory] [single] address: [mandatory] [multiple] phone: [mandatory] [multiple] fax-no: [optional] [multiple] e-mail: [optional] [multiple] nic-hdl: [optional] [single] remarks: [optional] [multiple] notify: [optional] [multiple] mnt-by: [optional] [multiple] changed: [mandatory] [multiple] source: [mandatory] [single] the new output would be person: [mandatory] [single] [look-up key] address: [mandatory] [multiple] * phone: [mandatory] [multiple] * fax-no: [optional] [multiple] * e-mail: [optional] [multiple] [look-up key] nic-hdl: [optional] [single] [look-up key] remarks: [optional] [multiple] * notify: [optional] [multiple] [inverse key] mnt-by: [optional] [multiple] [inverse key] changed: [mandatory] [multiple] * source: [mandatory] [single] * where [look-up key] indicates that you use this attribute as an argument to whois and [inverse key] indicates that you can use the attribute as an argument to whois -i. If you have anymore questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration.

3 3

Database Inconsistencies
by Carol Orange 07 Apr '97

07 Apr '97

Hi all, At the January meeting of the RIPE Database WG, I promised to post the reports on RIPE Database inconsistencies produced by the group of students at the Free University last fall. The reports are now on our ftp site and can be found in: "Problem Definition" ftp://ftp.ripe.net/ripe/dbase/info/consistency1.ps ftp://ftp.ripe.net/ripe/dbase/info/consistency1.txt "The RIPE Database: Constraints and possible Solutions" ftp://ftp.ripe.net/ripe/dbase/info/consistency2.ps ftp://ftp.ripe.net/ripe/dbase/info/consistency2.txt Greetings, -- Carol

1 0

"Fake" records in RIPE DB
by RIPE Database Administration 07 Apr '97

07 Apr '97

Hi, Gabor ! I have traced the source of the problem and have contacted those responsible. Thank you for informing us. If you have anymore questions, please contact <ripe-dbm(a)ripe.net>. Regards, Ambrose Magee ____________________________ RIPE Database Administration.

1 0

fake records in RIPE DB
by Gabor Kiss 07 Apr '97

07 Apr '97

Dear Hostmaster, FYI. Last week the following fake records was created in the name of you: inetnum: 194.117.120.201 - 194.117.120.201 netname: SAP-HUNGAROPHARMA descr: Hungaropharma RT country: HU admin-c: VG48-RIPE tech-c: VG48-RIPE status: ASSIGNED PA changed: hostmaster(a)ripe.net 960702 source: RIPE inetnum: 194.117.120.204 - 194.117.120.204 netname: SAP-FOEVAROSI descr: Foevarosi Vizmuevek Rt. country: HU admin-c: SB568-RIPE tech-c: SB568-RIPE status: ASSIGNED PA changed: hostmaster(a)ripe.net 960702 source: RIPE inetnum: 194.117.120.216 - 194.117.120.216 netname: SAP-MMG descr: MMG Automatikai Muevek Rt. country: HU admin-c: TJ308-RIPE tech-c: TJ308-RIPE status: ASSIGNED PA changed: hostmaster(a)ripe.net 960702 source: RIPE inetnum: 194.117.120.231 - 194.117.120.231 netname: SAP-MVM descr: MVM RT. country: HU admin-c: AK625-RIPE tech-c: AK625-RIPE status: ASSIGNED PA changed: hostmaster(a)ripe.net 960702 source: RIPE I think some stronger authentication/authorization method is necessary as well as better education of local registries. :-) Regards Gabor

2 1

Results of vote
by Carol Orange 24 Mar '97

24 Mar '97

Hi All, Recently I sent out a message requesting your input regarding the prioritization of the on the RIPE Database work items on the NCC to do list. We received input from 16 of the 133 members (12%) on the db-wg mailing list, and would like to thank those who participated for giving us input. Based on your responses, the list of work items should be performed in the following order. The votes have been weighted according to the preference shown for an item. item votes priority ---- ----- -------- 17. [36] 1 9. [33] 2 8. [30] 3 23. [28] 4 22. [25] 5 7. [21] 6 13. [18] 7 12. [17] 8 4. [10] 9 21. [8] 10 3. [6] 11 25. [1] 12 The list is attached again below for reference. We will be working in the order specified above during the coming months and will report on our progress at the Dublin meeting. We may deviate from this order from time to time when it will speed up our progress significantly. It's quite useful to know what your priorities are. To gather this information more effectively in the future, we'd also like any input or suggestions you may have on the manner that this poll was conducted. For example: Do you like being polled on these matters? Was the voting mechanism clear and easy to follow? Do you have suggestions to improve it? Any input you have will be taken into account in the future. -- Carol --------------------------------------------------------------------- The list of items is now labeled with the priority from the list above (the number in [] is the priority with [1] being the top priority, and [12] being the lowest). RIPE NCC Database Work List --------------------------- [11] 3. Check the content of admin-c field during the creation and modification of objects to assure the content refers to a person object (and not a role). [9] 4. Take steps to remove obvious garbage (e.g. "see remarks") from the admin-c, tech-c, and zone-c fields. In general, such fields should contain a valid NIC handle referring to a person or role object. However, which NIC handles are valid is not always obvious in a global registry and that definition falls under one of the open issues. [6] 7. Hierarchical authorization in the routing registry. Note: this item depends on a clear definition on how the hierarchy should be defined. It was however decided that as soon as the Routing WG comes with a definition with which they are satisfied, it should "just be" implemented, without further discussion by the Database WG. [3] 8. Hierarchical notification in the routing registry. This would be a notification method which would allow someone to be notified when routes are announced. Please see notes under number 7 above. [2] 9. Currently insufficient information is sent to the person in a "notify" field when an object is modified by that person. It was decided that complete information on the modification should be sent to the person in the notify field, regardless of the options used to send it, and regardless of who made the changes. This is primarily to facility consistent administrative services. If the person in the notify field made the update, only one message should be sent, but it should contain all information. [8] 12.Implement (borrow) extended as-in/as-out features described by Cengiz Alaettingoglu in the RPSL draft: ftp://ftp.ripe.net/internet-drafts/draft-ietf-rps-rpsl-00.txt [7] 13.In response to whois queries, show the name associated with the NIC handle admin-c, tech-c, and zone-c fields. [1] 17.When objects are created or updated, check the validity of all referenced objects, before accepting the modification. [10] 21.Logging of syntax errors to collect information which might be used to improve the user interface. [5] 22. Track object history including UTC time zone, and as accurately as possible, who made the change. [4] 23. Define and implement a referral mechanism for TLDs. [ ] 24. Verbose object descriptions with "whois -tv". This is underway, and on the list for completion. [12] 25. Syntactic cleanup: remove ripe-181 line continuation.

1 0

REMINDER - Request for Input
by Carol Orange 13 Mar '97

13 Mar '97

Hi again, Last Friday, I posted a message requesting your input on the database activities at the RIPE NCC in the coming months. Tomorrow was set as the deadline to respond. As the input to this point has not been overwelming, I thought I'd send you a reminder, and give you all the weekend to consider. The deadline has now been moved forward to Monday, March 17, 1997. Greetings and good weekend, -- Carol ---------------------------------------------------------------------- * * * V O T E H E R E * * * In the following, I give a short description of each work item. The numbering is consistent with those in the presentation slides. Thus there are gaps in the numbering for those open issues which need further discussion (B). In the following list, please assign a score to *exactly 5* of the items. Give a 1 to that which is most import to you, 2 to the next most important 3 4 5 for what you would like to see, but with a lower priority. You can use each of the above scores exactly once in the list below. If you have any questions, just send them to me. ________________________ | | Please respond before * Monday, March 17, 1997 * |________________________| Here is the list then: [ ] 3. Check the content of admin-c field during the creation and modification of objects to assure the content refers to a person object (and not a role). [ ] 4. Take steps to remove obvious garbage (e.g. "see remarks") from the admin-c, tech-c, and zone-c fields. In general, such fields should contain a valid NIC handle referring to a person or role object. However, which NIC handles are valid is not always obvious in a global registry and that definition falls under one of the open issues. [ ] 7. Hierarchical authorization in the routing registry. Note: this item depends on a clear definition on how the hierarchy should be defined. It was however decided that as soon as the Routing WG comes with a definition with which they are satisfied, it should "just be" implemented, without further discussion by the Database WG. [ ] 8. Hierarchical notification in the routing registry. This would be a notification method which would allow someone to be notified when routes are announced. Please see notes under number 7 above. [ ] 9. Currently insufficient information is sent to the person in a "notify" field when an object is modified by that person. It was decided that complete information on the modification should be sent to the person in the notify field, regardless of the options used to send it, and regardless of who made the changes. This is primarily to facility consistent administrative services. If the person in the notify field made the update, only one message should be sent, but it should contain all information. [ ] 12. Implement (borrow) extended as-in/as-out features described by Cengiz Alaettingoglu in the RPSL draft: ftp://ftp.ripe.net/internet-drafts/draft-ietf-rps-rpsl-00.txt [ ] 13. In response to whois queries, show the name associated with the NIC handle admin-c, tech-c, and zone-c fields. [ ] 17. When objects are created or updated, check the validity of all referenced objects, before accepting the modification. [ ] 21. Logging of syntax errors to collect information which might be used to improve the user interface. [ ] 22. Track object history including UTC time zone, and as accurately as possible, who made the change. [ ] 23. Define and implement a referral mechanism for TLDs. [ ] 24. Verbose object descriptions with "whois -tv". This is underway, and on the list for completion. [ ] 25. Syntactic cleanup: remove ripe-181 line continuation.

1 0

NCC Update and Request for Input
by Carol Orange 07 Mar '97

07 Mar '97

Hi Folks, The intention of this mail is to give you a short update of the current efforts taking place at the RIPE NCC with respect to database development and maintenance and to gather your input in setting our priorities for the spring of '97 which to our great relief is finally showing its face - even here in Amsterdam! As promised, we are hard at work getting the components in place to give you verbose object descriptions with a new whois flag. You can expect an announcement for this service in the very near future. Simultaneously, we are working on a number of bug fixes and minor requests. For example, we now allow the use of 4 digit years in the changed field. As promised, we are working towards a complete and revised version of the RIPE database documentation (ripe-153) which will include some new sections and appendices as well as reflect the feedback received from the user community. The comments received to date have been greatly appreciated and will be reflected in the next version. Feedback is still welcome and desired, so if you've got comments, do send them along. To prioritize our further activities in the spring of 1997, we would appreciate your input on the direction of database development which is most important in your view. At the meeting of the Database working group at the RIPE-26 meeting in January, I gave a presentation on the current list of open issues regarding the RIPE database. The slides in which the various items are described are posted in the following places: http://www.ripe.net/meetings/ripe/ripe-26/pres/db-update/ ftp://ftp.ripe.net/ripe/presentations/ripe-m26-orange-dbupdate.ps.gz At that meeting, noone could think of items of concern which were not already on the list. During the course of the discussion, we classified each issue as one of: A) just do it, and B) needs further discussion I promised to post the items under (A) for priority setting and under (B) for discussion. In this mail, I'd like to ask for your input on which items in the "To do" category (A) have your highest priority. To prevent excessive traffic, please send your response directly to me (orange(a)ripe.net). I will post the results I've received in the week of March 17, and the corresponding action plan we will follow at the NCC during the coming months. Greetings, Carol ------------------------------------------------------------------- * * * V O T E H E R E * * * In the following, I give a short description of each work item. The numbering is consistent with those in the presentation slides. Thus there are gaps in the numbering for those open issues which need further discussion (B). In the following list, please assign a score to *exactly 5* of the items. Give a 1 to that which is most import to you, 2 to the next most important 3 4 5 for what you would like to see, but with a lower priority. You can use each of the above scores exactly once in the list below. If you have any questions, just send them to me. ________________________ | | Please respond before * Friday, March 14, 1997 * |________________________| Here is the list then: [ ] 3. Check the content of admin-c field during the creation and modification of objects to assure the content refers to a person object (and not a role). [ ] 4. Take steps to remove obvious garbage (e.g. "see remarks") from the admin-c, tech-c, and zone-c fields. In general, such fields should contain a valid NIC handle referring to a person or role object. However, which NIC handles are valid is not always obvious in a global registry and that definition falls under one of the open issues. [ ] 7. Hierarchical authorization in the routing registry. Note: this item depends on a clear definition on how the hierarchy should be defined. It was however decided that as soon as the Routing WG comes with a definition with which they are satisfied, it should "just be" implemented, without further discussion by the Database WG. [ ] 8. Hierarchical notification in the routing registry. This would be a notification method which would allow someone to be notified when routes are announced. Please see notes under number 7 above. [ ] 9. Currently insufficient information is sent to the person in a "notify" field when an object is modified by that person. It was decided that complete information on the modification should be sent to the person in the notify field, regardless of the options used to send it, and regardless of who made the changes. This is primarily to facility consistent administrative services. If the person in the notify field made the update, only one message should be sent, but it should contain all information. [ ] 12. Implement (borrow) extended as-in/as-out features described by Cengiz Alaettingoglu in the RPSL draft: ftp://ftp.ripe.net/internet-drafts/draft-ietf-rps-rpsl-00.txt [ ] 13. In response to whois queries, show the name associated with the NIC handle admin-c, tech-c, and zone-c fields. [ ] 17. When objects are created or updated, check the validity of all referenced objects, before accepting the modification. [ ] 21. Logging of syntax errors to collect information which might be used to improve the user interface. [ ] 22. Track object history including UTC time zone, and as accurately as possible, who made the change. [ ] 23. Define and implement a referral mechanism for TLDs. [ ] 24. Verbose object descriptions with "whois -tv". This is underway, and on the list for completion. [ ] 25. Syntactic cleanup: remove ripe-181 line continuation.

1 0

route obj & notifications
by Schmitz＠RUS.Uni-Stuttgart.DE 14 Feb '97

14 Feb '97

Dear colleagues, yes, it is me again ;-) asking for comments on another element of hierarchical authorisation for route objects: The idea is to use a prefix based hierarchical scheme just as it is used for inetnum objects. However, it turned out that there are several difficulties in it. So, to get started the consensus at the wg session was: *not* to enforce a prefix based hierarchical scheme until these difficulties are solved, but instead generate notifications only. There are several proposals out there for notifications - to notify only if it is requested (by adding "notify" attributes to objects) - to leave inetnum objects out for the time being until their relation to route objects is clearly defined - notifications should only be done for creation of route objects not for changes or deletions (to prevent mail floods) - probably, the creator of route objects should also be notified of notifications for coordination purposes What are your feelings about this? Yes, no, maybe, missing items? Your comments are welcome! For further reference read/attack/tear apart/abuse the compilation at http://www.ripe.net/wg/routing/haro-d.html Regards Joachim _____________________________________________________________________________ Dr. Joachim Schmitz schmitz(a)noc.dfn.de DFN Network Operation Center Rechenzentrum der Universitaet Stuttgart ++ 711 685 5553 voice Allmandring 30 ++ 711 678 8363 FAX D-70550 Stuttgart FRG (Germany) _____________________________________________________________________________

1 0

route & aut-num objects
by Schmitz＠RUS.Uni-Stuttgart.DE 14 Feb '97

14 Feb '97

Dear colleagues, let's start with a very easy thing near to solution with hierarchical authorisation for route objects: the relation to aut-num objects. A route object references several other objects. Important to us are the references to the maintainer and the aut-num object (both are mandatory attributes of the route object): route: [mandatory] [single] descr: [mandatory] [multiple] origin: [mandatory] [single] ----> points to aut-num object ... mnt-by: [mandatory] [multiple] ----> points to mntner object changed: [mandatory] [multiple] source: [mandatory] [single] The aut-num object also contains a mandatory reference to a maintainer object. Up to now, the maintainers referenced in route object and aut-num object of same origin need not be the same. For hierarchical authorisation it would be nice to introduce a "mnt-lower" attribute in the aut-num object defining which maintainers may create route objects for the AS of the corresponding aut-num object. This allows control of one AS which parties may generate route objects with its origin. We already had consensus in the wg session that this is needed but did not yet decide precisely how to do it. The idea of using a "mnt-lower" attribute came up and I think it is easiest to implement. If there are no objections I will hand on the above proposal for imple- mentation in the database software in two weeks. For further reference read/attack/tear apart/abuse the compilation at http://www.ripe.net/wg/routing/haro-d.html Regards Joachim _____________________________________________________________________________ Dr. Joachim Schmitz schmitz(a)noc.dfn.de DFN Network Operation Center Rechenzentrum der Universitaet Stuttgart ++ 711 685 5553 voice Allmandring 30 ++ 711 678 8363 FAX D-70550 Stuttgart FRG (Germany) _____________________________________________________________________________

1 0