- db-wg - mailman.ripe.net

Announcement: RIPE Whois Database version 3.0 beta is out
by RIPE Database Administration 08 Dec '00

08 Dec '00

Dear All, [ Apologies for duplicate messages ] This mail is to inform you that version 3.0 beta of the RIPE Whois Database is out. Read on! You will be affected too. The RIPE Database is about to change its data format, from RIPE181 to RPSL (the Routing Policy Specification Language, see RFC2622). This will affect all object types, including "person", "domain" and "inetnum". The version 2 of the RIPE Database, the one currently in production on whois.ripe.net, will in the next months be replaced by version 3. All the entries in the RIPE Database will be translated into RPSL format, and that is the format you will get when you will be querying the Database. Don't be mistaken: for the moment, whois.ripe.net continues running with version 2 and answering to your queries in the old-fashioned way. But this will not last forever; in a few months, all will become RPSL. So if you use the Whois database, we encourage all of you to prepare for this change as soon as possible. We strongly suggest you to start to get used to RPSL and version 3.0 . In this aim, you can do the following: - Query rpsl.ripe.net at port 43. You will find there exactly the same objects that are in whois.ripe.net, mirrored live and in RPSL format. - If you have any script that runs by querying whois.ripe.net, test it by pointing it to rpsl.ripe.net and make it v3.0-compliant. - Try our CGI interface to the RPSL server: you can find it at http://www.ripe.net/ripencc/pub-services/db/reimp/. - Get to know the new update path: use our "sandbox" database for RPSL updates. Send your updates to <auto-rip(a)ripe.net> and query the corresponding database at rpsl.ripe.net, port 4343 . Remember that in a few months, that's the way the RIPE Database will be running; so the earlier you start working with it, the less likely you are to be unpleasantly surprised later. We are currently preparing additional resources to help ease the transition; we will announce them as soon as they are available, which will happen later this month. Please note that the software is still in a "beta" phase. So we need your feedback! If you happen to find something that looks like a bug when you are testing the RPSL database, don't hesitate... Send us a bug report at <dbrip(a)ripe.net> ! If you want to actively participate to the migration effort, you could join the "RIP migration Task force" and let yourself be heard. Just write to <ripe-dbm(a)ripe.net> and ask to join the Task Force. For the ones interested in viewing or trying the code, you can download it at: ftp://ftp.ripe.net/ripe/dbase/reimp/whois-rip-3.0beta1.tar.gz You can also join a discussion list related to version 3.0 beta of the RIPE Database Software. To subscribe to this list, send a message to <majordomo(a)ripe.net> with the line: subscribe db-beta in the body of your letter. It is a moderated mailing list so some delays in subscription are possible. Regards, The RIPE Database Group.

1 0

Re: Privacy, Broadband, & the Database
by Wilfried Woeber, UniVie/ACOnet 01 Dec '00

01 Dec '00

I've tried to sort some of the various aspects/expectations. Lets see.... => It depends on the level of "responsibility" and functionality granted to => and exercised by that end site. As I've said in a private mail already, => we should ask the question about the usefulness of "assigning" (in the => good old sense) very small amounts of addresses to sites which are tied => in to the services of their provider anyway. I guess most of the ADSL, => dial-up, cable-TV connection assignments sh/could be reviewed from that => point of view. = =The name and address info in the RIPE database needs to contain the =person (or organisation) responsible for a certain amount of IP space, =in my opinion. = =So, if an ISP wants to be responsible for the IP space it hands out to =customers, then the ISP should be free to fill in their own name and =address. This gives ISPs the freedom to fill the RIPE db with names =and addresses of customers (that saves them the hassle of dealing with =for example abuse coming from that site), or putting their own name =in it (meaning they have to respond actively to for example abuse reports, =which is the ISP's job anyway). 1) IP-Addresses are taken from a globally unique pool. Thus the global Internet community expects to have access to publicly accessible registry data to query the status (unassigned/reserved - allocated - assigned) for a particular address or address block or other unique resource. 2) The contact information listed for the individual resource serves 2 purposes: a) it offers info to get in contact with a human being to resolve connectivity problems (IP or application layer) between end sites b) it offers info to get in touch with a human being in case of operational problems that have an unwanted impact on the performance or security of networks other than the source of the problem (routing issues, spam, abuse). Ref: 1: While I still think that the open availability of that info is essential for the responsible and continued operation of the Internet, it can be argued whether contact info to get in touch with a human being has to be publicly available. In the long run, we probably have to deal with the authorities in one way or another (on a national and international level), to get the existence and maintainance of such a global registry recognized and endorsed. Ref: 2)a): While "we" certainly have become comfortable with the notion of being able to "talk" to everyone who is on the net (to resolve operational problems for our own applications or customers), we might have to let go here - and also to openly tell *our* customers, that we cannot help in case of problems when the "other end" opts to remain anonymous (by requesting contact info to be suppressed or, better yet, not to be collected at all). Ref: 2)b): I think it is - as much a resonable expectation held by all people and sites and backbones on the net that their network(s) can be protected against misuse (proactively), or that any incidents can be resolved quickly and easily, - as it might be reasonable to request anonymity on the net for end sites. I guess we have similar situations in real life, where there is a conflict of expectations and interests between the individual and the community, like . license plates on vehicles to follow up on violation of rules, . a public registry for real estate ownership to settle ownership claims . or to track down the owners in case the patch of land becomes a source of danger or nuisance for neighbours, . registries for companies, their legal representatives and financial backing (which many would certainly prefer not to disclose :-) . registries for the address of citizens in case they do not live up to contractual obligations So, would we be better off by following up on our "IRT-pointer" proposal and then offer the alternative to supply - *either* the end users's contact information - *or* the contact information for an IRT that has accepted responsibility for the operational use of that resource? In most cases, and for small end-user sites, this would be ISP staff, anyway. Having a certain resource listed in the registry without any contact information (neither individual contact nor IRT/ISP contact) should be seen as a _strong_ hint to exercise strong care in dealing with that site - by applying whatever means available to minimize the impact of misconfiguration and/or attacks originating from those sites. This could indeed limit the usefulness of those resources on a global scale, much like you are allowed to obtain and drive a vehicle without registration and thus without a license plate on private property - but this vehicle gets filtered or blocked on any public road :-) Comments? Wilfried. _________________________________:_____________________________________ Wilfried Woeber : e-mail: Woeber(a)CC.UniVie.ac.at UniVie Computer Center - ACOnet : Tel: +43 1 4277 - 140 33 Universitaetsstrasse 7 : Fax: +43 1 4277 - 9 140 A-1010 Vienna, Austria, Europe : RIPE-DB: WW144, PGP keyID 0xF0ACB369 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

1 0

Re: Privacy, Broadband, & the Database
by Wilfried Woeber, UniVie/ACOnet 30 Nov '00

30 Nov '00

=> I'm writing for your opinions, suggestions, and comments regarding updating => the RIPE database with assignments made for broadband (xDSL) customers. As => you may or may not be aware, ADSL is now being offered in the UK, which => has very strict data protection laws. We offer ADSL services as a NATted = =IP addresses and data related to them is international resource. =I don't think you should apply any law in UK when we talk about =RIPE(.int) resources. But that doesn't mean the people *collecting* customer related info *and transferring that data to somewhere abroad* would be exempt from local laws. Quite the opposite holds true for many (most?) countries... =I assume we are still supposed to maintain single nic-hdl per person. =What's the plan now, if my LIR would need to refer to the same contact =person? It doesn't make sense to direct any calls to your LIR, if the =problem lies at customer network at our LIR's addresses. It depends on the level of "responsibility" and functionality granted to and exercised by that end site. As I've said in a private mail already, we should ask the question about the usefulness of "assigning" (in the good old sense) very small amounts of addresses to sites which are tied in to the services of their provider anyway. I guess most of the ADSL, dial-up, cable-TV connection assignments sh/could be reviewed from that point of view. => phone: +44-845-272-0444 => remarks: Phone is Demon Internet Helpdesk = =The same rule should apply to all assignments. If a (home) network =contact person doesn't want to be 'listed', s/he can try to live =without real network and buy NAT from a provider. That's a very valid point. =If people think RIPE db contains too much sensitive information, =I would rather see the query results restricted somehow. For example =operators (LIRs) could have full query access, and the others =could merely receive pointers to the LIR for further information. This idea happens to pop always now and then, again and again. the last time I heard a proposal along those lines was at the Brisbane APNIC meeting. I then asked the following question: "So you are going to collect information, but you are going to hide it. Fair enough. Have you thought about criteria and/or mechanisms to disclose that information, either to entites with a "valid" interest and/or to law enforcement?" I did not receive a useful answer. Just as a virtual poll - would anyone in the RIPE region care to propose criteria and/or mechanism? Wilfried. _________________________________:_____________________________________ Wilfried Woeber : e-mail: Woeber(a)CC.UniVie.ac.at UniVie Computer Center - ACOnet : Tel: +43 1 4277 - 140 33 Universitaetsstrasse 7 : Fax: +43 1 4277 - 9 140 A-1010 Vienna, Austria, Europe : RIPE-DB: WW144, PGP keyID 0xF0ACB369 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Problems with windows? re-boot! Problems with unix? be root!

4 3

Re: Privacy, Broadband, & the Database
by Wilfried Woeber, UniVie/ACOnet 30 Nov '00

30 Nov '00

>The name and address info in the RIPE database needs to contain the >person (or organisation) responsible for a certain amount of IP space, >in my opinion. I agreee, and the important word here is "responsible". >So, if an ISP wants to be responsible for the IP space it hands out to >customers, then the ISP should be free to fill in their own name and >address. Doing that means that an assignment (*to the customer*, in the traditional sense) has not been performed. >This gives ISPs the freedom to fill the RIPE db with names >and addresses of customers (that saves them the hassle of dealing with >for example abuse coming from that site), or putting their own name >in it (meaning they have to respond actively to for example abuse reports, >which is the ISP's job anyway). This is very similar to my reasoning regarding the usefulness of performing an assignment for a (small) site which does not fully (independently and in a responsible way) manage their address space. Wilfried.

1 0

Privacy, Broadband, & the Database
by Sam Bradford 29 Nov '00

29 Nov '00

Hello all. (Apologies if you got a 'subscribe' message from me earlier. I wasn't thinking about Majordomo for some reason.) I'm writing for your opinions, suggestions, and comments regarding updating the RIPE database with assignments made for broadband (xDSL) customers. As you may or may not be aware, ADSL is now being offered in the UK, which has very strict data protection laws. We offer ADSL services as a NATted service (with one large entry in the RIPE database like for dial-up customers) and a non-NAT service that includes an IP assignment. Our concern is that a large percentage of our non-NAT ADSL customers are either home users or home businesses. Updating the RIPE database with an IP assignment requires an end-user Admin contact, and the person object requires a telephone number. Worst case scenario is that we update the database for a home business or a home user with the customer's telephone number, his teenage daughter enters a chat room, and you can imagine the rest... After brief consultation with the RIPE NCC Hostmasters, we have begun updating the database with our role object as Tech Contact and the customer as Admin contact. For the phone number, we are using our own, noting as much in a 'remarks' field. For example: inetnum: 62.49.8.64 - 62.49.8.79 netname: PYAR-ADSL descr: DEMON ADSL CUSTOMER country: GB admin-c: PY84-RIPE tech-c: DHG5-RIPE status: ASSIGNED PA notify: hostmaster(a)demon.net mnt-by: AS2529-MNT changed: hostmaster(a)demon.net 20001115 source: RIPE <snip DHG5-RIPE> person: Phil Yarranton address: GB phone: +44-845-272-0444 nic-hdl: PY84-RIPE remarks: Phone is Demon Internet Helpdesk notify: hostmaster(a)demon.net mnt-by: AS2529-MNT changed: hostmaster(a)demon.net 20001115 source: RIPE However, I'm not sure that is the best option, and could lead to inconsistency. At what point do we determine (and should we determine?) if an obvoius business customer is home-based or office-based? Should we have one set of rules for the database for leased line & ISDN customers and another for ADSL? Or one set for business customers (regardless of home-based or office-based) and another for home users? Looking forward to your comments. :-) Sam Bradford ----------------------------------------------------------------- sam bradford, senior hostmaster sam.bradford(a)demon.net Demon Internet / Thus plc . hostmaster(a)demon.net Tel: +44-845-272-0666 . . http://www.demon.net/ Fax: +44-20-8371-1285 t h u s http://www.thus.net/

3 3

About the Hermes Project
by Andrea Carmignani 06 Nov '00

06 Nov '00

Hi my name is Andrea Carmignani, I belong to the network research group of the university of Roma3. We developed a tool, known as Hermes, able to explore the Internet topology. Its based mainly on the information coming from the databases belonging to some Internet Resource Record like RIPE, RADB, VERIO, CABLE & WIRE, ARIN and BELL in addiction to the BGP routing information coming from the Route Projects of the Oregon University. Hermes is able to dynamically explore the Internet Topology, providing a Map representing the As and their interconnections. For any AS present on the map, it can also provide several information like: AS name, Maintainers, Routing Policies, list of the known border Router, list of the routes originated by the AS and the As macro which the AS belongs. We will be grateful if you could give us some feedback on the real utility of this tool and hopefully some suggestion able to improving it. Please visit the URL: http://www.dia.uniroma3.it/~hermes here you will find a more detailed paper, some screenshot galleries, a screencam of the tool working and you will able to use Hermes through the Applet provided. I think this is the right mailing list where to post this news if not I apologize otherwise Thanks in Advance Andrea

2 1

reimp.ripe.net (rpsl.ripe.net) server is temporarily unavailable
by Andrei Robachevsky 27 Oct '00

27 Oct '00

Dear Colleagues, Yesterday a machine that was used for the new prototype DB server crashed and could not be restarted due to hardware problems. We are working on this and will bring the prototype server up as soon as possible. Sorry for any inconvenience. Regards, Andrei Robachevsky DB Group Manager RIPE NCC

1 0

Re: mark
by Randy Bush 29 Sep '00

29 Sep '00

> -----BEGIN PGP SIGNED MESSAGE----- > > person: Mark Bold > address: Verio/UK sigh. apologies. i think i made this same address book stoopid once before. sorreeeee. randy

1 0

RIPE NCC to shut down rpslii.ripe.net on November 6
by RIPE Database Administration 22 Sep '00

22 Sep '00

Dear All, [This does *not* affect the RIPE whois database at whois.ripe.net] As you know, the RIPE NCC has provided a test RPSL database for a few years. This database software (available for queries at rpslii.ripe.net) was written in perl and was set up so that the community can experiment with RPSL. It was around since the time of RPSL drafts. As the reimplementation project is approaching the end, we have set up two instances of the new software at reimp.ripe.net. One of them is a mirror and the other a test database. The purpose of the new test database is exactly the same as of the old perl version. Therefore we decided to shut down the old RPSL perl server at rpslii.ripe.net. This will happen on November 6, 2000. The data accumulated in that database will be lost. You can use the new RIP Test Database already. Updates to this database should be sent to <auto-rip(a)ripe.net> and you can query this database at reimp.ripe.net port 4343. We would be very pleased if you could test the new implementation and let us know your opinions. Please send your comments and bug reports to the <db-beta(a)ripe.net> mailing list. To subscribe to this list, send a message to <majordomo(a)ripe.net> with the line: subscribe db-beta in the body of your letter. This is a moderated mailing list so some delays in subscription are possible. If you have any questions, please don't hesitate to contact <ripe-dbm(a)ripe.net>. Regards, Marek Bukowy ____________________________ RIPE Database Administration.

2 1

Re: auto-dbm quirks
by Steven Bakker 21 Sep '00

21 Sep '00

>>>>> On Thu, 21 Sep 2000, "MR" == Matthew Robinson wrote: MR> I have the "joy" of sitting behind an Exchange server and I can from MR> the 'Format' menu select plain text only which fixes this problem. I MR> suspect that there is more config on the server itself but as an MR> example this message should be plain text. It is, but the Exchange server itself apparently forces the multipart/ alternative, though only for clients that send their mail through the proprietary Exchange protocol. Straight SMTP to the server seems to work fine. In corporate-honcho type environments this may be dandy; keyboard monkeys are less than happy with it. Cheers, Steven (wishing nmh could do "virtual folders" on top of IMAP).

2 1