January 1997 - routing-wg - mailman.ripe.net

Global routing table size graph
by Erik-Jan Bos 23 Jan '97

23 Jan '97

Routing WG, during the Routing WG meeting I promissed to e-mail you the URL of the graph that shows the global routing table size since 1 January 1994. Measurements are done every hour on the hour, by a script I run on my workstation. This graph is updated automatically, every hour. The URL is: ftp://ftp.surfnet.nl/surfnet/net-management/ip/nets.ps The first real dip in the graph was after the Houston (?) IETF, where during the CIDRD working group meeting almost all ISPs present decided to start using CIDR. __ Erik-Jan.

1 0

Re: hierarchical route objects, part 1
by Schmitz＠RUS.Uni-Stuttgart.DE 16 Jan '97

16 Jan '97

Dear Curtis, you wrote: > > > Here is an example. The registries approve of the top level blocks, > > > that is the aggregates (or blocks that should be announced as > > > aggregates). Whoever the aggregate is allocated to creates the top > > > level route object and can delegate below that. For example, ANS has > > > 207.24/14. There were two route objects needed to handle multihomed > > > (a /22 and a /24) prefixes. The usual way a provider would delegate > > > would be to reference more than one maintainer in the route object > > > (the provider and the customer, and probably the other provider if the > > > route object had a unique origin AS). > > > > > This scheme sounds reasonable but it raises the question of coordination > > among different registries... > > I would not expect to see the same address space allocated by more > than one registry. All registries would have a top level 0/0 route > object which prevents registry of bogons. RIPE would have a hierarchy > under 0/0 and the RA would have a hierarchy tracking InterNIC assigned > addresses. I would not expect the two to overlap, therefore there is > no real coordination problem beyond initial setup or new blocks > allocated to a registry. > This is more difficult than it looks from the beginning. As an example we ourselves register all our objects (inetnum, routes, persons...) in RIPE database (because we are in Europe). However, among others we peer with MCI and MCI insists on having route objects in their registry in order to allow routing. If the route objects are secured by hierarchical authorisation in the RIPE database we would also like to have them secured in the MCI registry. Following the scheme above MCI would have to include inetnums (while it is a pure routing registry) *and* would have to generate top level route objects allowing specific access by us to our route objects (this is a very tiring job especially regarding "the swamp" 192.x). Of course, they could copy from RIPE database... (ask them why they don't...) I guess, this problem may only be resolved by the GRR (which is yet to come) From the ongoing disussion I conclude for myself (at the current state of the discussion) that we can not impose a strict security mechanism at the time being without breaking too many things. Maybe, I am too dumb to see a simple solution which might be just at hand... I will summarise the things said here on the mailing list for the routing wg session next week and try to think some more about it (I guess, Daniel was right from the start when he said that we should begin with a bare notification mechanism...) 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) _____________________________________________________________________________

2 1

RIPE 26 Routing WG Draft Agenda
by Schmitz＠RUS.Uni-Stuttgart.DE 16 Jan '97

16 Jan '97

- Routing Working Group Meeting - Proposed Agenda (1st draft) for RIPE-26, Jan 1997, Amsterdam A. Administrative issues (J.Schmitz) - volunteering of the scribe - agenda bashing - minutes from last meeting - open actions B. New Developments of RATools (D.Kessens) C. Hierarchical authorisation for route objects (J.Schmitz) D. Report on route aggregation by the RIPE NCC (D.Karrenberg, RIPE NCC) E. Report on routing stability (G.Winters, Merit) Y. General input from other WGs Z. AOB _____________________________________________________________________________ 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

Re: hierarchical route objects, part 1
by Schmitz＠RUS.Uni-Stuttgart.DE 16 Jan '97

16 Jan '97

Dear Curtis, dear David, you both wrote: > > > Your black hat example is also flawed. At the top of the heirachy can > > > be 0/0 registered to IANA and withdrawn (not announced). The > > > registries themselves can have top level objects below that. In order > > > to make any changes, you need to have been given authorization from a > > > higher level. You can then assign authorization to lower blocks to > > > other parties. > > > > This works for IP network objects since the registries need to add these > > objects manually anyway. > > > > This is not that obvious for 'route' objects. Are you proposing that the > > registries have to approve (manually) all the route objects that are > > in the route hierarchy directly below their own allocated space ? > > Tie the hierarchy to the inetnums. Continue to flog the InterNIC for > non-participation in the IRR. The maintainer of a database such as > the RADB that uses InterNIC as the number registry may have to take > InterNIC data (ftp) and load it into inetnum records. I wonder about the effort necessary to do this. Moreover, we should not forget about pure routing registries - they would be forced to include inetnums as well (and I wonder whether they are willing to do so) > > An inetnum references one or more maintainer (in mnt-by). To create a > route object, you must satisfy one of the criteria: > > 1. You must be in the maintainer field in a less specific route > object (used to create route objects for more specific prefixes; > hopefully these get aggregated somewhere ). > > 2. You must be in a maintainer for the exact inet-num and for the > AS that you are putting in the route object (used after the > initial top level route object is created). > > Once a top level route object is created, a more specific route object > can be created and the mnt-by field can reference more than one > maintainer. > > Here is an example. The registries approve of the top level blocks, > that is the aggregates (or blocks that should be announced as > aggregates). Whoever the aggregate is allocated to creates the top > level route object and can delegate below that. For example, ANS has > 207.24/14. There were two route objects needed to handle multihomed > (a /22 and a /24) prefixes. The usual way a provider would delegate > would be to reference more than one maintainer in the route object > (the provider and the customer, and probably the other provider if the > route object had a unique origin AS). > This scheme sounds reasonable but it raises the question of coordination among different registries... 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) _____________________________________________________________________________

2 1

Re: again: hierarchical route objects
by Schmitz＠RUS.Uni-Stuttgart.DE 15 Jan '97

15 Jan '97

Dear all, was my last mail too long? I did not yet see any comments... (or is it that disgusting :-( I hope not ;-) 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

hierarchical route objects, part 1
by Schmitz＠RUS.Uni-Stuttgart.DE 11 Jan '97

11 Jan '97

Dear colleagues, regarding hierarchical authorization of route objects in the RIPE database: from what I have heard there is a general feeling that it is needed and the basic scheme to implement it should follow the lines: * The root of the authorization tree is an AS-object (aut-num object). If it contains a "mnt-lower" attribute it controls all route-objects which have this AS as origin. * Then for route-objects the same rules apply as for inetnum-objects with respect to IP subranges: If a route-object contains a "mnt-lower" attri- bute it controls all more specific route-objects immediately below. * The authorization is checked against - more or less specific route-objects, or existence of the route-object itself with same origin (differing origin rejected) - if no route-objects exist: which authorization is specified for the autnum-object referred to by the origin attribute (rejected if this authorisation is not met) - if not even an autnum-object exists no action is taken However: there is still a problem that route-objects are somehow logically linked to allocated address space. The question how to deal with this is still open - I continue on this in a separate mail. Yet, the three rules for route-objects described above are a kind of common denominator(*) and moreover a very reasonable approach (these rules are also independent of the address space allocation relation to route-objects). If there are no further denials I suggest to implement it that way. Regards Joachim (*) Yes, I know: When aiming for the common denominator, be prepared for the occasional division by zero. _____________________________________________________________________________ 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) _____________________________________________________________________________

5 5

Re: hierarchical route objects, part 1
by Steven Bakker 10 Jan '97

10 Jan '97

==> From: Gabor Kiss <kissg(a)sztaki.hu> ==> Thu, 9 Jan 1997 11:28:52 +0100 CA> I disagree that this is a good idea. If I register the following route CA> object (which actually exists): CA> CA> route: 128.0.0.0/1 CA> descr: HALF-DEFAULT-ONE CA> origin: AS1800 CA> advisory: AS690 1:1800 2:1239 CA> mnt-by: MAINT-AS1800 CA> mnt-lower: MAINT-AS1800 CA> CA> nobody else can register any route objects. GK> This can be prevented by this object GK> route: 0.0.0.0/0 GK> descr: for authorisation purposes GK> origin: AS0 GK> mnt-lower: MAINT-INTERNIC GK> etc. GK> GK> Internic (or whatever else) may delegate maintainer rights as well GK> as allocates address ranges. That still doesn't help you if the 128.0.0.0/1 is already registered (which it is); I imagine the proposed auth scheme would look for the most specific enclosing route object, which is 128.0.0.0/1 and not 0.0.0.0/0 (for most routes nowadays anyway). Steven

2 1

again: hierarchical route objects
by Schmitz＠RUS.Uni-Stuttgart.DE 09 Jan '97

09 Jan '97

Dear all, this is quite some discussion (I really did not expect this). Obviously, I could have been doing some more thinking before writing my earlier mail... Now, I try to summarize some of the points. I think Havard made some good comments. He really had me reconsider the effect of a "mnt-lower" attribute in the aut-num object. Actually, "mnt-lower" makes perfect sense if some hierarchically lower object exists (which is valid for IP address ranges). From some logical point of view, aut-num objects may well be understood as being hierarchically higher than route objects. However, to create a route object in the database the current imple- mentation already demands a reference to an autnum object and a main- tainer object: route: [mandatory] [single] descr: [mandatory] [multiple] origin: [mandatory] [single] hole: [optional] [multiple] withdrawn: [optional] [single] comm-list: [optional] [multiple] advisory: [optional] [multiple] remarks: [optional] [multiple] notify: [optional] [multiple] mnt-by: [mandatory] [multiple] changed: [mandatory] [multiple] source: [mandatory] [single] By the mandatory mnt-by attribute some authorisation is already needed (as specified in the maintainer object). Adding a mnt-lower attribute to the *aut-num* object does not change much, and raises the question what objects are actually hierarchically below it - these are not necessarily route objects. I guess, this is one of the reasons why several people feel that a closer relationship to allocated IP address space should be established. As the mnt-lower attribute in an aut-num object seemingly does not lead to additional hierarchical authorisation we arrive at the question what we want to achieve with hierarchical authorization of route objects. For me the rationale behind it are similar to inetnum objects: to exert control. If some origin AS routes an IP range, no other AS shall claim this range or subranges. However, we must ask ourselves how reasonable this approach is because we run into several problems here (Cengiz, Gabor, Steven, and earlier Curtis discussed this): - There must be some starting point for the authorization - otherwise anybody may create route objects of IP address ranges where none do yet exist. This seems to tie the route objects closely to allocated address space (which again leads to problems - as noted before - be- cause several registries are pure routing registries). Moreover, route objects may be created in one registry which are forbidden by authorisation in another. This really calls for "the GRR" (where pure routing registries might be regarded as less authoritative). I fear we do not have a solution for this problem today. - Multihoming will be impossible if authorization is based upon IP address ranges alone because in this case only one route object of one origin is allowed in the registry for each IP range. - Situations may arise where IP subranges must be routed by another AS than specified in an less specific route object. To restrict access to these subranges the "hole" attribute might be applicable. Subranges defined by "hole" could be excluded from the authorization as long as no route object exists which covers this subrange. I can imagine that this is complicated regarding implementation in the database software. - Handing over route objects from one maintainer to another will become difficult - up to now this is solved by allowing existence of several route objects for the same address space but of differing origin. This is a very convenient solution thinking of cases where filters for router configurations are built from route objects in the database. If two identical route objects of different origin exist this may allow a smooth switch over from one AS to another. With an authorisation scheme solely based upon IP address ranges as specified above this would no longer be possible. - Others? While with inetnum-objects the "mnt-lower" attribute leads to strong control there was a suggestion by Daniel that for route objects noti- fication may be sufficient. On the one hand (taking it to the extremes) this might lead to cases where an organisation controls an inetnum object but others control the route objects of the same address range (this may be reasonable but also a real competitive problem). On the other hand is this restriction needed because then at least multi- homing will probably become impossible? Anyway: obviously, checking hierarchical authorization with route objects is quite complex regard- less of the fact whether notification only or restrictive authorization is applied. This whole thing became longer than I first intended. It seems there is more in it than we all thought first ... however, some of the problems we have run into look similar to those of inetnum objects ... may I ask for additional comments? Any further suggestions, problems,... ? 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

Re: hierarchical route objects, part 1
by Gabor Kiss 09 Jan '97

09 Jan '97

Dear Cengiz, > I disagree that this is a good idea. If I register the following route > object (which actually exists): > > route: 128.0.0.0/1 > descr: HALF-DEFAULT-ONE > origin: AS1800 > advisory: AS690 1:1800 2:1239 > mnt-by: MAINT-AS1800 > mnt-lower: MAINT-AS1800 > > nobody else can register any route objects. This can be prevented by this object route: 0.0.0.0/0 descr: for authorisation purposes origin: AS0 mnt-lower: MAINT-INTERNIC etc. Internic (or whatever else) may delegate maintainer rights as well as allocates address ranges. Gabor

2 1

Re: hierarchical route objects, part 1
by Havard.Eidnes＠runit.sintef.no 08 Jan '97

08 Jan '97

Hi, just a few quick comments to your proposal. > regarding hierarchical authorization of route objects in the > RIPE database: from what I have heard there is a general > feeling that it is needed and the basic scheme to implement it > should follow the lines: You say very little about the reasons and motivations for wanting the modified behaviour. I think that is important to state. I've made some guesses below -- correct them as appropriate. > * The root of the authorization tree is an AS-object (aut-num > object). If it contains a "mnt-lower" attribute it controls > all route-objects which have this AS as origin. Aren't the route objects with a given origin already controlled by the authorization specified in the aut-num object? (Sorry, it's been a while since I twiddled "my" route objects, so I'm a bit rusty on this now.) If I'm in error here and this is modified behaviour, what would the motivation be for it? > * Then for route-objects the same rules apply as for > inetnum-objects with respect to IP subranges: If a > route-object contains a "mnt-lower" attribute it controls > all more specific route-objects immediately below. The motivation seems to be to have the option of controlling the use of the given address space covered, e.g. to require intervention by the maintainer/"owner" of the address space should a customer want to leave and take the address space within this address block with him, or to prevent such a move. (?) Other than that I think your suggestions make sense. - Håvard

1 0