Re: 91.198.71.0 - 91.198.71.255
(Copied to APWG-Chairs and DP-TF in view of the policy implications) On 02/11/2007 16:36:35, Alex Le Heux <alexlh@ripe.net> wrote:
As it turns out, we had already noticed the change in registration data a few weeks ago and are currently in contact with the LIR. We have established procedures that are designed to track changes like this, and this case is already under investigation.
There appear to be seven other ranges and ASNs, similarly acquired. 193.33.128.0/23 AS42672 194.110.69.0/24 AS42811 91.193.40.0/22 AS42662 91.193.56.0/22 AS42672 91.194.140.0/23 AS43188 91.195.116.0/23 AS43702 91.196.232.0/22 AS43259 91.198.71.0/24 AS43603
We have audited the request and our policies and procedures were correctly applied when this range was assigned.
That's obviously very worrying - because it implies that any similar new application now would also be granted, and the claims that RIPE still has three years to go before IP4 exhaustion would then be seen as having been highly optimistic! I note your earlier comment about the data having been recently changed: so are you saying that the original application was valid just because it had an address in the RIPE region - even though that address may have been meaningless? If so, are you able to remind us what that address originally was (as at the time it would have been "public" data)? As you may know, the entity believed to be using these IP assignments is the notorious "Russian Business Network" about which much has recently been written: and http://en.wikipedia.org/wiki/Russian_Business_Network provides a convenient index to the more significant of those articles. The RIPE NCC cannot, of course, be concerned with any criminal activity that uses IP addresses it assigns, but I believe the RIPE NCC does need to be concerned about the obtaining by dishonest means of resources that are owned by the community and entrusted to the stewardship of RIPE NCC. During RIPE 55 I mentioned the earlier cases of shell companies set up apparently by a "Boris Mizhen" to apply for IP resources for spamming purposes (ie to avoid filters) and I understand that the LIR handling those applications (Merezha) has previously been linked to questionable assignments. For the record, the IP ranges and ASNs involved with that series of incidents, were as follows: 91.193.152.0/22 AS42719 91.193.192.0/22 AS42719 91.193.216.0/22 AS42719 91.193.88.0/22 AS42719 91.200.124.0/22 AS42719 91.200.132.0/22 AS42719 91.200.164.0/22 AS42719 91.200.176.0/22 AS42719 91.200.56.0/22 AS43791 91.200.60.0/22 AS43791 91.200.72.0/22 AS43799 91.200.80.0/22 AS43799 I (and others) mentioned at the Address Policy WG the probability of an imminent IPv4 landgrab - these two recent incidents seem to suggest that it has started. How well can RIPE policies stand up to such deliberate and abusive attacks? In particular, how protected would the community be against an LIR that intentionally submits applications which rely on data the LIR knows is bogus? The LIR is not identified on the WHOIS output - but if the only policing of the application is done by the LIRs (as I'm told that RIPE hostmasters are not allowed to question details supporting an application for resources) perhaps the identity of the LIR should be displayed against the IP range, so that patterns of dishonesty can quickly become visible?
We do sometimes assign resources to organisations for use outside our region, although it rarely happens, and such requests are handled very carefully as they are rather unusual.
I'd welcome some clarification on the policies involved there: is that just to entities ("organisations") located, at least nominally, within the RIPE service region: with operations outside that region - or would entities/organisations within other regions qualify? I ask because the other questionable incident that came to our attention recently was the use of 85.255.112.0/20 in California USA, when it was shown in the RIPE database as unassigned space: and we discovered that an assignment was, mysteriously, made of that space by the RIPE Hostmaster to "Inhoster" on the very day I pointed out the inappropriate use, and that assignment has also recently had its data changed: showing it as now being assigned to a "UkrTeleGroup". But still it is used - solely - in California USA.
It is a normal occurrence that a request is denied by one of the RIRs. If the reason for the denial is related to the location of the network, the end-user is then referred to the correct RIR. This is normally nothing to be suspicious about as there are many legitimate organisations that have operations in multiple RIR regions.
Given the (apparent) disparity in policy implementation between RIPE and the other RIRs - most of whose hostmasters are empowered to (and indeed do) check for and reject abusive applications it seems RIPE may soon be regularly targeted by organisations worldwide who cannot, for whatever reason, get resource assignments from their local RIR.
Could you tell us a little about the circumstances that brought this range to your attention? Any information might help us with our investigation.
We have been tracking the "Russian Business Network" and also "Boris Mizhen" for some time now, and we have been monitoring routing and other changes which tell us their traffic has moved to new routes. It is, however, unlikely that their actual location has changed, as traceroutes to IP addresses in the new ranges are indicative of the use of a "traceroute simulator" rather than of a real network path. Best regards -- Richard D G Cox <cio@spamhaus.org> CIO, The Spamhaus Project http://www.spamhaus.org
Dear Richard, Thank you for providing us with a list of specific resources that might be associated with this event. We will add them to our investigation. You can be sure that if we find fraudulent registrations, we will take appropriate action. Unfortunately neither the RIPE Policies nor our legal framework enables us to pursue action against any other type of abuse. All requests that we receive are treated equally, and where needed, we verify the identity of the requester to the best of our abilities. We also monitor the RIPE Database for unauthorised changes and take action where required. Usually these cases turn out to be legitimate company name changes or changes of ownership. Some of the changes in the RIPE Database that you have noticed are such cases. We do not however monitor the BGP routing table, as our mandate specifically does not extend to policing routing. It has been suggested several times over the past years that we should police routing. However, there has never been consensus in the community that we should actually do this. If you wish to open a discussion on this issue, the Address Policy Working Group is probably the best venue. A possible "land grab" for the remaining IPv4 address space is of course something that we keep a close watch on. So far, there have been no indications that one is ongoing. It is also highly unlikely that such a land grab would take the shape of a relatively small number of small PI assignments. The grand total of PI space that we assign is less than 2% of the total space that we assign and allocate, so such an event would be extremely noticeable. Also, as we publish our assignment and allocation data daily, there are more eyes than just ours watching, so a sudden change in these patterns would be very obvious to many people. Resources that the RIPE NCC allocates and assigns are, in principle, for use in the RIPE region. Our policies do no define precisely what "use in the RIPE region" means. Over the years the RIRs have developed a procedural framework for dealing with cross-region requests: - When the network is physically located in our region, even though connectivity might not go through our region, that network is eligible to receive resources from the RIPE NCC. - When a network is physically located outside our region but the routing originates inside our region it is also eligible. - When the request comes through a RIPE region LIR who is also the connectivity provider, it is also eligible. Note that these are guidelines, as by their nature these requests cover a large variety of situations. When needed, there is communication between the different RIRs about such requests. If you have noticed a disparity between the different RIRs that does not come from differences in the regional policies, I would be very interested to hear about this. I have also discussed your request for historical data with our legal team, and any data not currently available to the public requires a court order for the RIPE NCC to release it. This applies even to data that was published in the past, but is no longer publicly available. Finally, your suggestion to include the identity of the LIR in PI/AS assignment objects is an interesting one. If you wish to pursue it, it should be made to the Database and/or Address Policy Working Groups. Best Regards, Alex Le Heux RIPE NCC On Nov 4, 2007, at 17:50, Richard Cox wrote:
(Copied to APWG-Chairs and DP-TF in view of the policy implications)
On 02/11/2007 16:36:35, Alex Le Heux <alexlh@ripe.net> wrote:
As it turns out, we had already noticed the change in registration data a few weeks ago and are currently in contact with the LIR. We have established procedures that are designed to track changes like this, and this case is already under investigation.
There appear to be seven other ranges and ASNs, similarly acquired.
193.33.128.0/23 AS42672 194.110.69.0/24 AS42811 91.193.40.0/22 AS42662 91.193.56.0/22 AS42672 91.194.140.0/23 AS43188 91.195.116.0/23 AS43702 91.196.232.0/22 AS43259 91.198.71.0/24 AS43603
We have audited the request and our policies and procedures were correctly applied when this range was assigned.
That's obviously very worrying - because it implies that any similar new application now would also be granted, and the claims that RIPE still has three years to go before IP4 exhaustion would then be seen as having been highly optimistic!
I note your earlier comment about the data having been recently changed: so are you saying that the original application was valid just because it had an address in the RIPE region - even though that address may have been meaningless? If so, are you able to remind us what that address originally was (as at the time it would have been "public" data)?
As you may know, the entity believed to be using these IP assignments is the notorious "Russian Business Network" about which much has recently been written: and http://en.wikipedia.org/wiki/ Russian_Business_Network provides a convenient index to the more significant of those articles.
The RIPE NCC cannot, of course, be concerned with any criminal activity that uses IP addresses it assigns, but I believe the RIPE NCC does need to be concerned about the obtaining by dishonest means of resources that are owned by the community and entrusted to the stewardship of RIPE NCC.
During RIPE 55 I mentioned the earlier cases of shell companies set up apparently by a "Boris Mizhen" to apply for IP resources for spamming purposes (ie to avoid filters) and I understand that the LIR handling those applications (Merezha) has previously been linked to questionable assignments. For the record, the IP ranges and ASNs involved with that series of incidents, were as follows:
91.193.152.0/22 AS42719 91.193.192.0/22 AS42719 91.193.216.0/22 AS42719 91.193.88.0/22 AS42719 91.200.124.0/22 AS42719 91.200.132.0/22 AS42719 91.200.164.0/22 AS42719 91.200.176.0/22 AS42719 91.200.56.0/22 AS43791 91.200.60.0/22 AS43791 91.200.72.0/22 AS43799 91.200.80.0/22 AS43799
I (and others) mentioned at the Address Policy WG the probability of an imminent IPv4 landgrab - these two recent incidents seem to suggest that it has started. How well can RIPE policies stand up to such deliberate and abusive attacks? In particular, how protected would the community be against an LIR that intentionally submits applications which rely on data the LIR knows is bogus? The LIR is not identified on the WHOIS output - but if the only policing of the application is done by the LIRs (as I'm told that RIPE hostmasters are not allowed to question details supporting an application for resources) perhaps the identity of the LIR should be displayed against the IP range, so that patterns of dishonesty can quickly become visible?
We do sometimes assign resources to organisations for use outside our region, although it rarely happens, and such requests are handled very carefully as they are rather unusual.
I'd welcome some clarification on the policies involved there: is that just to entities ("organisations") located, at least nominally, within the RIPE service region: with operations outside that region - or would entities/organisations within other regions qualify? I ask because the other questionable incident that came to our attention recently was the use of 85.255.112.0/20 in California USA, when it was shown in the RIPE database as unassigned space: and we discovered that an assignment was, mysteriously, made of that space by the RIPE Hostmaster to "Inhoster" on the very day I pointed out the inappropriate use, and that assignment has also recently had its data changed: showing it as now being assigned to a "UkrTeleGroup". But still it is used - solely - in California USA.
It is a normal occurrence that a request is denied by one of the RIRs. If the reason for the denial is related to the location of the network, the end-user is then referred to the correct RIR. This is normally nothing to be suspicious about as there are many legitimate organisations that have operations in multiple RIR regions.
Given the (apparent) disparity in policy implementation between RIPE and the other RIRs - most of whose hostmasters are empowered to (and indeed do) check for and reject abusive applications it seems RIPE may soon be regularly targeted by organisations worldwide who cannot, for whatever reason, get resource assignments from their local RIR.
Could you tell us a little about the circumstances that brought this range to your attention? Any information might help us with our investigation.
We have been tracking the "Russian Business Network" and also "Boris Mizhen" for some time now, and we have been monitoring routing and other changes which tell us their traffic has moved to new routes. It is, however, unlikely that their actual location has changed, as traceroutes to IP addresses in the new ranges are indicative of the use of a "traceroute simulator" rather than of a real network path.
Best regards
-- Richard D G Cox <cio@spamhaus.org> CIO, The Spamhaus Project http://www.spamhaus.org
Best regards, Alex Le Heux RIPE NCC IP Resource Analyst
Hi, On Wed, Nov 07, 2007 at 04:05:30PM +0100, Alex Le Heux wrote:
Finally, your suggestion to include the identity of the LIR in PI/AS assignment objects is an interesting one. If you wish to pursue it, it should be made to the Database and/or Address Policy Working Groups.
Actually I think that this wouldn't be easy today (given that PI stands for "independent"), but with the new contract framework, we *do* have the instruments to tag PI/AS objects to "who is the currently-responsible LIR for this"? Whether we want (or even are allowed to) to publish this information is a different matter, and I'm not in a position to give a good answer on this. Gert Doering -- APWG chair -- Total number of prefixes smaller than registry allocations: 122119 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279
On 8 Nov 2007, at 17:57, Gert Doering wrote: [...]
Finally, your suggestion to include the identity of the LIR in PI/AS assignment objects is an interesting one. If you wish to pursue it, it should be made to the Database and/or Address Policy Working Groups.
Actually I think that this wouldn't be easy today (given that PI stands for "independent"), but with the new contract framework, we *do* have the instruments to tag PI/AS objects to "who is the currently-responsible LIR for this"?
Why would we care which LIR sent in a request for a PI assignment and why would we want to require "independent" networks to maintain a contractual relationship with an LIR? I can't speak for anyone else but I don't care who sends in the request. I care who is operating the network and how to contact them. Anything that diverts us from maintaining operationally relevant contact information is unhelpful. Identifying LIRs strikes me as a mechanism to label LIRs as either "good" or "bad" and doing this is at best likely to besmirch innocent customers who are unrelated to any abuse. At worst it is likely to cause them problems operating their networks. "Guilty by association" is something we should try to avoid. Regards, Leo Vegoda
Hi, just some more food for thought: On Fri, Nov 09, 2007 at 06:52:29PM +0100, Leo Vegoda wrote:
Actually I think that this wouldn't be easy today (given that PI stands for "independent"), but with the new contract framework, we *do* have the instruments to tag PI/AS objects to "who is the currently-responsible LIR for this"?
Why would we care which LIR sent in a request for a PI assignment and why would we want to require "independent" networks to maintain a contractual relationship with an LIR?
Because the RIR wants to know how to reach the PI holder. Either they have a direct relation - or if not, they need to go through a LIR. (At least that seems to be what the RIPE community and the NCC agrees to be the way forward). Whether or not this relationship should be made publically viewable is a completely different discussion, of course. [..]
Identifying LIRs strikes me as a mechanism to label LIRs as either "good" or "bad" and doing this is at best likely to besmirch innocent customers who are unrelated to any abuse. At worst it is likely to cause them problems operating their networks. "Guilty by association" is something we should try to avoid.
Actually it goes along with the responsibility thing of address distribution. If the LIR doesn't do a proper job in maintaining contact data for the end user, the blame falls back onto the LIR, and I think that this is actually not such a bad thing. Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 110584 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279
Hi Gert, On 20 Nov 2007, at 12:34, Gert Doering wrote: [...]
Actually I think that this wouldn't be easy today (given that PI stands for "independent"), but with the new contract framework, we *do* have the instruments to tag PI/AS objects to "who is the currently-responsible LIR for this"?
Why would we care which LIR sent in a request for a PI assignment and why would we want to require "independent" networks to maintain a contractual relationship with an LIR?
Because the RIR wants to know how to reach the PI holder. Either they have a direct relation - or if not, they need to go through a LIR.
(At least that seems to be what the RIPE community and the NCC agrees to be the way forward).
Whether or not this relationship should be made publically viewable is a completely different discussion, of course.
The RIPE NCC has always maintained records of which LIR sent in a request for a PI assignments. I assume that this would continue. I have doubts about whether publishing this information is appropriate.
Identifying LIRs strikes me as a mechanism to label LIRs as either "good" or "bad" and doing this is at best likely to besmirch innocent customers who are unrelated to any abuse. At worst it is likely to cause them problems operating their networks. "Guilty by association" is something we should try to avoid.
Actually it goes along with the responsibility thing of address distribution. If the LIR doesn't do a proper job in maintaining contact data for the end user, the blame falls back onto the LIR, and I think that this is actually not such a bad thing.
I doubt that most LIRs are in a position to maintain the end users's contact information as well as the end user. I understand that the RIPE NCC encourages PI holders to have their own maintainer on the inetnum for their address space. The current training material mentions this on slide 72: http://www.ripe.net/training/lir/material/lir_training_oct07-1.pdf I expect a very well designed web based tool to allow end users to manage their contact information would have a far more beneficial effect that a contractual requirement for the LIR to do it for the end user. Regards, Leo
Hi, On Tue, Nov 20, 2007 at 11:53:00PM +0100, Leo Vegoda wrote:
I expect a very well designed web based tool to allow end users to manage their contact information would have a far more beneficial effect that a contractual requirement for the LIR to do it for the end user.
All nice and shiny. But what are you going to do if the end user does *not* properly maintain their contact information? Having requirements without a way to apply pressure if the requirements are not met is not overly effective. Gert Doering -- NetMaster -- Total number of prefixes smaller than registry allocations: 110584 SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (89) 32356-444 USt-IdNr.: DE813185279
participants (4)
-
Alex Le Heux -
Gert Doering -
Leo Vegoda -
Richard Cox