[Gnso-ppsai-pdp-wg] For review - updated templates Cat B, questions 1 and 2
Stephanie Perrin
stephanie.perrin at mail.utoronto.ca
Mon Mar 3 18:40:55 UTC 2014
THanks for this response. Re your last line, In all risk assessment, the first thing we must do is accept that we live in a world of risk. Getting to Zero risk means being dead. Re your second last line, we do have to quantify the magnitude of risk increased by delay, or new risks created by delay. Otherwise, delay could be still ok.
So the operative questions are, how big is the risk? HOw likely is it? What is the impact? HOw is it sensitive to time ( ie does it compound)? can I tolerate it or do I have to transfer it? Are there reasonable mitigations that I can afford? Can I get somebody else to mitigate my risk, preferably at no cost to myself?
What I see is going on here, from my position of very limited knowledge of the day to day life of registrars, but a slightly larger experience of risk management (Michele will back me up here when I say I am not actually this modest, but for the purposes of this list I am trying to be well behaved, albeit overly talkative)....IS that we are asking registrars and proxy/privacy service providers to mitigate a few risks for other stakeholders. I don't find that fair, unless it can be proven that it is the most cost effective way to mitigate a risk that all stakeholders recognize exists at a high enough level to warrant significant cost and effort. Since we in NCSG purport to speak for civil society and the common person, that means there has to be a risk to them, as far as I am concerned, before new costs and risks start coming their way. I dont think the case has been made that the risk is high enough, or cannot be better mitigated other ways, and I think the registrars and p/p providers have made an eloquent case that loading the verification and the policing function on them is a workload out of scale with the purported benefits. I also guess it will drive new risks, worse than the old ones, as malfeasant actors adjust to new constraints.
Put more concisely: the risk is that people do bad things when they get a domain name. That risk cannot be mitigated by demanding verification of ID, it drives ID theft. It cannot be mitigated by reveal requests, response would likely be more spurious reveal requests, or revelation of bad data, or good data that still wont get you a warrant. Worst case scenario risks (my personal fav is trafficking of women from websites) are much better treated through standard traffic analysis techniques, or tackling the ISPs. (in the case of trafficking women, forget the domain registration data and follow the Johns).
Doubtless my views are still very uninformed but I have not seen evidence to change my mind yet....we need, I think, a comprehensive risk assessment where we list and evaluate all the risks. That would help a lot.
cheers SP
On 2014-03-03, at 10:20 AM, Williams, Todd wrote:
> Thanks Stephanie. To first respond to Michele’s point in the attached: as I mentioned below, I don’t think the argument depends on there being anything “inherently suspicious” or “nefarious” about P/P services or their users (any more so than tractor-trailers or their drivers are “inherently suspicious”), simply a recognition of the risks posed.
>
> Which then gets to your bigger point Stephanie that the “fundamental disagreement” between the “2013 RAA” camp and the “more” camp (as referenced below) is as to this question: whether there is in fact more “risk posed” by P/P services. But there are different ways to assess the “risk posed”, right? One is to ask: “Is the risk higher that users of P/P services will engage in abusive behavior?” I think that is what you are referring to when you say that we have a fundamental disagreement with rival studies. But I also think that we can put that disagreement to the side for now, because I don’t think we need to get that far when assessing the “risk posed.”
>
> Rather, I think there is a second question that we should also ask when assessing the “risk posed” by P/P services, which is this: “Once a user of a P/P service does engage in abusive behavior (however frequently or infrequently that happens), will the fact that they are using a P/P service increase the risk that such abuse will not be successfully addressed/corrected/stopped, etc.?” The answer to this second question has to be yes, for the reason that Steve mentioned in our last call: the P/P service introduces at least some element of delay to the enforcement process. I suppose that we can debate the magnitude of the risk posed by that delay, and I suppose that it will also be related to the parameters that we put in place for relay/reveal etc. procedures down the road. But the magnitude is not zero.
>
> From: Stephanie Perrin [mailto:stephanie.perrin at mail.utoronto.ca]
> Sent: Saturday, March 01, 2014 4:29 PM
> To: Williams, Todd
> Cc: James M. Bladel; Tim Ruiz; PPSAI
> Subject: Re: [Gnso-ppsai-pdp-wg] For review - updated templates Cat B, questions 1 and 2
>
> I don’t think we agree that there is more risk with P/P services, which is kind of a fundamental disagreement. We even have rival studies backing us up.
> As to the analogies, they are useful but limited, as they can capture our imaginations in ways which are not in accord with the facts. My analogy for a p/p registration would not be a tractor trailer licence, it would be a small car with tinted windows.
> Stephanie
> On Mar 1, 2014, at 3:52 PM, Williams, Todd <Todd.Williams at turner.com> wrote:
>
>
> I agree with James that using analogies/illustrations can be helpful to frame issues in these discussions, and I think his airport illustration does a nice job of that. But I would offer a different one:
>
> Couldn't we say that registering a "standard" domain (without a P/P service) is to using a P/P service as driving a car is to driving a tractor-trailer? Everybody recognizes that there is a baseline floor of verification that should be required before you can get a driver's license. But everybody also recognizes that because the potential risk from "bad" tractor-trailer drivers is greater than that of "bad" car drivers (I'm leaving "bad" undefined here intentionally, because it doesn't matter whether it's abusive/malicious/incompetent, etc.), some EXTRA level of verification is needed before you can get a license to drive a tractor-trailer. In other words, nobody would argue that the same test should be used to get a standard driver's license as to get a license to drive tractor-trailers, because the latter by definition carries more risk. So too with "standard" domain registrations vs. P/P registrations.
>
> To James's last point: note that the analogy doesn't depend on there being anything "inherently suspicious" about tractor-trailer drivers (in the normative sense); simply a recognition that what they are doing poses higher risks for the roadways. Note too that it wouldn't be much of an argument to say that because the extra verification required of tractor-trailer drivers wouldn't always catch "bad" drivers ahead-of-time, we should instead simply rely on the standard driver's license test.
>
> Of course, this doesn't address just how far beyond the 2013 RAA floor our verification/re-verification requirements should go, or what additional measures would or wouldn't be effective. This is just to say that if the choice we're wrestling with now is between the 2013 RAA vs. "more" (however "more" is defined), I don't understand the argument on the 2013 RAA side.
>
> Todd
> From: gnso-ppsai-pdp-wg-bounces at icann.org <gnso-ppsai-pdp-wg-bounces at icann.org> on behalf of James M. Bladel <jbladel at godaddy.com>
> Sent: Saturday, March 1, 2014 6:30:35 AM
> To: Tim Ruiz
> Cc: PPSAI
> Subject: Re: [Gnso-ppsai-pdp-wg] For review - updated templates Cat B, questions 1 and 2
>
> So my last sentence doesn't make much sense, unless I substitute "it" with the idea of having differing verification/validation standards for Registrars and PP services.
>
> Thanks--
>
> J.
>
> Sent from my iPad
>
> On Mar 1, 2014, at 9:35, "James M. Bladel" <jbladel at godaddy.com> wrote:
>
> In just one sentence, Tim has captured the essence of the problem. Which isn’t about determining the “right” answer, but attempting to strike a balance between security vs. ease of use and barriers to legitimate access.
>
> Here’s something we can all relate to: Back in 2001, some jerk tried to blow up a plane with his shoe. As a result, now 650 million passengers have to remove their shoes before boarding an airplane in the US. Is this reasonable? As a society, we have determined that it is, but would we feel the same if it was something more, like a full body search? Probably not. Our privacy service routinely investigates and suspends perhaps a thousand domain names in a year. Sounds like a lot, and it certainly keeps the Abuse team busy, but ultimately represents a tiny fraction of the tens of millions of domain names under management.
>
> In any event, I see no compelling reason why the verification requirements for a PP service should be any different than those of a registrar. First, as I believe some in this thread have pointed out, it would allow the service to leverage code, personnel, & processes developed by the affiliated registrar. And it creates a perception that there is something inherently suspicious about subscribing to a PP service, or wanting the equivalent of an unlisted phone number, which simply isn’t the case.
>
> Thanks—
>
> J.
>
>
>
>
> From: Tim Ruiz <tim at godaddy.com>
> Date: Friday, February 28, 2014 at 21:39
> To: John Horton <john.horton at legitscript.com>
> Cc: PPSAI <gnso-ppsai-pdp-wg at icann.org>
> Subject: Re: [Gnso-ppsai-pdp-wg] For review - updated templates Cat B, questions 1 and 2
>
> So to make your invsetigation of 150 cases easier (which I question in any event) millions of users are needlessly hassled. Makes perfect sense in today's world I guess.
>
> Tim
>
>
>
> On Feb 28, 2014, at 5:32 PM, "John Horton" <john.horton at legitscript.com> wrote:
>
> Hi all,
>
> Verification and re-verification of registration data would be enormously useful and important in identifying, mapping and deterring malfeasance. Just to share our background to put our comments in context, we've assisted in over 150 drug or supplement investigations by conducting cybercrime research, and each project typically involved research into dozens, hundreds or even thousands of Whois records plus corresponding IP/NS/MX etc. information. There are numerous instances in which either 1) the accurate Whois data (including, accurate data behind a Whois privacy/proxy service) "broke open" the case, or 2) submitting a WDRPS complaint in instances where we could show that the Whois data was inaccurate resulted in modified Whois information that then "broke open" the case, either by virtue of the modified Whois information itself, or from derivative information (e.g., additional reverse queries on Whois, name server, IP address or other records). Keep in mind too that sometimes showing that the Whois record is falsified results in the suspension of the domain name, which also has the effect of stopping the harmful use of that particular domain name. Verification would result in some instances of inaccurate registration data becoming accurate, or alternatively, of discontinuing registration services.
>
> In the interests of brevity, I hope that summary is enough explanation, but if anyone still doesn't understand how (or agree that) verified registration data -- or by extension, verification and some sort of periodic re-verification -- is useful, I'm happy to provide a couple of real life examples of investigations we've worked on where either the a) accuracy of the Whois record or b) response to the inaccuracy finding was extremely useful, although I'll modify the domain names. Again, I'm happy to provide real-life examples, with redacted information. It's not just occasionally or mildly useful. It's enormously important.
>
> One additional point: keep in mind that when researching criminal networks, there are typically multiple (hundreds or even thousands) of domain names at play, and even if -- as Tim pointed out -- the verified email and phone number have nothing to do with the person's real identity, good cybercrime research across the thousands of Whois records can often result in derivative information pointing to the real identity of the criminal entities.
>
> As to the point that the domain name isn't harmful but the content may be, I suspect that there are minds that won't be changed in this group on both sides of that argument. :) But, I'd point out that that train has left the station, so to speak: Section 3.18 of the 2013 RAA clearly contemplates harmful use of a domain name.
>
> Thanks,
>
> John Horton
> President, LegitScript
> <image001.jpg>
>
> Follow LegitScript: LinkedIn | Facebook | Twitter | YouTube | Blog | Google+
>
>
> On Fri, Feb 28, 2014 at 12:36 PM, Tim Ruiz <tim at godaddy.com> wrote:
> It doesn't. I can use perfectly good information, including a verifiable phone number and email address, that has nothing to do with who I really am. As we have tried to argue before, unsuccessfully, is that all verification does is push the "miscreants" to be better at obfiscating who they are (and it just isn't that hard). As you said, it only results in making it difficult for everyone for the acts of a few.
>
> Tim
>
>
> On Feb 28, 2014, at 2:07 PM, "Stephanie Perrin" <stephanie.perrin at mail.utoronto.ca> wrote:
>
> My apologies I totally mis-read that. So how does verification catch that then?
> On 2014-02-28, at 1:52 PM, John Horton wrote:
>
>
> Well, because absent an accurate Whois record, it can be difficult to know who to hold accountable.
>
> Stephanie, to clarify: I was saying that 95% of Whois data in a certain sub-category of criminal or miscreant behavior (spam, malware, phishing) is inaccurate (not "accurate").
>
> John Horton
> President, LegitScript
> <image001.jpg>
>
> Follow LegitScript: LinkedIn | Facebook | Twitter | YouTube | Blog | Google+
>
>
> On Fri, Feb 28, 2014 at 9:44 AM, Stephanie Perrin <stephanie.perrin at mail.utoronto.ca> wrote:
> I agree, it is all about risk...but what risk are we really talking about? I dont understand why a P/P provider should be forced to take on more risk than other registrars. Further,why should the registrar be accountable for verified data, once the original data verification is done. If John is correct and in 95% of cases the data from the P/P service provider was proven accurate, then how does any amount of data verification solve the problem? The accountability for miscreant behaviour of all kinds rests with the domain name user. IF the data is inaccurate, ramp up the penalties if it can be shown that the data was rendered inaccurate for the purposes of fraudulent activity.
> At the risk of sounding overly philosophical, It seems to me that the Internet ecosystem is somehow being held to account for the actions of individuals. It is the individuals that should be held to account. Not the domain name, or the company that issued it. Particularly, I think that if products sold are tainted, then there is plenty of other consumer protection law that applies...why are we trying to solve that problem?
> Cheers Stephanie perrin
>
> On 2014-02-28, at 12:29 PM, Carlton Samuels wrote:
>
>
> ..which seems to me all about risk management on part of the provider. Its the results that matter.
>
> So, for all the possible permutations, in line with those enumerated by Volker, might it not be more useful to refer 'verified credentials' as a requirement on the provider, allow them to accept the business risk and leave it to them to decide how to do it.......and, inherently, the risks acceptable to them for provisioning the service?
>
> -Carlton
>
>
> ==============================
> Carlton A Samuels
> Mobile: 876-818-1799
> Strategy, Planning, Governance, Assessment & Turnaround
> =============================
>
>
> On Fri, Feb 28, 2014 at 6:29 AM, Volker Greimann <vgreimann at key-systems.net> wrote:
> Hi John,
>
> I am having a bit of a hard time understanding your point here.
>
> You are describing three different cases here, two of which will not benefit from verification in the least bit and one might, but only in some cases:
>
> a) The data is accurate, but stolen: Here verification would not uncover any issues with the data as it is essentially correct and will most likely be identified as accurate.
> b) The data is false: Here, depending on the methods used, the inaccuracy may be uncovered and would lead to an automated request to provide updated data or deactivation after a set time. Remember, in order to keep providing services in a sensible manner, this needs to be automated in some form, i.e. no individual record would likely see any manual review.
> c) The data is already accurate: If the data is already correct, what purpose does verification fulfill? The data cannot become more accurate. Verification in this case seems like an exercise in self-gratification.
>
> That said, even if there is a benefit to be derived from verification, such benefits are achieved once verification concludes. Re-verification of already verified data fulfills no purpose whatsoever. So if a set of data has already been verified by the registrar, there is no need for the p/p provider to again verify the same data. Only if no verification is or can be performed on the registrar level does verification by providers come into play.
>
> Volker
>
> Am 28.02.2014 00:32, schrieb John Horton:
> Thanks, Marika. I also wanted to provide a comment pertaining to Question 2 in the attachments (relating to periodic checks).
>
> In a few of the recent discussions, there's been some reference to criminals always or nearly always being untruthful in their Whois records (even if privacy-protected), leading to the conclusion that there is little purpose in having a registrar or any third party have to verify or re-verify the information (especially if it is difficult to prove that the data is falsified). I wanted to share our experience and observations on that point, in the hope that it's relevant to future discussion regarding Question 2.
>
> Our consistent observation has been that when it comes to a particular sub-category of criminal activity, spam, phishing, malware, and so forth, it's probably safe to say that that statement is true -- the registrant's Whois information is nearly always inaccurate. Even in cases, such as some where we've worked with law enforcement, when the Whois record for a domain name involved in spam, phishing or malware is privacy-protected and is subsequently unmasked, the Whois record is still not accurate behind the privacy curtain. There are probably exceptions, but that's what we've seen well over 95% of the time. On occasion, it's a real address and phone number, just not one genuinely connected to the registrant.
>
> But there are other types of criminal activity where the Whois record is not so regularly obfuscated. For example, we investigate a lot of websites selling tainted dietary supplements that end up containing some toxin or adulterant that harms people. In those cases, we've overwhelmingly seen that even if the Whois record is privacy-protected, the trend is that the underlying Whois record is accurate. The same has been true for illegal or counterfeit medical device websites that we've researched. On illegal Internet pharmacies not engaged in spam, it's probably 50-50. (It might be a shell corporation, but that's still valuable information.)
>
> One important point to consider is that the Whois registration can be relevant information from a banking perspective for commercial entities. That is, some banks are going to look at an online merchant's domain name registration record and if it's either inaccurate or protected, they may require disclosure, or ask about any discrepancy, which can be an incentive for criminals selling products online who nevertheless want to get paid via credit card to have an accurate Whois. Hackers, malware providers and spammers will find a way around that, but they don't necessarily constitute "most" criminal activity.
>
> The point here is, I think verification can still be a useful and necessary tool in either scenario, even if it doesn't uncover useful information a portion of the time. I realize that only pertains to a portion of the issues related to Question 2, but I hope that our observations on that are relevant.
>
> Thanks,
>
> John Horton
> President, LegitScript
> <image001.jpg>
>
> Follow LegitScript: LinkedIn | Facebook | Twitter | YouTube | Blog | Google+
>
>
> On Wed, Feb 26, 2014 at 2:39 AM, Marika Konings <marika.konings at icann.org> wrote:
> Dear All,
>
> Following our call yesterday, please find attached the updated templates for Category B – questions 1 & 2. Please review these templates to make sure the WG discussions have been accurately reflected and feel free to share any comments / edits you may have with the mailing list. We've created a page on the wiki where we'll post the templates that have been finalised for now (noting that for some of these the WG will need to come back to the template at a later date), see https://community.icann.org/x/ihLRAg.
>
> The WG will continue its deliberations on Category B – Question 2 next week. Some of the questions that came up during the conversation yesterday and which you are encouraged to share your views on (and/or add additional questions that need to be considered in this context) are:
> What would be the arguments for not using the same standards / requirements for validation and verification as per the 2013 RAA?
> Should there be a requirement for re-verification, and if so, what instances would trigger such re-verification?
> In case of affliction between the P/P service and the registrar, if the registration information has already been verified by the registrar, should this exempt the P/P provider from doing so?
> Should the same requirements apply to privacy and proxy services or is there a reason to distinguish between the two?
> Best regards,
>
> Marika
>
> _______________________________________________
> Gnso-ppsai-pdp-wg mailing list
> Gnso-ppsai-pdp-wg at icann.org
> https://mm.icann.org/mailman/listinfo/gnso-ppsai-pdp-wg
>
>
>
> _______________________________________________
> Gnso-ppsai-pdp-wg mailing list
> Gnso-ppsai-pdp-wg at icann.orghttps://mm.icann.org/mailman/listinfo/gnso-ppsai-pdp-wg
>
> --
> Bei weiteren Fragen stehen wir Ihnen gerne zur Verfügung.
>
> Mit freundlichen Grüßen,
>
> Volker A. Greimann
> - Rechtsabteilung -
>
> Key-Systems GmbH
> Im Oberen Werk 1
> 66386 St. Ingbert
> Tel.: +49 (0) 6894 - 9396 901
> Fax.: +49 (0) 6894 - 9396 851
> Email: vgreimann at key-systems.net
>
> Web: www.key-systems.net / www.RRPproxy.netwww.domaindiscount24.com / www.BrandShelter.com
>
> Folgen Sie uns bei Twitter oder werden Sie unser Fan bei Facebook:
> www.facebook.com/KeySystemswww.twitter.com/key_systems
>
> Geschäftsführer: Alexander Siffrin
> Handelsregister Nr.: HR B 18835 - Saarbruecken
> Umsatzsteuer ID.: DE211006534
>
> Member of the KEYDRIVE GROUP
> www.keydrive.lu
>
> Der Inhalt dieser Nachricht ist vertraulich und nur für den angegebenen Empfänger bestimmt. Jede Form der Kenntnisgabe, Veröffentlichung oder Weitergabe an Dritte durch den Empfänger ist unzulässig. Sollte diese Nachricht nicht für Sie bestimmt sein, so bitten wir Sie, sich mit uns per E-Mail oder telefonisch in Verbindung zu setzen.
>
> --------------------------------------------
>
> Should you have any further questions, please do not hesitate to contact us.
>
> Best regards,
>
> Volker A. Greimann
> - legal department -
>
> Key-Systems GmbH
> Im Oberen Werk 1
> 66386 St. Ingbert
> Tel.: +49 (0) 6894 - 9396 901
> Fax.: +49 (0) 6894 - 9396 851
> Email: vgreimann at key-systems.net
>
> Web: www.key-systems.net / www.RRPproxy.netwww.domaindiscount24.com / www.BrandShelter.com
>
> Follow us on Twitter or join our fan community on Facebook and stay updated:
> www.facebook.com/KeySystemswww.twitter.com/key_systems
>
> CEO: Alexander Siffrin
> Registration No.: HR B 18835 - Saarbruecken
> V.A.T. ID.: DE211006534
>
> Member of the KEYDRIVE GROUP
> www.keydrive.lu
>
> This e-mail and its attachments is intended only for the person to whom it is addressed. Furthermore it is not permitted to publish any content of this email. You must not use, disclose, copy, print or rely on this e-mail. If an addressing or transmission error has misdirected this e-mail, kindly notify the author by replying to this e-mail or contacting us by telephone.
>
>
>
>
> _______________________________________________
> Gnso-ppsai-pdp-wg mailing list
> Gnso-ppsai-pdp-wg at icann.org
> https://mm.icann.org/mailman/listinfo/gnso-ppsai-pdp-wg
>
> _______________________________________________
> Gnso-ppsai-pdp-wg mailing list
> Gnso-ppsai-pdp-wg at icann.org
> https://mm.icann.org/mailman/listinfo/gnso-ppsai-pdp-wg
>
>
> _______________________________________________
> Gnso-ppsai-pdp-wg mailing list
> Gnso-ppsai-pdp-wg at icann.org
> https://mm.icann.org/mailman/listinfo/gnso-ppsai-pdp-wg
>
>
> _______________________________________________
> Gnso-ppsai-pdp-wg mailing list
> Gnso-ppsai-pdp-wg at icann.org
> https://mm.icann.org/mailman/listinfo/gnso-ppsai-pdp-wg
>
> _______________________________________________
> Gnso-ppsai-pdp-wg mailing list
> Gnso-ppsai-pdp-wg at icann.org
> https://mm.icann.org/mailman/listinfo/gnso-ppsai-pdp-wg
> _______________________________________________
> Gnso-ppsai-pdp-wg mailing list
> Gnso-ppsai-pdp-wg at icann.org
> https://mm.icann.org/mailman/listinfo/gnso-ppsai-pdp-wg
>
> <RE [Gnso-ppsai-pdp-wg] For review - updated templates Cat B, questions 1 and 2.eml>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mm.icann.org/pipermail/gnso-ppsai-pdp-wg/attachments/20140303/9a29e55a/attachment-0001.html>
More information about the Gnso-ppsai-pdp-wg
mailing list