[db-wg] RIPE Policy Proposal 2017-02 Validates Database Attributes

Hi all,

Agreed that a yearly email ping-pong is not really a good verification
of abuse contact address. It's trivial to let RIPE NCC mails through
and throw away the rest.

But, one might ask, what's the point of a database if the data is not
reliable? What's the point of all that contact info if we can't trust
it?

Why would we want to store (and serve) data that is not correct?

Data validation is standard industry practice, pretty much everywhere.
You can't even buy a bar of chocolate without validating your email
address. One might ask, why would RIPE NCC not follow even the most
basic industry practices while validating data of this importance?

Cheers,
Agoston


On Thu, Sep 7, 2017 at 6:54 PM, Sascha Luck [ml] via db-wg
<db-wg at ripe.net> wrote:
>
>
> ---------- Forwarded message ----------
> From: "Sascha Luck [ml]" <dbwg at c4inet.net>
> To: db-wg at ripe.net
> Cc: anti-abuse-wg at ripe.net
> Bcc:
> Date: Thu, 7 Sep 2017 17:54:04 +0100
> Subject: Re: [db-wg] RIPE Policy Proposal 2017-02 Validates Database Attributes
> All,
>
> I will discuss this here as I do not accept the Anti-Abuse WG as
> a forum for this proposal. For one thing, this proposal affects
> every ripedb user - in fact, as this entails changes to how the
> NCC provides services, the services-wg would be an even better
> venue. For another, given the "population" and culture of
> "debate" on the AAWG, any "consensus" reached there would be so
> worthless as to be farcical. (If anyone wants amplification on
> this, https://www.ripe.net/ripe/mail/archives/anti-abuse-wg/
> provides ample evidence.
>
> To the meat of the proposal:
>
> Firstly, this proposal conflicts somewhat with NWI-7 in this very
> WG. (Another reason why it should have gone here in the first
> place) With the upcoming possibility of delegating abuse-c: to
> (customers') resource objects, who bears any consequences of
> these customers not replying to this proposed email?
>>
>> Rationale
>> a. Arguments supporting the proposal
>
>
>> Accurate and validated information in the RIPE Database is
>> essential to establish a trusted and transparent environment in
>> which all network operators can operate safely.
>
>
> abuse-c: is a contact object, just as admin-c: or tech-c: and
> (correct me if I'm wrong) audited at the same time as the rest of
> the contact information. What makes it so unique that this
> verification is not enough?
>
>> Accurate and
>> validated information helps Internet troubleshooting at all
>> levels, but it also helps to attribute malicious online
>> activities that undermine this trusted environment.
>
>
> See above.
>
>> The lack of reliable accurate and validated information in the
>> database negatively impacts legitimate uses of the RIPE Database,
>> including:
>
>
> An *email adress* that doesn't reply once a year does NOT equate
> to a "lack of reliable accurate and validated information". I
> find this statement somewhat insulting to the NCC team who do
> make the effort to keep the ripedb data accurate and do audit
> resource holders.
> There is an issue with the reliability of out-of-region and
> legacy resource data but as the NCC has no "enforcement" powers
> over these resource holders, in these cases this proposal
> snatches at thin air.
>>
>> Assuring the security and reliability of the network by
>> identifying points of contact for IP addresses for network
>> operators, ISPs, and certified computer incident response teams;
>
>
> "org:", "admin-c:", "tech-c:", "mnt-by:" and, yes, "abuse-c:" exist.
>
>> Ensuring that IP address holders are accountable, so individuals,
>> consumers and the public are empowered to resolve abusive
>> practices that impact safety and security;
>> Assisting businesses, consumer groups, healthcare organisations
>> and other organisations that are combating fraud (some of which
>> have mandates to electronically save records) to comply with
>> relevant legal and public safety safeguards;
>
>
> The contact object that does (or *should*) stand for the person(s) who
> can speak for a LIR, legally, is "admin-c:". "abuse-c:" is some role
> account in a NOC or even a ticket system unlikely to have any
> decision-making power. An attempt to make these roles (perhaps
> even personally) responsible for the behaviour of a LIR and its
> customers is counter-productive. I for one would flatly refuse to
> do any abuse report handling under these circumstances.
>
>> Complying with national, civil and criminal due process laws in
>> support of investigations and providing justice for victims.
>
>
> Would the proposers please amplify exactly which law or due
> process is violated by the NCC not sending an email once a year
> to an abuse-c:? Myself, and I'm sure the NCC legal team, would be
> interested to know.
>
> On a technical note, email is neither a secure nor a reliable
> transport for such verification. In the day of blocklists and
> large email providers imposing arbitrary restrictions on email
> senders it is not guaranteed that a verification email reaches
> the intended address or that a reply reaches the sender. The NCC
> ARC procedure, however employs both email and personal contact
> via telephone to verify the accuracy of the ripedb information.
>
> In toto, this proposal would impose unneccessary work on LIRs and
> the NCC, using unsuitable means to rectify a non-existing issue, and I therefore oppose it.
>
>