Whois reform to take four years, cost up to $107 million A YEAR, and may still be pointless

ICANN’s proposed post-GDPR Whois system could cost over $100 million a year to run and take up to four years to build, but the Org still has no idea whether anyone will use it.

That appears to be the emerging conclusion of ICANN’s very first Operational Design Phase, which sought to translate community recommendations for a Standardized System for Access and Disclosure (SSAD) into a practical implementation plan.

SSAD is supposed to make it easier for people like trademark owners and law enforcement to request personal information from Whois records that is currently redacted due to privacy laws such as GDPR.

The ODP, which was originally meant to conclude in September but will now formally wrap up in February, has decided so far that SSAD will take “three to four years” to design and build, costing between $20 million and $27 million.

It’s calculated the annual running costs at between $14 million and $107 million, an eye-wateringly imprecise estimate arrived at because ICANN has pretty much no idea how many people will want to use SSAD, how much they’d be prepared to pay, and how many Whois requests they will likely make.

ICANN had previously guesstimated startup costs of $9 million and ongoing annual costs around the same level.

The new cost estimates are based on the number of users being anywhere between 25,000 and three million, with the number of annual queries coming in at between 100,000 and 12 million.

And ICANN admits that the actual demand “may be lower” than even the low-end estimate.

“We haven’t been able to figure out how big the demand is,” ICANN CEO Göran Marby told the GNSO Council during a conference call last month.

“Actual demand is unknowable until well after the launch of the SSAD,” an ICANN presentation (pdf) states. The Org contacted 11 research firms to try to get a better handle on likely demand, but most turned down the work for this reason.

On pricing, the ODP decided that it would cost a few hundred bucks for requestors to get accredited into the system, and then anywhere between $0.45 and $40 for every Whois request they make.

Again, the range is so laughably broad because the likely level of demand is unknown. A smaller number of requests would lead to a higher price and vice versa.

Even if there’s an initial flurry of SSAD activity, that could decline over time, the ODP concluded. In part that’s because registries and registrars would be under no obligation to turn over records, even if requestors are paying $40 a pop for their queries.

It’s also because SSAD would not be mandatory — requestors could still approach contracted parties directly for the info they want, for low or no cost, if they think the price of SSAD is too high or accreditation requirements too onerous.

“There’ll always be a free version of this for everybody,” Marby said on the conference call.

In short, it’s a hell of a lot of money for not much functionality. There’s a better than even chance it could be a huge waste of time and money.

An added complication is that the laws that SSAD is supposed to address, mainly GDPR, are likely to change while it’s being implemented. The European Union’s NIS2 Directive stands to move the goalposts on Whois privacy substantially, and not uniformly, in the not-too-distant future, for example.

This is profoundly embarrassing for ICANN as an organization. Created in the 1990s to operate at “internet speed”, it’s now so bloated, so twisted up it its own knickers, that it’s getting lapped by the lumbering EU legislative process.

The ODP is set to submit its final report to ICANN’s board of directors in February. The board could theoretically decide that it’s not in the interest of ICANN or the public to go ahead with it.

Marby, for his part, seems to be thinking that there could be some benefit from a centralized hub for submitting Whois requests, but that it should be simpler than the current “too complex” proposal, and funded by ICANN.

My take is that ICANN is reluctant to move ahead with SSAD as it’s currently proposed, but because top-down policy-making is frowned upon its hands are tied to make the changes it would like to see.

ICANN teases prices for private Whois lookups

ICANN has started to put some flesh on the bones of the forthcoming (?) SSAD system for accessing private Whois records, including teasing some baseline pricing.

During a session at ICANN 72 last week, staffers said responses to recent requests for information put the cost of having an identity verified as an SSAD user at about $10 to $20.

Those are vendor wholesale prices, however, covering the cost of looking at a government-issue ID and making sure it’s legit, and do not include the extra administration and cost-recovery charges that ICANN plans to place on top.

The verification fee would have to be renewed every two years under ICANN’s proposal, though the verification vendors are apparently pushing for annual renewals.

The fee also would not include the likely per-query charge that users will have to pay to request the true personal data behind a redacted Whois record.

It’s not currently anticipated that any money would flow to registrars, CEO Göran Marby said.

SSAD, the Standardized System for Access and Disclosure, is currently undergoing Operational Design Phase work in ICANN, with monthly webinar updates for the community.

ICANN expects to reveal more pricing details on the December webinar, staffers said.

ICANN adds another six months to Whois reform roadmap

ICANN says that its preparatory work for possible Whois reforms will take another six months.

The Operational Design Phase for the System for Standardized Access and Disclosure will now conclude “by the end of February 2022”, ICANN said this week.

That’s after the Org missed its original September deadline after six months of work.

ICANN program manager Diana Middleton said at ICANN 72 last week that ODP had been delayed by various factors including surveys taking longer than expected and throwing up more questions than they answered.

A survey of Governmental Advisory Committee members due September 17 was extended until the end of October.

But she added that ICANN intends to throw its first draft of the output — an Operational Design Assessment — at its technical writers by the end of the month, with a document going before the board of directors in early February.

SSAD is the proposed system that would funnel requests for private Whois data through ICANN, with a new veneer of red tape for those wishing to access such data.

The ODP is ICANN’s brand-new process for deciding how it could be implemented, how much it would cost, and indeed whether it’s worthwhile implementing it at all.

It’s also being used to prepare for the next round of new gTLDs, with a 13-month initial deadline.

The longer the current ODP runs, the greater the cost to the eventual SSAD user.

Whois rule changes that nobody likes get approved anyway

ICANN’s Generic Names Supporting Organization Council has approved a handful of changes to Whois policy, despite the fact that pretty much nobody was fully on-board with the proposals and how they were made.

The new recommendations call for a new field in Whois records to flag up whether the registrant is a private individual, whose privacy is protected by law, or a legal entity like a company, which have no privacy rights.

But the field will be optional, with no obligation for registries or registrars to use it in their Whois services, which has angered intellectual property interests, governments and others.

The working group that came up with the recommendations also declined to find that Whois records should come with an anonymized registrant email address as standard. This absence of change was also adopted by the Council, causing more disappointment.

In short, nothing much is happening to Whois records for the foreseeable future as a result of these policy changes.

But the process to arrive at this conclusion has highlighted not just the deep divisions in the ICANN community but also, some argue, deficiencies in the ICANN process itself.

The Expedited Policy Development Process working group that has since 2018 been looking at the interaction between Whois and privacy protection law, primarily the European Union’s General Data Protection Regulation, had been asked two final questions earlier this year, to wrap up its long-running work.

First, should registrars and registries be forced to distinguish between legal and natural persons when deciding what data to publish in Whois?

Second, should there be a registrant-based or registration-based anonymized email published in Whois to help people contact domain owners and/or correlate ownership across records?

The answer on both counts was that it’s up to the registry or registrar to decide.

On legal versus natural, the EPDP decided that ICANN should work with the technical community to create a new field in the Whois standard (RDAP), but that there should be no obligation for the industry to use it.

On anonymized email addresses, the working group recommendations were even hand-wavier — they merely refer the industry to some legal advice on how to implement such a system in a GDPR-compliant way.

While this phase of the EPDP’s work was super-fast by ICANN standards (taking about nine months) and piss-weak with its output, it nevertheless attracted a whole lot of dissent.

While its tasks appeared straightforward to outsiders, it nevertheless appears to have inherited the simmering tensions and entrenched positions of earlier phases and turned out to be one of the most divisive and fractious working groups in the modern ICANN period.

Almost every group involved in the work submitted a minority statement expressing either their displeasure with the outcome, or with the process used to arrive at it, or both. Even some of the largely positive statements reek of sarcasm and resentment.

EPDP chair Keith Drazek went to the extent of saying that the minority statements should be read as part and parcel of the group’s Final Report, saying “some groups felt that the work did not go as far as needed, or did not include sufficient detail, while other groups felt that certain recommendations were not appropriate or necessary”.

This Final Report constitutes a compromise that is the maximum that could be achieved by the group at this time under our currently allocated time and scope, and it should not be read as delivering results that were fully satisfactory to everyone.

The appears to be an understatement.

The Intellectual Property Constituency and Business Constituency were both the angriest, as you might expect. They wanted to be able to get more data on legal persons, and to be able to reverse-engineer domain portfolios using anonymous registrant-baed email addresses, and they won’t be able to do either.

The Governmental Advisory Committee and Security and Stability Advisory Committee both expressed positions in line with the IPC/BC, dismayed that no enforceable contract language will emerge from this process.

Councilor Marie Pattullo of the BC said during the GNSO Council vote last Wednesday that the work “exceeds what is necessary to protect registrant data” and that the EPDP failed to “preserve the WHOIS database to the greatest extent possible”.

The “optional differentiation between legal and natural persons is inadequate”, she said, resulting in “a significant number of records being needlessly redacted or otherwise being made unavailable”. The approved policies contain “no real policy and places no enforceable obligations on contracted parties”, she said.

IPC councilor John McElwaine called the EPDP “unfinished work” because the working group failed to reach a consensus on the legal/natural question. The IPC minority statement had said:

Requiring ICANN to coordinate the technical community in the creation of a data element which contracted parties are free to ignore altogether falls far short of “resolving” the legal vs. natural issue. And failing to require differentiation of personal and non-personal data fails to meet the overarching goal of the EPDP to “preserve the WHOIS database to the greatest extent possible” while complying with privacy law.

But McElwaine conceded that “a minority of IPC members did favor these outputs as being minor, incremental changes that are better than nothing”.

The BC and IPC both voted against the proposals, but that was not enough to kill them. They would have needed support from at least one councilor on the the other side of the GNSO’s Non-Contracted Parties House, the Non-Commercial Stakeholders Group, and that hand was not raised.

While the NCSG voted “aye”, and seemed generally fine with the outcome, it wasn’t happy with the process, and had some stern words for its opponents. It said in its minority statement:

The process for this EPDP has been unnecessarily long and painful, however, and does not reflect an appreciation for ICANN’s responsibility to comply with data protection law but rather the difficulty in getting many stakeholders to embrace the concept of respect for registrants’ rights…

With respect to the precise issues addressed in this report, we have stressed throughout this EPDP, and in a previous PDP on privacy proxy services, that the distinction between legal and natural is not a useful distinction to make, when deciding about the need to protect data in the RDS. It was, as we have reiterated many times, the wrong question to ask, because many workers employed by a legal person or company have privacy rights with respect to the disclosure of their personal information and contact data. The legal person does not have privacy rights, but people do.

While welcoming the result, the Registrars Stakeholder Group had similar concerns about the process, accusing its opponents of trying to impose additional legal risks on contracted parties. Its minority statement says:

it is disappointing that achieving this result was the product of significant struggle. Throughout the work on this Phase, the WG revisited issues repeatedly without adding anything substantially new to the discussion, and discussed topics which were out of scope. Perhaps most importantly, the WG was on many occasions uninterested in or unconcerned with the legal and financial risks that some proposed obligations would create for contracted parties in varying jurisdictions or of differing business models, or the risks to registrants themselves.

The Registries Stakeholder Group drilled down even more on the “out of scope” issue, saying the recommendation to create a new legal vs natural field in Whois went beyond what the working group had been tasked with.

They disagreed with, and indeed challenged, Drazek’s decision that the discussion was in-scope, but reluctantly went ahead and voted on the proposals in Council in order to finally draw a line under the whole issue.

The question of whether the legal vs natural question has been in fact been resolved seems to be an ongoing point of conflict, with the RySG, RrSG and NCSG saying it’s finally time to put the matter to bed and the IPC and BC insisting that consensus has not yet been reached.

The RySG wrote that it is “well past time to consider the issue closed” and that the EPDP had produced a “valuable and acceptable outcome”, adding:

The RySG is concerned that some have suggested this issue is not resolved. This question has been discussed in three separate phases of the EPDP and the result each time has been that Contracted Parties may differentiate but are not required to do so. This clearly demonstrates that this matter has been addressed appropriately and consistently. A perception that this work is somehow unresolved could be detrimental to the ICANN community and seen as undermining the effectiveness of the multistakeholder model.

Conversely, the BC said the report “represents an unfortunate failure of the multistakeholder process” adding that “we believe the record should state that consensus opinion did not and still does not exist”.

The IPC noted “a troubling trend in multistakeholder policy development”, saying in a clear swipe at the contracted parties that “little success is possible when some stakeholders are only willing to act exclusively in their own interests with little regard for compromise in the interest of the greater good.”

So, depending on who you believe, either the multistakeholder process is captured and controlled by intransigent contracted parties, or it’s unduly influenced by those who want to go ultra vires to interfere with the business of selling domains in order to violate registrant privacy.

And in either case the multistakeholder model is at risk — either “agree to disagree” counts as a consensus position, or it’s an invitation for an infinite series of future policy debates.

Business as usual at the GNSO, in other words.

Alice’s Registry disappears down the rabbit hole

One of the oldest domain registrars appears to be on its way out.

San Francisco-based Alice’s Registry has been hit with a breach notice and termination warning by ICANN after apparently being incommunicado for over a year.

According to ICANN, they last spoke in August 2020, when AR indicated that it was thinking about “shutting down the registrar business”.

Since then, the web site has stopped working and ICANN can’t get through on the telephone.

The breach notice claims past-due fees and a failure to operate a working Whois service, and gives the registrar until November 1 to pay up or get its contract terminated.

Alice’s Registry is one of the oldest registrars, founded in 1999, but it’s never had more than a few thousand names under management. Its founder, Rick Wesson, has been involved in the ICANN community since pretty much the beginning.

Almost no security researchers asking for Whois records – Tucows

Security researchers are not asking for private Whois records in anywhere near the numbers you might have been led to believe, according to data released this week by Tucows.

The registrar revealed that it received just one request from the security community between September 2020 and the end of August 2021. That’s not even 1% of the total.

Over the same period, the “commercial litigators” category, presumably including intellectual property interests going after suspected cybersquatters, were behind 87% of requests.

About 9% of requests came from law enforcement agencies, Tucows said.

The company said that it disclosed private registrant data in 74% of cases. It denied the requests in 9% of cases. Other requests were incomplete or abandoned.

Tucows has been offering a Tiered Access service for its Whois records since the General Data Protection Regulation came into effect in May 2018. It has received 4,478 requests since then.

Price of Whois lookups could rise as ICANN delays reform work

ICANN has delayed the conclusion of work on Whois reform, potentially increasing the cost of requesting domain registration data in future.

Back in March, its board of directors gave the Org six months to complete the Operational Design Phase of the so-called SSAD, or System for Standardized Access and Disclosure, but that deadline passed this week.

It appears that ICANN is not even close to concluding its ODP work. No new deadline has been announced, but ICANN intends to talk to the community at ICANN 72 next month.

SSAD is a proposal created by the community and approved — not without controversy — by the GNSO Council. It would essentially create a centralized clearinghouse for law enforcement and intellectual property interests to request private registrant data from registries and registrars.

The ODP is a new process, never before used, whereby ICANN clarifies the community’s intentions and attempts to translate policy recommendations into a roadmap that is feasible and cost-effective to implement.

It seems this process suffered some teething troubles, which are partially responsible for the delays.

But it also appears that ICANN is having a hard time finding potential service provider partners capable of building and operating SSAD all by themselves, raising the prospect of a more complex and expensive piecemeal solution.

It had 17 responses to a recent RFI, but no respondent said it could cover all the bases.

The key sticking point, described by some as a “chicken and egg” problem, is figuring out how many people are likely to use SSAD and how often. If the system is too expensive or fails to deliver results, it will be used less. If it works like a charm and is cost-effective, query volumes would go up.

So ICANN is challenged to gaze into its crystal ball and find a sweet spot, balancing cost, functionality and usage, if SSAD is to be a success. So far, its estimates for usage range from 25,000 users making 100,000 requests a year to 3 million users making 12 million requests.

That’s how far away from concluding its work ICANN is.

Confounding matters, the longer ICANN drags its feet on the ODP phase, the more expensive SSAD is likely to be for the end users who will ultimately wind up paying for it.

In a webinar last week, CEO Göran Marby said that the SSAD project is meant to recover its own costs. Whatever ICANN is spending on the ODP right now is expected to be recouped from access fees when SSAD goes live.

“This should not cost ICANN Org anything,” he said. “The costs should be carried by the user.”

ICANN is working on the assumption that SSAD will eventually happen, but if the ODP decides not to implement SSAD, ICANN will have to eat the costs, he indicated.

When the ICANN board approved this ODP, it did not specify how much money was being allocated to the project.

A second and separate ODP, looking at the next round of new gTLDs, was earlier this month given $9 million to conduct an anticipated 10-month project.

More privacy headaches? UK to withdraw from GDPR

The UK is to craft its own privacy legislation, after Brexit enabled it to extricate itself from the EU’s General Data Protection Regulation, potentially causing headaches for domain name companies.

While it’s still in the very early pre-consultation stages, the government announced today that it wants “to make the country’s data regime even more ambitious, pro-growth and innovation-friendly, while still being underpinned by secure and trustworthy privacy standards.”

The country looks to be heading to a new privacy regime that registries and registrars doing business there will have to comply with, particular with regard to Whois services, in other words.

But it might not be too bad — the government is talking up plans to make “data adequacy” deals with third countries to enable the easy, legal transfer of private data across borders, which is always useful in the context of domain names.

While the UK is no longer in the EU, most EU laws including GDPR were grandfathered in and are still in effect.

As Kabul falls, Whois could present a danger to ordinary Afghans

With Afghanistan falling to the Taliban this week, there’s potential danger to .af registrants — both in terms of losing domain services and of Whois being used for possibly deadly reprisals.

At time of writing, it’s been four days since the fall of Kabul. The uneasy truce between NATO and Taliban forces has failed to prevent scenes of chaos at the city’s main airport and the PR machine of so-called “Taliban 2.0” is in full bluster.

The new Taliban is, its spokespeople suggest, more tolerant of western liberal values and more supportive of human rights than its brutal, pre-9/11 incarnation.

Few believe this spin, and there have been multiple reports of 1990s-style oppression, including revenge killings and the suppression of women’s rights, across the country.

With all that in mind, a blog post about .af domain names may seem trivial, but it’s not my intention to trivialize.

I’m as appalled as any right-minded observer by the situation on the ground in Afghanistan and the neglect that led to it. But I believe .af could prove a learning moment in the ongoing conversation about Whois privacy.

The .af ccTLD has been managed since not long after the US-led invasion by the country’s Ministry of Communications and IT as the Afghanistan Network Information Center.

The registry had previously been managed for free from London by NetNames, with an admin contact in Kabul, according to the report of the 2003 IANA redelegation, which happened at a time when Afghanistan was still under a transitional government heavily overseen by the foreign governments behind the invasion.

Domain policy for .af was created in 2002, and it includes provisions for an open, freely available Whois database that is still in effect today.

Domains registered via overseas registrars appear to be benefiting from the impact of the EU’s General Data Protection Regulation, which redacts personal information, but this obviously does not apply in Afghanistan.

This means the names, addresses, phone numbers and email addresses of .af registrants are available for querying via various Whois interfaces, including the registry’s own, which is managed by New Zealand-based back-end CoCCA.

Using a combination of web searches and Whois queries, it is possible to find personally identifiable information of registrants, including names and addresses, at local human rights groups, as well as local news media and technology providers supportive of human rights causes.

If the reports of Taliban fighters conducting house-to-house searches for enemies of the new state are accurate, the easy availability of this personal data could be a serious problem.

To a great extent, this could be a case study in what privacy advocates within the ICANN community are always warning about — public access to Whois data gives oppressive regimes a tool to target their oppression.

And as we have seen this week, oppressive regimes can appear almost literally overnight.

While it seems unlikely there’s anyone from the old Afghan ministry still in control of the registry, I think .af back-end provider CoCCA, as well as Whois aggregators such as DomainTools, should have a long think about whether it’s a good idea to continue to provide open access to .af Whois records at this time.

Fortunately, there doesn’t appear to be a great many .af domains under management. DomainTools reckons it’s under 7,000.

At the other end of the scale of seriousness, overseas .af registrants may also see issues with their names due to the Taliban takeover.

It seems incredible today, but in 2001 a Taliban decree restricted internet access to a single computer at a government ministry. Others in government could apply to use this computer by sending a fax to the relevant minister.

While it seems impossible that such a Draconian restriction could be reintroduced today, it still seems likely that the Taliban will crack down on internet usage to an extent, including introducing morality or residency restrictions to .af regs.

.af is currently open to registrants from anywhere in the world, with no complex restrictions and .com-competitive prices.

Many multinational corporations have registered .af names for their local presence.

The string “af” has in recent years become social media shorthand for “as fuck”, and a small number overseas registrants appear to be using it as a domain hack in that context — type “corrupt.af” into your browser and see what happens.

Others seem to be using .af, where short domains are still available, as shortcuts to their social media profiles.

I don’t believe ICANN will need to get directly involved in this situation. Its Whois query tool does not support .af, and IANA presumably won’t need to get involved in terms of redelegation any more than it would following a general election or a coup d’état.

Domainers at risk as EnCirca takes over deadbeat registrar’s customer base

Customers of defunct registrar Pheenix risk losing their domains because the company was not properly escrowing its registrant data, according to the registrar taking over their domains.

EnCirca, which is taking over up to 6,000 domains previously registered with Pheenix, says the registrar’s shoddy escrow practices mean some of these domains may not be reunited with their rightful owners.

Pheenix “failed to properly escrow domain ownership information for many of the domains utilizing WHOIS proxy services”, EnCirca recently wrote, adding:

We anticipate that many domains will remain unclaimed due to bounced emails or inoperable proxy services. Locating rightful owners will be problematic since the data escrow is often devoid of any identifying ownership information.

To try to mitigate the problem, EnCirca is offering affected registrants the chance to prove ownership by filling out a form and uploading other evidence, such as Pheenix receipts or bank statements.

EnCirca added that because Pheenix disappeared still owing money to registries, the registries may be forcing renewal or restore fees that will then be passed on registrants.

If your domains were at or near expiration, restoring them could be complex and pricey or impossible.

If you’re affected, you can find information here.

Most or all Pheenix customers are likely to be domain investors. It was a drop-catcher, which once had over 500 dummy registrars in its expansive dropnet, most of which it subsequently de-accredited.

But it went AWOL last May, not responding to ICANN or paying its dues, apparently disappearing from the face of the Earth.

ICANN terminated its accreditation in May this year, and initiated a bulk transfer to EnCirca a couple weeks ago (which it only disclosed this week).

EnCirca has experience handling this kind of problem, which is presumably why ICANN gifted it the bulk transfer. In 2018 it took on the domains 49 of Pheenix’s shell registrars, which it says were suffering from the same escrow problems.