You don't need to wait until 2022.
As I explain you before, anyone can send 158.6 million TCP whois packets in under an hour, using a few IPs and a single $2 VPS.
Verisign might notice the uptick in those packets. It has been in the business for a few decades and regularly sees such attempts to mine the WHOIS. It also rate limits requests from single IP addresses and may block them. That means that more disposable IP addresses would be necessary.
Then there's the problem of a single dataset. All that your dataset would represent would be a single snapshot of domain names in your list. You would not know how many have been transferred out or transferred in to Epik. You also don't know how many registrations Epik has lost through deletions or gained via new registrations. This means that you have to create multiple datasets for comparison and do this for each gTLD you wish to check. From the legacy gTLD set, there are .COM/NET/ORG/BIZ/INFO/MOBI/ASIA/CAT/COOP/JOBS/MUSEUM/NAME/PRO/TEL/TRAVEL/XXX and the relatively inactive .POST. There are also over 1,100 new gTLDs.
You just need the 158.6m com from the zone file. The same with other gtlds, you can get them on czds.
This brings up another problem. The number of domain names under management by a registry is not always the same as those in the zone file. Almost every gTLD has a number of dark domain names. These are domain names without nameservers. The link below shows how the .COM and .NET are actually slightly larger than the zone files.
https://www.verisign.com/en_US/channel-resources/domain-registry-products/zone-file/index.xhtml
Some of those domain names are in their pending-delete phase when their nameservers are removed and they are about to drop within five days. Others are intentionally dark due to legal action or action by their registrants.
As for the CZDS, not all registries grant access to zone file requests and there are frequent outages while registries renew access requests every 90 days or so. It was a perfectly good specification but ICANN managed to break one of the most important aspects of it (continual access to the zone file unless the registry deactivated it).
Some of the registries, especially the new gTLDs, are moving away from the old WHOIS system to the RDAP system (
https://www.icann.org/rdap). That gives the registries a lot more control over access than the WHOIS system.
The only ones missing would be cctld.
Some of the ccTLD registries make the gTLDs seem extremely open by comparison. Registries such as DEnic (the .DE ccTLD registry) doesn't publish anything other than the domain name status and the nameservers via WHOIS. Eurid, (the .EU registry), will provide the registrar via WHOIS but will provide more data via web-based WHOIS. Other registries don't even provide WHOIS. But the big problem with ccTLDs is that access to the zone files varies. Many ccTLD registries will not provide access to their zone files.
Anyone interested doesn't have to wait until 2022.
You can see the transaction reports for .COM here: (
https://www.icann.org/resources/pages/com-2014-03-04-en ) They are in comma separated variable format and should be readable as a text file or in any Open Source spreadsheet software. As you can see, they provide much more information than simply domain names under management for each registrar. (number of nameservers, number of new registrations by years registered, number of renewals by years renewed, number of transfer gains, number of transfer losses, number of deletions, number of deletions for which the registrar did not have to pay (deleted grace). They are statistics rather than actual domain names.
What people outside the domain name industry do not understand is that domain name registrations at a registrar level tend to be incredibly sticky. People, for reasons best known to themselves, tend to register domain names and keep them with the same registrar for the lifetime of the domain name. What has happened with Epik over the last few weeks is that some portfolio operators (they own large numbers of domain names) have changed their nameservers to those of Dan, Afternic, Bodis and others. They may not have actually changed the registrar for their domain names. Most of the domain names on Epik are for sale. These domain names do not have developed websites.
The latest (May 2021) .COM report only shows 701,454 domain names (in total) being transferred. Transfer from one set of nameservers to another is typically higher. Those kinds of generally transfer show up in the changed WHOIS record and in the zone files. Tracking all these changes, even with the registry reports, and understanding them is a complex business.
Regards...jmcc