Methodology
Every conclusion here comes from fixed rules applied to public records (classifier version 1.0.0). No AI or language model decides anything. Each change keeps the records it rests on, and when a rule changes, earlier conclusions are kept as history.
The sources
- ARIN-Issued: ARIN's daily mailing-list report of Add and Remove lines. Read from the public archive; each message is stored with its SHA-256.
- ARIN's transfer log (NRO format): every transfer since 2009 with date, type, both organizations and both registries. The strongest evidence there is.
- ARIN's delegated-extended statistics: one file a day listing every ASN and address block ARIN administers and its status (assigned, allocated, reserved or available), checked against ARIN's MD5.
- RDAP, found through the IANA bootstrap: the current registry and organization for changes as they're reported. Only organization-level details are kept; points of contact are dropped before anything is saved.
ARIN's WhoWas service isn't queried automatically (ARIN asks that it not be used for bulk research). A conclusion from manual WhoWas research is marked as such, with who did it, when and why.
Why "Add" doesn't always mean allocation
When a resource changes hands, ARIN re-registers it to the recipient, and ARIN-Issued prints that as Add. So an Add can be a transfer within ARIN, a merger or acquisition, a transfer from another registry, or even a transfer out to another registry. Only when no transfer explains it, and ARIN's statistics show the resource moving from ARIN's unissued pool (reserved or available) to a registration dated around the report, does the tracker call it an apparent new issuance, or an apparent reissuance when the tracker saw it removed or released from hold earlier.
What "Remove" really is
Since 2026-01-01, not one of 1,736 Remove lines was a registration being removed. 1,735 were changes to ARIN's own pool records:
- Hold released: a reserved (held) resource ARIN moved to the status it calls 'available'. ARIN often does this in batches, then reissues them over the following weeks.
- ARIN pool record: ARIN removing a reserved block's record because part of it was just issued (the same report adds the issued pieces), or re-cutting its reserved blocks.
The actual removals from registration (assigned or allocated to reserved) don't appear on ARIN-Issued at all. The tracker finds them by comparing each day's statistics with the day before, and lists them as Not in ARIN-Issued. Transfers out to other registries are often missing from ARIN-Issued too, and are found the same way.
Why "Remove" doesn't tell you the reason
Public records don't say whether a registration was returned voluntarily, revoked (for example for non-payment) or reclaimed. The tracker never guesses: removals are Removed (return, revocation or reclamation not publicly distinguishable).
How transfers are matched
A change matches a transfer record when the ranges overlap and the transfer happened within 48 hours of the report day (ARIN batches its reports and the log is in UTC). When several records overlap, one that matches the range exactly wins; several records with the same parties, registries and type that together cover the range exactly are one transfer recorded in pieces. Anything else is conflicting and goes to review rather than being guessed.
Blocks sold in pieces
A transfer of part of a block, while the rest stays behind, is marked as a piece of that block. The block is the one ARIN's statistics listed on the earliest day the tracker has, or the day before the transfer, whichever is larger, since ARIN often re-lists what's left in smaller records after a sale. Each block sold in pieces gets its own page with every piece, seller and buyer, and the share of it transferred so far. The transfer itself is classified as usual.
How the statistics are compared
ARIN publishes each day's statistics file at about 09:00 Eastern, so it can already include that morning's changes. The tracker therefore compares the file from the day before a report with the file from the day after. When another change to the same range was reported the day before or after, the comparison may include it, and the conclusion is marked observed only. Re-cuts of a block into differently sized records, with the same holder and status, are recognised as a change of representation, not of substance.
Why reserved or 'available' doesn't mean you can get it
Reserved space includes ARIN's hold pool and blocks set aside for policy. 'Available' is ARIN's own label in its statistics. Neither is a promise that ARIN will issue the resource, or to whom; ARIN publishes no definitive live list of what applicants can receive. The tracker only ever says possible reissue candidate.
Hold estimates
ARIN's published rule holds returned, revoked and reclaimed ASNs and IPv6 space for six calendar months, and IPv4 space for 60 days, before reissuing them. The tracker adds that to the day it saw the removal (six months after 31 August is 28 or 29 February). It's the earliest the rule allows, not a date ARIN will act on. Of the 15 resources the tracker saw both removed and released from hold, the hold lasted 7 to 145 days (median 20).
Confidence
- Confirmed
- An explicit authoritative record identifies the event, such as a matching transfer record.
- Strong inference
- ARIN's statistics before and after support one interpretation, but no record of the transaction exists.
- Observed only
- The change is known, but public evidence doesn't establish the cause (or another change nearby clouds it).
- Conflicting
- Sources disagree, or the records can't be reconciled. Queued for review.
Confidence is never a percentage. Recent conclusions are provisional for seven days, while the following days' statistics, the transfer log and RDAP are checked again; then they settle.
Classifications
| Classification | Inventory effect | Since 2026-01-01 |
|---|---|---|
| Intra-ARIN transfer | No effect on ARIN applicant inventory | 750 |
| Inter-RIR transfer into ARIN | Transferred into ARIN administration; not a new issuance | 50 |
| Inter-RIR transfer out of ARIN | Transferred out of ARIN administration; not offered to ARIN applicants | 631 |
| Merger/acquisition or reorganization | No effect on ARIN applicant inventory | 0 |
| Apparent new issuance | Likely issued from ARIN-controlled inventory | 2,786 |
| Apparent reissuance of previously removed resource | Likely issued from ARIN-controlled inventory | 400 |
| Removed (return, revocation or reclamation not publicly distinguishable) | Possible future reissue candidate after hold | 1,572 |
| Registration or organization record change | No effect on ARIN applicant inventory | 0 |
| Range split/merge or reporting-format change | No effect on ARIN applicant inventory | 0 |
| No substantive resource change found | No effect on ARIN applicant inventory | 172 |
| Unresolved change | Inventory effect unknown | 3 |
| Released from hold (ARIN lists it as 'available') | ARIN's statistics now list it as 'available'; reissue not confirmed | 833 |
| Change to ARIN's own pool record | No effect on ARIN applicant inventory | 732 |
Freshness and limits
- The tracker reads its sources once a day, at about 14:30 UTC, after ARIN has published the day's report (about 03:30 UTC), its transfer log (about 04:00) and its statistics file (about 13:00). Each day's report appears the following afternoon (UTC), already compared with ARIN's statistics, and settles after seven days.
- It covers reports from 2026-01-01 on. Earlier removals aren't known to it, so some reissues count as new issuances and some releases from hold have no removal date.
- Routing (BGP) is not used: a resource can be registered without being routed, and routing proves neither ownership nor availability.
- It's an independent project and is not affiliated with or endorsed by ARIN, the NRO or any Regional Internet Registry.
This site independently analyzes publicly available Internet number-resource records. It is not affiliated with or endorsed by the American Registry for Internet Numbers (ARIN), the Number Resource Organization (NRO), or any Regional Internet Registry. Classifications and hold-expiration dates are evidence-based interpretations and do not constitute confirmation that a resource is available for issuance or transfer.