iOS Artifacts

SIM Card Changes on iPhone: What CellularUsage.db and CommCenter Record

Validated4n6·

"Was a different SIM card in this phone?" comes up in drug investigations, stalking investigations, homicide investigations and any investigation where a suspect may have swapped his/her telephone number. iOS keeps a record of the SIM cards it has seen. The question is what the timestamps next to them actually mean.

I swapped SIM cards in and out of a test iPhone on a logged timeline, then compared that timeline against the two main SIM artifacts: CellularUsage.db and com.apple.commcenter.data.plist. Both recorded every SIM. Neither recorded every swap.

Key findings

  • Both sources listed all three SIM cards ever used in the phone, with each card's ICCID and phone number.
  • Each source keeps one timestamp per SIM. Earlier insertions and removals of the same card were overwritten.
  • The CellularUsage.db time for the second SIM matched when the phone powered on with it, not when it was inserted.
  • In com.apple.commcenter.data.plist, the ts value for a removed SIM matched when it was ejected. For the SIM in the phone, it matched when it was inserted.
  • For the oldest SIM, the two sources gave dates almost two years apart. They are recording different events.
  • The phone sent an iMessage with no SIM card in it at all, over Wi-Fi.

Where the data lives

SourcePathWhat it holds
CellularUsage.db/private/var/wireless/Library/Databases/CellularUsage.dbTable subscriber_info: ICCID (subscriber_id), phone number (subscriber_mdn), last_update_time, slot
CommCenter data plist/private/var/wireless/Library/Preferences/com.apple.commcenter.data.plistPersonalWallet: one entry per ICCID with phone number (mdn), timestamp (ts, Unix seconds), SIM type and flags

Test setup

ItemDetail
DeviceApple iPhone 7 (A1660)
iOS version15.8.2 (19H384)
Test date6 Mar 2025, Eastern Standard Time (UTC-5). Times below are local.
SIM cardsThree Verizon SIMs, called SIM A, SIM B and SIM C below. SIM C was used in the phone in earlier years and was not part of this test.
ExtractionFull file system, Cellebrite Inseyets UFED
Analysis toolsiLEAPP 2.1.0, Cellebrite Inseyets Physical Analyzer 10.5.0.1016, plist viewer

Actions on the device

The phone started powered off with no SIM card in it.

TimeAction
12:28Inserted SIM A and powered on. Status showed "No Service."
12:31Outgoing call attempt. Did not connect.
12:35Joined an iPhone hotspot over Wi-Fi. Outgoing call connected (50 seconds).
12:37Left the hotspot
12:38Outgoing call over cellular (10 seconds)
12:39Ejected SIM A. "No SIM Card Installed" message.
12:40Reinserted SIM A. Phone connected to Verizon.
12:42Outgoing call (11 seconds)
12:43Ejected SIM A
12:44Inserted SIM B. Phone connected to Verizon.
12:45Outgoing call (21 seconds)
12:46Powered off
12:47Powered on
12:48Outgoing call (20 seconds)
12:49Ejected SIM B
12:50Joined the hotspot with no SIM in the phone
12:51Sent an iMessage over Wi-Fi with no SIM. It was delivered.
12:53Reinserted SIM A
12:54Left the hotspot
12:56Connected to the extraction tool

What CellularUsage.db recorded

The subscriber_info table had one row for each SIM.

SIMlast_update_time (UTC)LocalMatches
SIM A2025-03-06 17:53:0112:53:01Final reinsertion of SIM A
SIM B2025-03-06 17:47:4612:47:46Power-on with SIM B, not its 12:44 insertion
SIM C2022-06-18 11:53:46Earlier use, outside this test

SIM A went into the phone three times during the test: 12:28, 12:40 and 12:53. Only the last one survived. SIM B was inserted at 12:44 and used for a call at 12:45, but its row shows 12:47:46, a minute after the phone was restarted.

That fits what BlackBag reported when they first wrote up this database in 2016: many actions can change last_update_time, so it should not be read as "the SIM was inserted at this time." BlackBag also found the table only kept the last three SIM cards on the iOS versions they tested. This phone had exactly three, so I couldn't check whether that limit still applies on iOS 15.

What com.apple.commcenter.data.plist recorded

Under PersonalWallet, the plist had one dictionary per ICCID. Each held an info dictionary with the phone number (mdn), a Unix timestamp (ts), the SIM type and several flags.

SIMtsUTCLocalMatches
SIM A17412835802025-03-06 17:53:0012:53:00Final reinsertion of SIM A
SIM B17412833482025-03-06 17:49:0812:49:08SIM B ejected
SIM C17152896582024-05-09 21:20:58Earlier use, outside this test

So ts doesn't mean the same thing for every entry. For the SIM in the phone, it lined up with insertion. For the SIM that had been taken out, it lined up with removal. The simplest reading is that it marks the last time that SIM's state changed, but one test isn't enough to call that a rule.

iLEAPP 2.1.0 parses this plist and labels the value "Timestamp Unknown," which is a fair warning.

The flags

KeySIM A (in phone)SIM B (removed)SIM C (old)
typesimsimsim
no_srcnot presentTrueTrue
label-idGUID presentnot presentnot present
eap_akaTrueTrueTrue
cb_ver50.0.150.0.146.0.1

Only the SIM in the phone had a label-id, and only the SIMs not in the phone had no_src set to True. That makes those two keys a useful hint about which SIM was installed at extraction, though I haven't seen Apple document either one. cb_ver looks like a carrier settings version: SIM C, last used in 2024, has an older value than the two SIMs used in this test. Joshua Hickman has shown that the same plist also maps each SIM to a UUID that appears in call records (ZLOCALPARTICIPANTUUID in CallHistory.storedata), which is how you tie a call to a line on a dual-SIM phone.

The two sources disagree, and that's useful

For SIM C, CellularUsage.db says June 2022 and the plist says May 2024. Those can both be right, because they record different events. Neither one is a history. Each is a single snapshot of the last thing that happened to that SIM, as that file saw it. Put them side by side and you get two data points instead of one.

What neither source recorded

  • Repeated swaps. SIM A's first insertion at 12:28, its removal at 12:39, its reinsertion at 12:40 and its removal at 12:43 left no separate record in either source.
  • The time a SIM was in the phone. Neither source gives a start and end. You can't tell from these files alone that SIM B was in the phone from 12:44 to 12:49.
  • Communication without a SIM. The 12:51 iMessage went out with no SIM installed. The 12:35 call connected over Wi-Fi while the phone showed "No Service" (likely Wi-Fi Calling, though the call records weren't examined in this test). A missing or inactive SIM doesn't mean the phone wasn't communicating.

How to use this in casework

  1. Pull the SIM list from both CellularUsage.db and com.apple.commcenter.data.plist. Every ICCID and phone number is a lead, especially ones the user never mentioned.
  2. Report the timestamps for what they are: the last update each file recorded for that SIM. Don't call either one an insertion time unless you can corroborate it.
  3. Compare the two sources. Different dates for the same SIM give you two points on its timeline.
  4. Use label-id and no_src as hints for which SIM was installed at extraction, and confirm against the SIM tray and the carrier.
  5. To tie calls or messages to a specific line, use the SIM UUID mapping for calls and destination_caller_id in sms.db for messages.
  6. For a full history of when a line was active in this handset, get carrier records. The device keeps only the latest state.

Limits of this test

  • One device, one iOS version (15.8.2), physical SIMs only. eSIM and dual-SIM behavior was not tested.
  • All three SIMs were on the same carrier.
  • The device had only three SIMs, so the three-SIM limit BlackBag reported was not tested.
  • Call history, unified logs and other possible SIM change artifacts were not examined.
  • Action times were written down by hand, so allow a few seconds of difference.

Run the same test on your own device and iOS version before you rely on this in a report.

Bottom line

An iPhone remembers which SIM cards it has seen, and it will give you their ICCIDs and phone numbers. It doesn't remember every time they went in or came out. Each SIM gets one timestamp per file, and that timestamp can mean insertion, removal or a restart depending on what happened last. Use these artifacts to find the SIMs, then use carrier records to build the timeline.

Sources

Testing completed by Major Ricky Johnson, Richland County Sheriff's Department.