ELINT · Electronic Intelligence

Six false base station scenarios, six denials of service

NIST's draft white paper on false base stations ran six corner cases against commercial handsets in a laboratory 5G network. Every one produced a working denial of service, and in several the handset never returned to the legitimate cell until the rogue transmitter was switched off.

Cell selection in 5G is decided by signal strength, and the device cannot authenticate a cell before it connects to it. A draft white paper from the US National Institute of Standards and Technology now puts a number on what that costs: six for six. Every false base station scenario the authors evaluated against commercial handsets produced a working denial of service [1].

What was tested

NIST Cybersecurity White Paper 36G, False Base Station, was published as an initial public draft on 30 September 2026 by Michael Bartock and Jeffrey Cichonski of NIST with Parisa Grayeli and Sanjeev Sharma of MITRE [2]. It is the seventh instalment in the National Cybersecurity Center of Excellence’s CSWP 36 series, which also covers subscriber identifier concealment and temporary identity reallocation [4]. The comment window is stated inconsistently across NIST’s own pages: the publication record gives 30 October 2026, while the announcement, the project page and secondary coverage give 31 October 2026 [2][3][4][10].

The testbed is modest and entirely off the shelf. A single Ettus Research USRP B210 software-defined radio, the srsRAN open-source base station software and the Open5GS open-source core, both at 3GPP Release 17, on a laptop under Ubuntu, against Samsung Galaxy S22 and S23 handsets at Release 16 [1]. In five of the six scenarios the false base station impersonates both the radio access network and the core’s mobility management function; in the sixth, only the radio side. Every scenario advertises a tracking area identifier different from the home network’s [1].

What the results say

The six scenarios differ only in how the rogue cell declines to complete the connection. One ignores the radio connection request. One rejects it outright. One answers the radio setup and then ignores the registration message. One returns a forged registration acceptance. One returns a forged registration rejection. The last tampers with the initial non-access stratum message to claim that the device supports neither integrity protection nor ciphering [1].

Two outcomes are reassuring. The forged registration acceptance was ignored by the handset, which is mutual authentication working as designed. The tampered initial message was rejected by the core because 5G does not permit null integrity [1]. A decade of work on the confidentiality side did what it was built to do, and the device was not compromised in any of the six cases [1].

It was, however, denied service in all six. NIST groups the outcomes into two severity classes [1]:

  • A continuous registration loop, in which the handset retries without pause and the draft notes rapid battery depletion. This followed a radio-level rejection, a forged rejection carrying a “PLMN not allowed” cause, and the tampered initial message [1].
  • An intermittent registration cycle, in which the handset makes five attempts and then waits roughly twelve minutes before repeating. Back-off timers reduce the battery cost but not the outage [1].

The operationally important detail is not the battery drain. It is that in the first, second, third and sixth scenarios the handset kept returning to the rogue cell and did not attempt the legitimate one until the rogue transmitter was turned off [1]. The device had a working network available and could not reach it. NIST reports no power measurements behind the battery-depletion characterisation [1].

Why the protections that worked do not help here

The gap is structural rather than a defect in any implementation. Broadcast system information, paging, early radio connection setup, identity requests and rejection messages are all sent before a trusted security context exists, so they carry no integrity protection. Subscriber concealment and the 5G authentication and key agreement procedure do not cover cell selection at all, and moving to 5G standalone does not change that, because the cell remains unauthenticated [7].

3GPP has a study on this. Technical Report 33.809, Study on 5G security enhancements against False Base Stations, identified the relevant threat classes, including pre-connection fake signalling, cell blocking, forced downgrade and fake network map data, and is the document NIST’s draft cites for its threat model [1]. Its candidate mitigations include digitally signed broadcast system information and network-side detection using device measurement reports [7]. The report’s latest version, 18.1.0, is dated 20 September 2023 [5]. The baseline security specification, Technical Specification 33.501, reached version 20.3.0 on 28 September 2026 but does not mandate the false base station protections [6][7]. Three years of release cycles have moved the baseline and left the study where it was.

The attack that does not need a base station

Running a rogue cell is the conservative end of the technique. Overshadowing injects crafted signals on top of a legitimate base station’s transmission, timed and powered so that the receiver decodes the attacker’s bits instead, in either direction [7]. In 2025 a team from the Singapore University of Technology and Design published a framework that sniffs and injects messages before authentication without any rogue base station, reporting about 80 per cent sniffing accuracy and 70 to 90 per cent injection success at up to 20 metres across five tested handsets; GSMA acknowledged the work under identifier CVD-2024-0096 [9]. Separate 2026 work on signalling storms against the random access channel, validated on a physical testbed rather than in simulation, exploits the same root condition NIST’s scenarios turn on: the base station allocates radio resources before the user is fully authenticated [8].

The regional picture

Malaysian enforcement has targeted fraud delivery rather than denial of service, but the hardware class is the same. On 29 July 2026 the Malaysian Communications and Multimedia Commission reported an operation in Johor Bahru against a vehicle fitted with a fake base transceiver station that broadcast signals mimicking a licensed operator and delivered SMS messages carrying links outside the official network. A 32-year-old man was found driving it. The Commission said it had conducted five such operations to mid-2026, three in Genting Highlands and two in Johor, and cited section 239(1)(a) of the Communications and Multimedia Act 1998 and regulation 16(1)(b) of the Communications and Multimedia (Technical Standards) Regulations 2000 [11]; a 2025 operation in Kuala Lumpur seized 45 items of equipment valued at about MYR100,000 [12]. The point for a regional reader is supply: this equipment already circulates in vehicles in the Klang Valley and in Johor.

What to do

NIST’s draft directs its recommendations at standards bodies, operators and researchers [1]. For an organisation, the usable items are these:

  • Disable 2G and 3G fallback through device management profiles or device settings where the option exists. The draft lists this as a device-configurable protection, because older networks lack robust authentication and are easier to impersonate [1].
  • Ensure concealed subscriber identifiers are in use rather than the null scheme, so the permanent identifier is not exposed during the initial handshake [1].
  • If you operate a private 5G network, deploy false base station detection and consider an out-of-band or dual-connectivity path, satellite or Wi-Fi, for resilience. This is NIST’s recommendation to private operators [1].
  • On the network side, watch for measurement reports naming cells or physical cell identifiers the operator never deployed, clusters of downgrades from 5G to legacy technologies, abnormal random access patterns, and spikes in registration rejections that do not match planned changes [7].
  • Comment on the draft before the window closes at the end of October 2026, through the project page form [4].

For anyone whose continuity plan treats a mobile handset as the fallback when the fixed network fails, the draft is a reason to revisit that assumption. Confidentiality on the radio layer has been defended well and the evidence here says so. Availability has not been, and a single off-the-shelf radio is enough to demonstrate it. In electronic intelligence terms the distinction matters: an adversary who cannot read your traffic can still decide whether you have a network, and nothing in the current baseline obliges the handset to notice.

Sources

Every R3KONX article cites its primary material. 12 sources, in order of first citation. Links open the original publication.

  1. CSWP 36G, False Base Station: Applying 5G Cybersecurity and Privacy Capabilities (Initial Public Draft) NIST National Cybersecurity Center of Excellence · 2026-09-30
  2. CSWP 36G, False Base Station: Applying 5G Cybersecurity and Privacy Capabilities — publication record NIST Computer Security Resource Center · 2026-09-30
  3. Draft 5G White Paper Now Available: False Base Station NIST · 2026-09-30
  4. 5G Cybersecurity project page NIST National Cybersecurity Center of Excellence · 2026-10-01
  5. TR 33.809: Study on 5G security enhancements against False Base Stations (FBS) 3GPP · 2023-09-20
  6. TS 33.501: Security architecture and procedures for 5G System 3GPP · 2026-09-28
  7. Fake Base Station Attacks in 5G: Why Overshadowing Still Bypasses the Air Interface P1 Security · 2026-07-13
  8. Experimental Validation and Mitigation of RRC Storm Attacks in 5G Cellular Networks arXiv (2608.16441) · 2026-08-17
  9. Novel 5G Attack Bypasses Need for Malicious Base Station SecurityWeek · 2025-08-18
  10. NIST NCCoE releases draft 5G cybersecurity white paper on false base stations Australian Cyber Security Magazine · 2026-10-01
  11. MCMC busts fake mobile tower SMS scam syndicate in Johor Bahru Malay Mail · 2026-07-29
  12. Malaysian Regulator Shares Video of Raid That Seized SMS Blaster Gear Worth $24,000 Commsrisk · 2025-09-08

Researched and written by the R3KONX analysis desk from the cited primary material. Two caveats apply. The NIST draft is an initial public draft and its findings may change before final publication, and the draft's own publication page and NIST's announcement give different comment deadlines, both of which are published below. Figures on sniffing and injection success rates come from a 2025 research disclosure, not from the NIST testbed, and are labelled as such. Corrections to event@r3konx.asia.

R3KONX 2027

Work in this domain? So does the programme

4–6 May 2027, World Trade Centre Kuala Lumpur. Curated talks, hands-on training and The BattleGrid.