Almost every impersonation scam starts with a call or a text that shows a number or sender name the victim trusts, and the networks that carry those messages are the only parties who can stop spoofed traffic at scale. Regulators have responded by turning scam prevention into network obligations: caller ID authentication and robocall mitigation in the United States, numbering misuse rules and know your customer duties in the United Kingdom. This guide is for fraud, security and regulatory teams at network operators, wholesale carriers and messaging aggregators, and for enterprises that rent numbers or send bulk SMS and need to know what their providers must do. It complements the Vishing and Smishing Guide, which covers how organisations protect their own staff and customers, and does not repeat it.
Key takeaways
- Caller ID authentication proves origin, not intent. STIR/SHAKEN lets the originating carrier sign the caller ID with an attestation level of A, B or C (RFC 8588), so a signed call can still be a scam from a fully attested number the scammer rented.
- US rules are detailed and dated. 47 CFR Part 64 Subpart HH required STIR/SHAKEN in IP networks by 30 June 2021, gives gateway and intermediate providers their own deadlines, and requires every provider to run a robocall mitigation programme and answer traceback requests within 24 hours.
- Blocking is explicitly permitted, with safeguards. 47 CFR 64.1200(k) lets voice providers block invalid, unallocated and unused numbers and use reasonable analytics, and 64.1200(p) requires mobile providers to block texts from numbers on a reasonable do-not-originate list.
- Number leasing is the weak point regulators are targeting. Ofcom fined Tismi £437,500 in October 2026 for sub-allocating numbers without adequate due diligence or misuse monitoring, exposing about 366,000 numbers to potential misuse.
- SMS rules in the UK are still being set. Ofcom's consultation on combatting mobile messaging scams closed on 28 January 2026 and Ofcom stated it planned a final decision in summer 2026; check the current status before relying on any specific obligation.
Where scams enter the network
| Entry point | Scam pattern | Primary network control |
|---|---|---|
| International gateway | Calls from abroad presenting domestic numbers, often impersonating banks, police or government | Gateway attestation, blocking domestic caller IDs arriving on international routes, traceback |
| Wholesale and VoIP origination | Rented numbers and SIP trunks used for high-volume impersonation calls | Customer due diligence, traffic monitoring, robocall mitigation programme |
| Sub-allocated number ranges | Numbers passed down a reseller chain until no one knows the end user | Know your customer at every layer, misuse reporting and blocking |
| Business messaging aggregators (A2P SMS) | Smishing from alphanumeric sender IDs imitating brands | Sender ID registration and protection, content and URL filtering, aggregator vetting |
| Person-to-person SMS from SIM farms | Bulk smishing from many prepaid SIMs | Volume and velocity rules, SIM box detection, rapid number suspension |
| Account and SIM servicing | Port-out and SIM swap to intercept one-time codes | Strong authentication for account changes. See the SIM Swap Fraud Guide. |
Attackers use these channels for ATT&CK T1566.004 Spearphishing Voice in enterprise attacks and for mobile T1660 Phishing, but the network operator sees them as traffic patterns, not techniques. Controls work best when they combine origin evidence (signatures, route), customer evidence (who rented the number) and behaviour (volumes, call duration, complaint rates).
Caller ID authentication with STIR/SHAKEN
STIR/SHAKEN is a set of IETF and industry standards in which the originating provider signs the caller ID in a SIP call using a PASSporT token, and downstream providers verify the signature. RFC 8588, an IETF Standards Track document, defines the SHAKEN extension and its two key claims.
| Attestation | Meaning (RFC 8588) | How a terminating provider should treat it |
|---|---|---|
| A, full | The provider can fully attest to the calling identity: it knows the customer and that the customer may use the number | Strong origin evidence; still apply analytics, because a scammer can be a known customer |
| B, partial | The provider originated the call but cannot fully attest to the calling identity | Neutral; combine with route and behaviour |
| C, gateway | The provider received the call from a gateway that does not support PASSporT; the lowest level | Weak; typical of international and legacy traffic, so domestic caller IDs on C should be scrutinised |
| No signature | Call passed through a non-IP or non-compliant segment | Treat as unverified |
origid: RFC 8588 also defines an opaque per-call identifier for the network element where the call started, which supports traceback. It recommends a unique value per call so that observers cannot correlate calls to a network element.
US obligations under 47 CFR Part 64 Subpart HH
- Section 64.6301: voice service providers had to implement STIR/SHAKEN in their IP networks by 30 June 2021, subject to extensions and exemptions, holding an SPC token and STI certificate.
- Section 64.6302: intermediate providers share the 30 June 2021 deadline, and gateway providers must authenticate caller ID for calls that use US North American Numbering Plan numbers in the caller ID field.
- Section 64.6303, non-IP networks: deadlines of 30 June 2021 for voice service providers, 30 June 2023 for gateway providers and 31 December 2023 for non-gateway intermediate providers receiving calls directly from an originating provider; providers either upgrade to SIP with STIR/SHAKEN or document participation in a non-IP standards effort.
- Section 64.6305, robocall mitigation: every provider needs a robocall mitigation programme including a commitment to respond fully to traceback requests within 24 hours, must file in the Robocall Mitigation Database and recertify annually on or before 1 March. Since 11 April 2023 providers may accept calls directly from a gateway provider only if that provider is listed and not de-listed.
Call and text blocking rules
Blocking stops traffic that authentication cannot: calls from numbers that should never originate traffic and patterns that analytics show to be highly likely illegal. In the US, 47 CFR 64.1200 sets what is allowed.
- Number-based voice blocking, 64.1200(k)(1) and (2): providers may block calls purporting to come from invalid numbers, numbers not allocated to any provider, allocated but unused numbers, inbound-only numbers whose subscriber asked for blocking, and numbers that reasonable analytics show are highly likely tied to one-ring scams.
- Analytics-based blocking, (k)(3): terminating providers may block using reasonable analytics if they consider caller ID authentication information, let consumers opt out, apply analytics without discrimination and charge no line-item fee.
- Blocking without opt-out, (k)(11): permitted only for patterns highly likely to be illegal, with human oversight, ending once calls are known to be likely lawful, and with disclosure.
- Safeguards: never block 911 calls and make reasonable efforts not to block public safety calls; publish a single complaint contact, give status updates within 24 hours, cease blocking after a credible erroneous-blocking claim, return an appropriate response code such as SIP 603+, and provide a subscriber's list of blocked calls within 3 business days on request.
- Text blocking, 64.1200(p): mobile wireless providers must block texts purporting to come from a North American number on a reasonable do-not-originate list, which may hold subscriber-requested, invalid, unallocated and allocated-but-unused numbers; a list that leaves out obvious numbers may be deemed unreasonable.
Do-not-originate lists for enterprises: organisations with inbound-only numbers, such as bank helplines printed on cards, can ask their provider to place those numbers on a do-not-originate list so any outbound call or text claiming to come from them is blocked. This is one of the cheapest anti-impersonation controls available.
UK approach: numbering misuse, KYC and messaging
The UK relies on Ofcom's General Conditions, numbering rules and guidance rather than a single caller ID authentication mandate.
- Numbering due diligence: Ofcom's October 2026 decision against Tismi cited General Conditions B1.6, B1.8 and B1.9(c) and its Good Practice Guide on sub-allocated and assigned numbers. The failings ran from November 2022 to January 2025: sub-allocating numbers without adequate due diligence, no monitoring for misuse, and no documented process for misuse reports. Ofcom directed number blocking in June 2025 and the £437,500 fine included a 30% discount for admitting liability.
- Further investigations: Ofcom announced investigations on 24 September 2026 into whether two telecoms companies failed to prevent misuse of their numbers, and has another investigation open into possible number misuse. Treat numbering KYC as an active enforcement priority.
- Mobile messaging: the consultation on new rules and guidance for mobile operators and business messaging aggregators aims both to prevent scammers accessing messaging services and to stop them once they have access, drawing on existing sector practice. When checked on 27 October 2026 the page still showed "Pending Statement".
- Data sharing: Ofcom also publishes material on data sharing between providers to combat fraud; sharing confirmed scam numbers and sender IDs is now an expected practice.
For the wider telecoms security regime see the Telecoms Security Laws Map and the UK Telecommunications Security Act Guide.
SMS filtering and sender ID protection
Smishing is harder to authenticate than voice: there is no widely deployed signature for SMS, and alphanumeric sender IDs can be set by whoever submits the message. Controls therefore sit with the operator firewall and the aggregators that connect businesses to networks.
| Control | What it does | Owner | Limitation |
|---|---|---|---|
| Aggregator onboarding and KYC | Verifies the business sending traffic and its right to use a brand name | Aggregator, operator contract terms | Reseller chains dilute checks |
| Sender ID protection lists | Blocks protected brand sender IDs unless they arrive on authorised routes | Operators, industry registries | Scammers switch to look-alike names or long numbers |
| SMS firewall content and URL filtering | Blocks known scam URLs, new domains and templates | Mobile operator | Must handle privacy and false positives; scammers vary text |
| Velocity and SIM farm detection | Flags bulk P2P traffic from many SIMs or one location | Mobile operator | Legitimate group messaging can trigger it |
| Customer reporting loop | Customers forward scam texts (7726 in the UK) to feed filters | Operators | Depends on reporting volume and speed of action |
| Do-not-originate for texts | Blocks texts from numbers that never send them (US 64.1200(p)) | Mobile provider | Covers numbers only, not sender names |
Enterprises that send SMS should register their sender IDs with every route they use, never send links in one-time code messages, and use a consistent short domain so customers can learn what genuine messages look like. Artificially inflated traffic is a separate fraud against SMS senders; see the Phone Number Fraud guide.
Choosing controls by role
- Large mobile and fixed operators: sign and verify all IP traffic, analytics-based blocking with an appeals process, SMS firewall, sender ID protection, and data sharing with banks and other operators.
- Wholesale carriers and gateway providers: the highest-value control is refusing or terminating customers who generate scam traffic; customer due diligence, traffic profiling and fast traceback response matter more than consumer features.
- Smaller providers and resellers: focus on KYC for every number sub-allocated, monitoring complaints per customer and a documented misuse process, which is exactly where the Tismi failings lay.
- Enterprises: do-not-originate registration for inbound-only numbers, sender ID registration, contractual KYC and misuse clauses with telephony and messaging suppliers, and an internal route for staff to report spoofing of company numbers. See the Vishing and Help Desk Social Engineering Awareness Guide.
How to implement
- Map traffic sources: list every interconnect, gateway, wholesale customer and messaging aggregator, with the share of complaints each generates.
- Fix the customer layer: introduce KYC for every customer that receives numbers or sends A2P traffic, including resellers, and require them to pass checks down the chain.
- Authenticate: in jurisdictions with STIR/SHAKEN, sign all originated IP calls with the correct attestation and verify all inbound signatures; record attestation in call records.
- Block in monitor mode first: run number-based and analytics rules in log-only mode for several weeks, review samples by hand, then enable blocking for the highest-confidence rules.
- Publish redress: stand up the complaint contact, response codes and unblock process before blocking at scale, so false positives can be reversed within the required time.
- Close the loop: feed customer reports, bank fraud data and traceback results into rules; suspend offending customers and numbers promptly.
- Keep rollback: every blocking rule should be versioned and individually switchable, so a rule that blocks a hospital or government line can be removed in minutes.
Detection and measurement
- Attestation coverage: share of originated and terminated calls carrying a valid signature, by attestation level.
- Scam complaint rate per customer: complaints or 7726-style reports per million calls or messages for each wholesale customer and aggregator.
- Traceback response time: time to respond fully, against the 24-hour US requirement.
- Time to suspend: median time from first credible report to blocking a number or sender ID.
- False blocking rate: successful erroneous-blocking claims per million blocked calls or texts, and time to unblock.
- Sub-allocation KYC coverage: share of sub-allocated numbers with a verified end user on record.
Framework mapping
| Control | NIST CSF 2.0 | NIST SP 800-53 Rev. 5 | Regulatory anchor |
|---|---|---|---|
| Caller ID signing and verification | PR.DS-02 | SC-8 Transmission Confidentiality and Integrity | 47 CFR 64.6301 to 64.6303; RFC 8588 |
| Traffic monitoring and analytics | DE.CM-01 | SI-4 System Monitoring | 47 CFR 64.1200(k)(3) |
| Message filtering | DE.CM-01 | SI-8 Spam Protection | 47 CFR 64.1200(p); Ofcom messaging consultation |
| Customer and reseller due diligence | GV.SC-07 | SA-9 External System Services, SR-6 Supplier Assessments | Ofcom GC B1, Good Practice Guide |
| Traceback and data sharing | RS.CO-03 | IR-4 Incident Handling, IR-6 Incident Reporting | 47 CFR 64.6305 |
Common failure patterns
- Treating attestation A as "safe": scammers rent numbers from providers that know them as customers. Signatures prove origin, analytics decide intent.
- KYC only at the first layer: numbers sub-allocated through resellers with no end-user checks, the core failing in the Tismi case.
- No misuse process: reports arrive at a generic inbox with no owner, so the same numbers keep calling for weeks.
- Blocking without redress: aggressive rules block legitimate callers such as clinics, then get switched off entirely after complaints.
- Sender ID protection on some routes only: scammers simply buy the unprotected route.
- Enterprises never registering their numbers: inbound-only bank or help desk numbers stay spoofable because no one asked for do-not-originate treatment.
Practitioner checklist
- Inventory interconnects, gateways, wholesale customers and aggregators with complaint rates for each.
- Require documented KYC for every customer and reseller that receives numbers or sends A2P traffic.
- Sign and verify all IP calls and record the attestation level in call records.
- Scrutinise domestic caller IDs arriving with gateway attestation or on international routes.
- Run blocking rules in monitor mode, then enable with a published complaint and unblock process.
- Maintain do-not-originate lists for voice and text, and offer them to enterprise customers.
- Answer traceback requests inside 24 hours and measure it.
- Protect sender IDs on every route and filter scam URLs at the SMS firewall.
- Track Ofcom's messaging scams statement and update obligations when it is published.
Standards and references
- eCFR: 47 CFR Part 64 Subpart HH, Caller ID Authentication
- eCFR: 47 CFR 64.1200, Delivery restrictions
- IETF: RFC 8588, PASSporT Extension for SHAKEN
- FCC: Combating Spoofed Robocalls with Caller ID Authentication
- Ofcom: Telecoms provider fined £437,500 for failing to prevent misuse of its phone numbers
- Ofcom: Combatting mobile messaging scams
- Ofcom: Scam calls and messages
- NIST: SP 800-53 Rev. 5
Related NHI Mgmt Group resources: Vishing and Smishing Guide · SIM Swap Fraud Guide · Telecom Network Security Guide · Impersonation Scam Awareness Guide