Keep trunks clean, stop fraud before it hits your backbone#
You sell capacity and reliability. The fastest way to torch both is letting junk flow across your trunks: robodial bursts, call pumping to high-rate destinations, wangiri callbacks, PBX compromises hammering premium prefixes. Costs spike. ASR tanks. Disputes land on your desk.You can't babysit every customer's dialer. You can police what enters your network.Where CallerAPI sits
Right at your SBC / trunk edge, pre-routing, pre-billing. For inbound, pre-present to enterprise trunks. For outbound, pre-origination. You ask one question: "Do we accept this call on this trunk?" We return a tier (SPAM / BUSINESS / UNKNOWN) and the why (raw data, reason codes). You enforce with SIP policy—reject, throttle, or pass. That’s it.
What we're actually fixing#
Cost leakage: block call pumping and IRSF before a single second lands in CDRs.
Quality: junk drags ASR/NER down; keep the graph clean for your paying customers.
Ops sanity: fewer disputes, clearer RCAs, fewer "your trunks are compromised" tickets.
How you run it (boring and predictable)#
| Case | Default action | What actually happens | Why this is sane |
|---|
| SPAM | Reject | Deny at edge (e.g., 607/403), log reason, short temp-block for repeats. | Kills obvious fraud and pumping fast; protects high-cost routes. |
| BUSINESS, but with recent reports | Throttle / Step-up | Cap CPS/concurrent, require verified CLI, tighten geo/prefix rules. | Legit traffic survives with friction; abusers give up. |
| UNKNOWN | Allow | Route normally, monitor early metrics (ACD, ASR, CPS). | Most new senders are fine. |
Start conservative. Put the gray zone in Throttle/Step-up, not hard Reject. Tighten after a week of data.
Day-1 to Day-30 - the playbook#
1
Week 1 - Pilot and test
Score all inbound/outbound attempts at the SBC. Don't enforce yet. Pipe tiers + reasons to NOC. Build two live lists: must-pass (emergency, verified enterprise ranges) and must-block (known abusers, toxic prefixes).
2
Week 2 - Turn on enforcement
Enforce SPAM = Reject on high-risk international prefixes and newly onboarded trunks. Keep BUSINESS (with reports) = Throttle/Step-up (CPS caps, concurrent caps, verified CLI). Expose reason strings to support.
3
Week 3 - Tune by trunk group & destination
Different customers and geos behave differently. Tighten on premium destinations; loosen for authenticated enterprise with stable patterns. If throttle pass-through is high, you’re being too strict.
4
Week 4 - Automate hygiene
Auto-escalate repeat gray offenders to verification. Auto-extend temp blocks for SPAM repeats. Keep signed decision logs - dispute enders.
What to measure (and how to read it)#
CPS spikes / concurrent channels - spiky = bad; throttle until it smooths.
ASR / NER - should improve as noise drops. If flat, widen SPAM.
ACD to high-rate prefixes - pumping is short. Rising ACD = cleaner routes.
Chargebacks / disputes - should fall. If not, expose reasons earlier in support.
Quick win: enforce on international high-rate prefixes and newly onboarded trunks first. That’s where the money leaks.
Edge cases you will hit#
Cost & Data Control - "Daily Spam Report" for flat spend#
Not every trunking provider wants per-attempt checks. Some want predictable cost and a simple ingest: one pull a day, curated intelligence, trunk-wide.
What you get
A daily set of confirmed spam/abuse numbers and campaigns, validated by our systems + human review. Tenant-scoped dataset for your edge.
Why it exists
Protect premium routes without variable bills. One (or a few) fetches per day. Load it. Enforce. Done.
How you use it
Pull today's JSON, push to SBC caches/ACLs, apply the same SPAM / BUSINESS / UNKNOWN policy. Curated list; no per-attempt metering.Endpoint: dedicated Daily Spam Report per tenant.
Fields: number, created date, violated date, consumer state, subject (spam category), robocall (yes/no).
Cutover time: posted before 13:00 UTC daily, covers prior 24h.
Validation: algorithmic scoring + human review on high-impact entries.
Retention: last 24h available (30-day window on request).
License: trunk-wide use at your edge.
Pricing: flat monthly fee (no per-event surprises).
Purpose: cost control with enough fidelity to keep trunks clean.
Ask us to enable Daily Spam Report for your tenant. We'll hand you a hardened URL, auth, and a sample file so your team can wire the nightly job.
You keep control of region and retention on your side.
Daily Report - budget certainty; great for global hygiene and premium-route protection.
Real-time - when you need per-attempt context (trunk group, destination, CPS profile) and want to throttle the gray zone instead of blunt-blocking it.
How to talk about this with customers#
Be straight: "We protect the trunk edge. We block obvious fraud, throttle the gray, and we log every decision with reason + timestamp." Publish reason codes in your portal. Angry tickets turn into quick fixes. Your upstream carriers notice the improvement.
Talk to an expert
Make your telco spam-free by integrating CallerAPI.