Techbanq

Telecom · BSS and OSS

At a telecom’s scale, the hard part is counting every event exactly once

Moving millions of messages is a solved problem; the difficulty is that each of those subscriber events also has to be rated on the right plan, billed once and reflected in the network, or the bill and the service drift apart. A pipeline that is fast but occasionally double-charges or drops an event under load turns a scale achievement into a revenue-assurance problem and a trust one.

Serving Saudi Arabia from our Riyadh office and our engineering hubs in Lahore and Karachi

Market
Saudi Arabia
Written for
CTO, COO
Decision
Keep billing, provisioning and the network correct at subscriber scale
Languages
EN · AR

Information security for telecom systems

  • ISO

    ISO/IEC 27001

    ISO/IEC 27001 certified

Certifications & registrations

What must hold

What a telecom’s operating systems have to hold true at scale

These are the guarantees a subscriber pipeline has to keep every day, not on a good day. Each is easy at low volume and only tested when the traffic is real.

Every event counted once

An idempotent pipeline where a retried or duplicated message is recognised and not billed twice or dropped.

What fails without it: Revenue leaks on one side and disputed double charges on the other.

Correct under peak load

The rating and charging path keeps its guarantees at the busy hour, not only in a quiet test.

What fails without it: A system that is correct until precisely the moment it is busiest.

Business and network agree

What the business system sold and what the network provisioned are reconciled, so an active plan is a working service.

What fails without it: A subscriber paying for what they cannot use, or using what they are not billed for.

History is reconstructable

Usage and change records are retained so a dispute or a regulator’s query can be answered with evidence.

What fails without it: A bill nobody can defend when it is challenged.

The real difficulty

Scale is not the problem; scale with correctness is

Handling volume, in the abstract, is a solved problem: queues, partitions and horizontal scale move a great many messages reliably. The difficulty in a telecom is that each of those millions of events also has to be exactly right — rated on the correct plan, billed once, and reflected in the network so the service a subscriber pays for is the service they actually get. A pipeline that is fast but occasionally double-charges, or that quietly drops an event when it is under pressure, converts a scale achievement into a revenue-assurance problem on one side and an angry-customer problem on the other.

That is why the interesting engineering sits in the business and network support layers rather than the app on the phone, which is the visible tenth. The order, the rating, the charging and the provisioning have to agree continuously at peak; and when they disagree — a plan sold but not provisioned, usage recorded but not rated — the system has to detect and reconcile the gap rather than let it surface as a complaint or a hole in the month-end numbers. We build for that reconciliation as much as for the flow, because at this scale the exceptions are not rare events; they are a steady rate that a design either handles or accumulates.

The pipeline

From a subscriber’s action to a bill that can be defended

The path each event takes, and the systems that carry each step. The branch is the reconciliation loop — the part that catches the disagreements this scale produces steadily.

  1. 01

    Capture the event once

    The network and OSS emit usage and signalling; capture each with an identity so a retry is not counted as a second event.

    • OSS
    • Mediation
  2. 02

    Rate and charge

    Apply the correct plan to the event exactly once, whatever the load at the moment it arrives.

    • Rating
    • Charging
    • BSS
  3. 03

    Provision and reflect

    Ensure what was sold is active in the network, so an activated plan is a service the subscriber can use.

    • Provisioning
    • Network inventory
  4. 04

    Bill and retain

    Produce the bill and keep the usage and change history a dispute or an audit will need.

    • Billing
    • Data retention

The reconciliation loop

Where the business and network systems disagree, a continuous check catches the gap and routes it to be resolved, because at this scale a mismatch left alone becomes either leaked revenue or a subscriber whose service does not match their bill.

The app is the easy tenth; the pipeline behind the bill is the job.

Our role

What we engineer and what stays with your network team

  1. 01

    Correctness at the pipeline

    We build the idempotent rating and charging path and the reconciliation that keeps the business and network systems honest, because that is where scale turns into either assurance or leakage.

  2. 02

    Integration, not the network

    We work with your OSS and network inventory rather than running the network; the network, the licences and the spectrum are yours, and we build the software layer that reads from and reconciles against them.

  3. 03

    A system your team scales

    We leave the pipeline, its guarantees and its runbooks documented and in your team’s hands, because a revenue-critical system only we can operate is a risk, not an improvement.

The network, the licences and the commercial plans stay entirely with you; we build the software that turns subscriber events into a bill and a service that agree, and that can be defended when either is questioned.

Common questions

What buyers ask

How do you stop revenue leakage between the network and billing?+

By making the two systems reconcile continuously rather than trusting that they match. Leakage happens in the gaps — usage the network recorded that the billing system never rated, a plan the business sold that the network never provisioned, an event counted twice or not at all under load. We build a reconciliation that compares what each system believes and surfaces every discrepancy as something to resolve, so the gap is caught while it is small rather than discovered as a shortfall at month-end. At this scale the mismatches are a steady rate, so the answer is a process that expects them, not an assumption that a clean pipeline will not produce any.

Can you modernise our billing without a risky migration of live subscribers?+

Yes, and for a live subscriber base that is the only responsible way. Rather than a single cutover that has to be right for every account at once, the new system runs alongside the old, rating and reconciling against it before it carries a single real bill. Subscribers are moved in controlled groups with the ability to fall back, and each group is proven correct against the incumbent system before the next moves. A billing migration that goes wrong is felt immediately by every affected customer, so the method is built to make each step reversible rather than to bet the base on one switchover night.

Where does the difficulty really sit — the customer app or the systems behind it?+

Almost entirely behind it. A subscriber app is a well-understood build, and a good one matters for experience, but it is the visible tenth of the system. The difficulty and the commercial risk are in the rating, charging, provisioning and reconciliation that have to be correct at scale — the parts a customer never sees until they go wrong, at which point they see a bill they dispute or a service that does not match what they bought. A telecom that invests in the app and neglects the pipeline behind it has polished the easy part and left the hard one to fail quietly at the busy hour.

Begin with your systems

A technical review of one revenue-critical flow — rating, charging or provisioning — that identifies where an event can be lost, double-counted or left unreconciled between the business and network systems, before any rebuild is scoped.

Start the conversation