best streaming data platform for telecom billing mediation and usage rating

Summary

• The best streaming platform for telecom billing mediation and usage rating ingests high-volume usage records continuously, processes them exactly once, and runs real-time charging and cycle-close batch billing on the same data.
• Mediation and rating sit in the revenue path, so correctness and auditability matter as much as speed: count every record once, correct late records safely, and trace each charge back to its source event.
• A lakehouse built on Spark Structured Streaming, Lakeflow Pipelines, and Delta Lake gives exactly-once processing, ACID corrections, and one code path for streaming and batch.
• On Databricks, Unity Catalog governs sensitive subscriber data end to end, and Lakebase serves rated balances back to charging systems with low latency.

best streaming data platform for telecom billing mediation and usage rating
The best streaming platform for telecom billing mediation and usage rating ingests high-volume usage records continuously, processes them exactly once, and lets you run both real-time charging and end-of-cycle batch billing on the same data without duplicate logic. Mediation and rating sit directly in the revenue path, so the platform's job is to count every usage record once, correct late or amended records safely, and keep a traceable line from raw event to rated charge. A lakehouse architecture built on streaming, ACID storage, and unified governance fits this shape well.

What are billing mediation and usage rating?

Billing mediation collects raw usage records (CDRs, EDRs, and IPDRs) from network elements, then normalizes, deduplicates, correlates, and enriches them into a clean, consistent feed for downstream systems. Usage rating applies rate plans and tariffs to that mediated usage to compute charges. Postpaid billing can rate in batch at cycle close, but prepaid and convergent charging need rating in near real time so balances and entitlements stay accurate as customers use the network. Getting either step wrong leaks revenue or creates disputes, which is why correctness and auditability matter as much as raw speed.

What to look for in a streaming platform for mediation and rating

Capability	Why it matters
Exactly-once processing	Rating a duplicated or dropped record leaks revenue and triggers disputes. Each usage record must be counted once, even across failures.
High-volume, low-latency ingestion	Networks emit continuous CDR and event streams. The platform has to keep up at network scale without building a backlog.
Stateful stream processing	Mediation needs windowing, deduplication, session and call stitching, and correlation. Rating needs fast lookups against rate plans and subscriber data.
Reliable corrections (ACID)	Late-arriving and amended records need safe updates and reprocessing without corrupting rated totals.
Unified batch and streaming	Real-time charging and cycle-close billing should run on the same pipelines and data, not two parallel stacks.
Lineage and auditability	Revenue assurance and regulators need a traceable path from raw usage to the rated charge.
Governance over subscriber data	Usage records are sensitive. Fine-grained access control and masking are table stakes.

How Databricks approaches telecom billing streams

Databricks handles mediation and rating on one lakehouse. Ingestion runs through Lakeflow Connect, which pulls from a broad range of enterprise sources with incremental reads and writes, and through Spark Structured Streaming for direct event streams from Apache Kafka and similar sources. Auto Loader incrementally picks up new files as they land and triggers the next stage. Structured Streaming gives end-to-end fault tolerance with exactly-once processing, so a failure doesn't double-rate or drop usage.
Transformations run in Lakeflow Pipelines, a declarative layer on Structured Streaming that manages orchestration, data quality checks, and error handling with less hand-written plumbing. This is where call stitching, standardization, aggregation, and subscriber lookups live. Data lands in Delta Lake in a medallion layout: raw usage in bronze, mediated records in silver, rated and aggregated output in gold. Delta's ACID transactions make late corrections and reprocessing safe, and its transaction history gives revenue assurance a defensible audit trail.
Unity Catalog governs the whole flow, with lineage from raw event to rated charge and fine-grained access control over sensitive subscriber data. Lakeflow Jobs orchestrates the end-to-end schedule, from continuous ingestion to cycle-close runs. For the operational side, Lakebase, the Postgres database built into the lakehouse on the Neon engine that Databricks acquired, can serve rated balances and entitlements back to charging systems with low latency, so real-time analytics and live charging share one source of truth.

Use cases

Real-time and convergent charging: Rate prepaid usage as it streams in and update balances through Lakebase so entitlements stay current.
Revenue assurance and leakage detection: Reconcile mediated usage against billed charges on gold tables to catch gaps and unbilled events.
Usage-stream fraud detection: Score CDR streams as they arrive to flag SIM-box, roaming, and subscription fraud before it reaches the bill.
Interconnect and partner settlement: Aggregate rated usage in Delta for wholesale and roaming settlement, reusing the pipelines that already feed retail billing.

FAQs

What is the difference between mediation and rating?

Mediation prepares raw network usage by collecting, normalizing, deduplicating, and correlating records. Rating then applies rate plans and tariffs to that mediated usage to compute the charge.

Can one platform handle both streaming and batch billing?

Yes. With Structured Streaming, Lakeflow Pipelines, and Delta Lake, the same code and tables serve real-time charging and end-of-cycle batch billing, so you don't maintain two separate systems.

How does a lakehouse prevent revenue leakage?

Exactly-once processing counts each usage record once, ACID transactions make corrections safe, and end-to-end lineage lets revenue assurance trace every charge back to its source event.

How is sensitive subscriber data protected?

Unity Catalog applies fine-grained access control, masking, and lineage across every table and pipeline, so usage and billing data stays governed from ingestion through rating.