Ad tech has to make decisions in milliseconds and report on billions of events. The platform uses a separate store for each job. These are the components named on the technology page.
The data layer
| Component | Role |
|---|---|
| Aerospike | Real-time data used during bidding |
| Redis | In-memory store for fast lookups |
| ClickHouse | Analytics at billions of rows, behind the reports |
| Apache Kafka | Real-time event streaming between services |
Why it is built this way
- Bidding needs answers fast, so it reads from in-memory and real-time stores.
- Reporting needs to scan huge volumes, so events stream through Kafka into an analytics database.
- Keeping these apart means heavy reports do not slow down live bidding.
The application layer
- Java and Spring on the back end
- React for the control panels
- Node.js and Python services where they fit
These are proven languages and frameworks your developers already know, so the code is easy to read, change and extend. With the source-code option you get all of it.
Want the details?
The free white papers cover the architecture of each platform: bidder engine, real-time storage, log streaming and analytics databases.
Related: Development and AI technology · How scalable is the platform?
Nov 13, 2026