How Nussknacker runs
Running Nussknacker involves three related choices: the processing mode, the engine, and the deployment target.
Together, they describe how a scenario processes data, what technology executes it, and where it runs.
| Answers | Examples | |
|---|---|---|
| Processing mode | How does the scenario process data? | Streaming, Request-Response |
| Engine | What technology executes the scenario? | Lite, Flink |
| Deployment target | Where and how is the engine run? | Designer, Kubernetes, a Flink cluster, a managed Flink service |
Batch processing is planned and will extend the set of available processing modes in the future.
These dimensions are related, but they describe different aspects of running Nussknacker. A processing mode determines the way a scenario interacts with incoming and outgoing data. The engine provides the runtime technology needed to execute that scenario. A deployment target is a configured destination where a scenario using that engine can be deployed.
Not every combination is possible. For example, Request-response scenarios use the Lite engine, while Streaming scenarios can use Lite or Flink. Each engine in turn has its own possible deployment targets.
A single Designer can work with multiple deployment targets, so scenarios using the same engine do not all have to run in the same place.
Where Designer runs
Designer is the authoring and management part of Nussknacker. It can run in Nussknacker Cloud or be operated by the customer, either on-premises or in the customer's cloud infrastructure.
Where Designer runs is independent of where scenarios are executed. A single Designer can deploy scenarios to multiple deployment targets running in different locations and operated by different parties. For example, Designer running in Nussknacker Cloud can deploy scenarios to a deployment target running entirely in the customer's infrastructure.
Designer itself typically has relatively modest resource requirements because scenario data is processed on deployment targets rather than by Designer. The exception is an embedded deployment target, where the Lite or Flink engine runs directly inside Designer; in that case Designer also performs the scenario processing.
Learn more:
- Processing Modes - how scenarios process data.
- Engines - the runtime technologies used to execute scenarios.
- Deployment Targets - where and how scenarios can be run.