Engines comparison
A bit of history
Nussknacker was originally built as Flink job's authoring tool. Although it was great fit for complex, stateful scenarios but also simple ones, we noticed, that sometimes maintenance of Flink can be painful. Furthermore, Flink often requires a lot of resources, which can be costly. We have introduced second streaming engine called Streaming-Lite which delivers most of Nussknacker's favours, but at the same time offers cheap and lightful setup.
When Streaming-Flink?
- You need to implement decision algorithms, even highly branched
- Your business logic requires data aggregation
- You want to match against event patterns
- You have already running Flink cluster
When Streaming-Lite?
- You need to implement decision algorithms, even highly branched
- Lightweight infrastructure is relevant for you
- You have already running K8 cluster
Features
| Streaming-Flink | Streaming-Lite | |
|---|---|---|
| real-time streaming | ✔️ | ✔️ |
| scalable processing (both horizontally/vertically) | ✔️ | ✔️ |
| high availability | ✔️ | ✔️ |
| high throughput | ✔️ | ✔️ |
| low latency | ✔️ | ✔️ |
| resilience | ✔️ | ✔️ |
| metrics | ✔️ | ✔️ |
| stateful processing | ✔️ | ❌ |
| ease of operations | medium | easy |
| environment setup difficulty | medium | easy |