-
Notifications
You must be signed in to change notification settings - Fork 31
Jetstream Capabilities
shmurthy62 edited this page Feb 11, 2015
·
3 revisions
-
Java Application framework
-
Abstractions for composing stream processing components (Channels and Processors) and applications.
-
Cloud Friendly
-
Events in the form of unstructured data (key value pairs)
-
Restful monitoring of pipeline stages
- Easy to integrate with Ganglia or other systems like Ganglia.
-
Pipeline of channels and processors
- Channels are adaptors which adapt to external sources and sinks of data
- Processors are implementations of stream processors
- Supports dynamic pipeline changes at run time
- Wired at run time using Spring XML syntax
-
Stream Processing Pipeline
- Within application boundary
- Across application boundary seamlessly connected via cluster messaging
- Configuration driven
- Pipeline rewiring without application restart
-
Pluggable Default CEP Engine (Esper)
- SQL like language for specifying processing rules
- Analysis over rolling and tumbling windows of time
- Filtering and Joining streams
- Grouping and Ordering output
- Event Mutation
- Event declaration through Spring XML syntax
- EPL extensions to route events to specific stages in the pipeline, publish messages over event partitions and create dimensions.
-
Clustering
- Affinity and Round Robin event scheduling from producer to consumer
- Pluggable event schedulers on producer side
- consistent hashing and modulo algorithms for affinity event scheduling
- Weighted Round Robin and Weighted Random
- Automatic traffic rebalancing with faults and cluster resizing
- Auto discovery of cluster nodes
- Pluggable transports
-
**Hot deployment and run time compilation of SQL **