6.6 데이터 처리방법 비교#
Comparison#
구분 |
Apache storm |
Apache Storm Trident |
Apache Flink |
Apache Kafka |
Apache Kafka Streams |
Spark streaming |
Samza |
---|---|---|---|---|---|---|---|
Streaming model |
native |
micro-batching |
micro-batching |
native |
native |
micro-batching |
native |
API |
Compositional |
Compositional |
Declarative |
Declarative |
Declarative |
Declarative |
Compositional |
Guarantee |
At-least-once |
Exactly-once |
Exactly-once |
Exactly-once |
Exactly-once |
Exactly-once |
At-least-once |
Fault Tolerance |
Record ACKs |
Record ACKs |
Checkpointing |
Log-based |
Log-based |
RDD based Checkpointing |
Log-based(kafka 활용) |
State Management |
Not build-in |
Dedicated Operators |
Stateful Operators |
Not build-in |
Stateful Operators |
Dedicated Dstream |
Stateful Operators |
Latency |
Very low |
Medium |
Low |
Low |
Low |
Medium |
Low |
Throughput |
Low |
Medium |
Low |
Medium |
Medium |
High |
Medium |
Maturity |
High |
High |
Low |
High |
High |
High |
Medium |