Need advice about which tool to choose?Ask the StackShare community!
Apache Flink vs Druid: What are the differences?
Introduction
This Markdown code provides a comparison between Apache Flink and Druid, highlighting the key differences between these two technologies.
Processing Model: Apache Flink is a stream processing technology that provides both batch and stream processing capabilities. It supports event-time and processing-time semantics, and allows for stateful operations, windowing, and event-time processing. On the other hand, Druid is a real-time analytical database designed to handle high-throughput, low-latency querying of time-series data. It relies on a columnar data store and is optimized for OLAP style queries.
Data Ingestion: Apache Flink supports various data sources and connectors, allowing data to be ingested from multiple systems. It provides built-in connectors for Kafka, Hadoop file systems, and more. Flink also supports custom sources and sinks, making it flexible for different use cases. Druid, on the other hand, relies on a specific data ingestion architecture where data is ingested through real-time and batch ingestion processes. It provides connectors for popular data sources like Kafka, Hadoop, and more.
Query Capabilities: Apache Flink provides a rich set of operators and functions for processing and analyzing data. It supports SQL queries, batch and stream processing, and complex event processing. Flink also allows for iterative processing and machine learning with its built-in libraries. Druid, on the other hand, focuses on OLAP-style queries for time-series data. It is optimized for fast aggregations and filtering on large datasets, making it suitable for real-time analytics.
Scalability and Fault-tolerance: Apache Flink is designed to scale horizontally and can handle large amounts of data. It provides fault-tolerance through its distributed and reliable streaming architecture, allowing for high data availability and resilience in the face of failures. Druid is also designed to scale horizontally and can handle large datasets. It provides fault-tolerance through data replication and distributed query processing, ensuring high availability and reliability.
Data Storage and Indexing: Apache Flink does not have its own storage layer and can work with various storage systems like Hadoop Distributed File System (HDFS) or Amazon S3. It does not provide indexing capabilities out of the box. Druid, on the other hand, has its own columnar storage format and indexing capabilities built-in. It uses inverted indexes and bitmap indexes to optimize queries and speed up data retrieval.
Use Cases: Apache Flink is commonly used for real-time stream processing, batch processing, and building complex event processing applications. It is widely used in industries like e-commerce, finance, and telecommunications. Druid, on the other hand, is commonly used for real-time analytics and powering interactive dashboards. It is well-suited for use cases like monitoring, ad tech, and IoT analytics.
In Summary, Apache Flink is a flexible stream processing technology with batch processing capabilities, while Druid is a real-time analytical database optimized for time-series data querying. Flink provides more general-purpose processing capabilities, while Druid is specialized for OLAP-style queries on large datasets.
We have a Kafka topic having events of type A and type B. We need to perform an inner join on both type of events using some common field (primary-key). The joined events to be inserted in Elasticsearch.
In usual cases, type A and type B events (with same key) observed to be close upto 15 minutes. But in some cases they may be far from each other, lets say 6 hours. Sometimes event of either of the types never come.
In all cases, we should be able to find joined events instantly after they are joined and not-joined events within 15 minutes.
The first solution that came to me is to use upsert to update ElasticSearch:
- Use the primary-key as ES document id
- Upsert the records to ES as soon as you receive them. As you are using upsert, the 2nd record of the same primary-key will not overwrite the 1st one, but will be merged with it.
Cons: The load on ES will be higher, due to upsert.
To use Flink:
- Create a KeyedDataStream by the primary-key
- In the ProcessFunction, save the first record in a State. At the same time, create a Timer for 15 minutes in the future
- When the 2nd record comes, read the 1st record from the State, merge those two, and send out the result, and clear the State and the Timer if it has not fired
- When the Timer fires, read the 1st record from the State and send out as the output record.
- Have a 2nd Timer of 6 hours (or more) if you are not using Windowing to clean up the State
Pro: if you have already having Flink ingesting this stream. Otherwise, I would just go with the 1st solution.
Please refer "Structured Streaming" feature of Spark. Refer "Stream - Stream Join" at https://spark.apache.org/docs/latest/structured-streaming-programming-guide.html#stream-stream-joins . In short you need to specify "Define watermark delays on both inputs" and "Define a constraint on time across the two inputs"
Pros of Druid
- Real Time Aggregations15
- Batch and Real-Time Ingestion6
- OLAP5
- OLAP + OLTP3
- Combining stream and historical analytics2
- OLTP1
Pros of Apache Flink
- Unified batch and stream processing16
- Easy to use streaming apis8
- Out-of-the box connector to kinesis,s3,hdfs8
- Open Source4
- Low latency2
Sign up to add or upvote prosMake informed product decisions
Cons of Druid
- Limited sql support3
- Joins are not supported well2
- Complexity1
Cons of Apache Flink
Sign up to add or upvote consMake informed product decisions
What is Druid?
What is Apache Flink?
Need advice about which tool to choose?Ask the StackShare community!
What companies use Druid?
What companies use Apache Flink?
Sign up to get full access to all the companiesMake informed product decisions
What tools integrate with Apache Flink?
Sign up to get full access to all the tool integrationsMake informed product decisions
Blog Posts
+3
+6
+4
+7
+8+6
+12