Several timing database
https://www.cnblogs.com/harrychinese/p/time_series_db.html
================================
be used as a timing database:
=========== =====================
[timing] TimescaleDB, based on PostgreSQL, supports SQL.
[timing] KairosDB, based on Cassandra, do not support SQL.
[GM] CrateDB, based on Elastic Search, but supports SQL ANSI
[timing] InfluxDB, is ranked first on the timing database db-engines, the latest version of the open source cluster function is not commercial version supports additional concurrent queries poor performance.
[GM] Kudu, column storage (class parquet), supports java API update data, more praise is to support upsert. SQL queries can be supported by impala or spark.
Simple Comments (based on the underlying technology to do a review, not to be a real test)
TimescaleDB based on PostgreSQL, may be suitable for a large amount of data that is not the case, but provide a rich SQL features
KairosDB, based on Cassandra, operation and maintenance should be relatively simple, scalability is also should be good, write performance is estimated CrateDB worse than the other does not support SQL.
CrateDB based Elastic Search, write performance should be good, should be good scalability is also estimated that SQL support and read performance will be worse, it supports full-text retrieval.
Contrast db-engines website:
https://db-engines.com/en/system/CrateDB%3BKairosDB%3BTimescaleDB
Compare Crate official:
http://go.cratedb.com/rs/832-QEZ-801/images/CrateDB-Cassandra-MongoDB-Comparison.pdf
================================
support SQL stream processing framework
============ ====================
most stream processing programs, the data temporarily stored in kafka usually in the format recommended Json / Avro, schema recommended Oracle Goldgate (OGG) data format.
SQL supports stream processing framework:
1. the Spark Streaming: can write complex SQL, and other databases such as DB do the Join.
2. Kafka's KSQL: Kafka and public cluster, no additional computing cluster.
3. PipelineDB: Based PostgreSQL extended, Cluster Edition be charged may be written directly to the data stream pipelinedb (temporary data stream in the form of FOREIGN TABLE pipelinedb), and then processed by pipelinedb SQL;. stream data may also hit the kafka, and then by pipelinedb extension to deal with.
================================
be used as a timing database:
=========== =====================
[timing] TimescaleDB, based on PostgreSQL, supports SQL.
[timing] KairosDB, based on Cassandra, do not support SQL.
[GM] CrateDB, based on Elastic Search, but supports SQL ANSI
[timing] InfluxDB, is ranked first on the timing database db-engines, the latest version of the open source cluster function is not commercial version supports additional concurrent queries poor performance.
[GM] Kudu, column storage (class parquet), supports java API update data, more praise is to support upsert. SQL queries can be supported by impala or spark.
Simple Comments (based on the underlying technology to do a review, not to be a real test)
TimescaleDB based on PostgreSQL, may be suitable for a large amount of data that is not the case, but provide a rich SQL features
KairosDB, based on Cassandra, operation and maintenance should be relatively simple, scalability is also should be good, write performance is estimated CrateDB worse than the other does not support SQL.
CrateDB based Elastic Search, write performance should be good, should be good scalability is also estimated that SQL support and read performance will be worse, it supports full-text retrieval.
Contrast db-engines website:
https://db-engines.com/en/system/CrateDB%3BKairosDB%3BTimescaleDB
Compare Crate official:
http://go.cratedb.com/rs/832-QEZ-801/images/CrateDB-Cassandra-MongoDB-Comparison.pdf
================================
support SQL stream processing framework
============ ====================
most stream processing programs, the data temporarily stored in kafka usually in the format recommended Json / Avro, schema recommended Oracle Goldgate (OGG) data format.
SQL supports stream processing framework:
1. the Spark Streaming: can write complex SQL, and other databases such as DB do the Join.
2. Kafka's KSQL: Kafka and public cluster, no additional computing cluster.
3. PipelineDB: Based PostgreSQL extended, Cluster Edition be charged may be written directly to the data stream pipelinedb (temporary data stream in the form of FOREIGN TABLE pipelinedb), and then processed by pipelinedb SQL;. stream data may also hit the kafka, and then by pipelinedb extension to deal with.