[
https://issues.apache.org/jira/browse/FLINK-22994?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
lynn1.zhang updated FLINK-22994:
--------------------------------
Description:
h1. BackGround
Flink maintain the udf result as binary, like BinaryStringData. When invoking
nesting udf like select useless(int_ip_2_string(ip)), the result of
int_ip_2_string(ip) will be serialization and deserialization.
Below is the Generated Code
!image-2021-06-15-15-18-12-619.png! this issue want to improve it as below
!image-2021-06-15-15-19-01-103.png!
h1. Performance Compare
Condition: Source = Kafka，Schema = PB with snappy；Flink Slot =
1、taskmanager.memory.process.size=4g; Linux Core = Intel(R) Xeon(R) Gold 5218
CPU @ 2.30GHz
UDF Introduction:
* ipip: input: int ip, output: map ip_info, map size = 14.
* ip_2_country: input map ip_info, output: string country.
* ip_2_region: input map ip_info, output: string region.
* ip_2_isp_domain: input map ip_info, output: string isp.
* ip_2_timezone: input map ip_info, output: string timezone.
h5. The throughput without udf invoke: 764.50 w/s
!image-2021-06-15-15-27-26-739.png!
h5. The throughput with udf invoke: 183.24 k/s
!image-2021-06-15-15-28-28-137.png!
h5. The throughput with udf nesting invoke: 41.42 k/s
!image-2021-06-15-15-29-09-773.png!
h5. The throughput with udf nesting invoke after the issue: 174.41 k/s
!image-2021-06-15-15-30-14-775.png!
was:
h1. BackGround
Flink maintain the udf result as binary, like BinaryStringData. When invoking
nesting udf like select useless(int_ip_2_string(ip)), the result of
int_ip_2_string(ip) will be serialization and deserialization.
Below is the Generated Code
!image-2021-06-15-15-18-12-619.png! this issue want to improve it as below
!image-2021-06-15-15-19-01-103.png!
h1. Performance Compare
Condition: Source = Kafka，Schema = PB with snappy；Flink Slot =
1、taskmanager.memory.process.size=4g; Linux Core = Intel(R) Xeon(R) Gold 5218
CPU @ 2.30GHz
UDF Introduction:
* ipip: input: int ip，output: map ip_info，map size = 14。
* ip_2_country: input map ip_info，output: string country。
* ip_2_region: input map ip_info，output: string region。
* ip_2_isp_domain: input map ip_info，output: string isp。
* ip_2_timezone: input map ip_info，output: string timezone。
h5. The throughput without udf invoke: 764.50 w/s
!image-2021-06-15-15-27-26-739.png!
h5. The throughput with udf invoke: 183.24 k/s
!image-2021-06-15-15-28-28-137.png!
h5. The throughput with udf nesting invoke: 41.42 k/s
!image-2021-06-15-15-29-09-773.png!
h5. The throughput with udf nesting invoke after the issue: 174.41 k/s
!image-2021-06-15-15-30-14-775.png!
> Improve the performance of invoking nesting udf
> -----------------------------------------------
>
> Key: FLINK-22994
> URL: https://issues.apache.org/jira/browse/FLINK-22994
> Project: Flink
> Issue Type: Improvement
> Components: Table SQL / Planner
> Affects Versions: 1.12.4
> Environment: h5.
> Reporter: lynn1.zhang
> Priority: Major
> Labels: pull-request-available
> Attachments: image-2021-06-15-15-18-12-619.png,
> image-2021-06-15-15-19-01-103.png, image-2021-06-15-15-27-26-739.png,
> image-2021-06-15-15-28-28-137.png, image-2021-06-15-15-29-09-773.png,
> image-2021-06-15-15-30-14-775.png
>
>
> h1. BackGround
> Flink maintain the udf result as binary, like BinaryStringData. When invoking
> nesting udf like select useless(int_ip_2_string(ip)), the result of
> int_ip_2_string(ip) will be serialization and deserialization.
> Below is the Generated Code
> !image-2021-06-15-15-18-12-619.png! this issue want to improve it as below
> !image-2021-06-15-15-19-01-103.png!
> h1. Performance Compare
> Condition: Source = Kafka，Schema = PB with snappy；Flink Slot =
> 1、taskmanager.memory.process.size=4g; Linux Core = Intel(R) Xeon(R) Gold 5218
> CPU @ 2.30GHz
> UDF Introduction:
> * ipip: input: int ip, output: map ip_info, map size = 14.
> * ip_2_country: input map ip_info, output: string country.
> * ip_2_region: input map ip_info, output: string region.
> * ip_2_isp_domain: input map ip_info, output: string isp.
> * ip_2_timezone: input map ip_info, output: string timezone.
> h5. The throughput without udf invoke: 764.50 w/s
> !image-2021-06-15-15-27-26-739.png!
> h5. The throughput with udf invoke: 183.24 k/s
> !image-2021-06-15-15-28-28-137.png!
> h5. The throughput with udf nesting invoke: 41.42 k/s
> !image-2021-06-15-15-29-09-773.png!
> h5. The throughput with udf nesting invoke after the issue: 174.41 k/s
> !image-2021-06-15-15-30-14-775.png!
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
