You are viewing a plain text version of this content. The canonical link for it is here.

Posted to jira@kafka.apache.org by "Chia-Ping Tsai (Jira)" <ji...@apache.org> on 2025/04/12 15:15:00 UTC

[jira] [Updated] (KAFKA-19134) Implementation of a Simple, Low-Latency Bounded Shared Buffer

     [ https://issues.apache.org/jira/browse/KAFKA-19134?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Chia-Ping Tsai updated KAFKA-19134:
-----------------------------------
    Attachment: tmp.patch

> Implementation of a Simple, Low-Latency Bounded Shared Buffer
> -------------------------------------------------------------
>
>                 Key: KAFKA-19134
>                 URL: https://issues.apache.org/jira/browse/KAFKA-19134
>             Project: Kafka
>          Issue Type: New Feature
>            Reporter: Chia-Ping Tsai
>            Assignee: Chia-Ping Tsai
>            Priority: Major
>         Attachments: tmp.patch
>
>
> Using the shared buffer in the network threads can reduce the GC count (and elapsed time). The attached patch offers a simple memory pool based on ConcurrentSkipListMap, and in my small env (1000000000 records), the GC time is reduced from 127.808 ms to 93.608 ms), and the created bytes are reduced from 40GB to 2GB.
>  
>  



--
This message was sent by Atlassian Jira
(v8.20.10#820010)