[
https://issues.apache.org/jira/browse/HDDS-8756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17886209#comment-17886209
]
Ethan Rose commented on HDDS-8756:
----------------------------------
I think the last benchmarking we did shows there are easier fixes to be done in
terms of buffer handling and network protocol inefficiencies that will likely
yield better results than this feature. Tasks like HDDS-11483 and HDDS-10338
are examples. [~ritesh] or [~duongnguyen] may be able to share more detailed
information here. This feature may not be bad to have, but it might be a
pre-mature optimization at this time. It assumes that the bottleneck is the
disk, and I am not sure if that is the case currently.
> Write-back cache using SSD storage
> ----------------------------------
>
> Key: HDDS-8756
> URL: https://issues.apache.org/jira/browse/HDDS-8756
> Project: Apache Ozone
> Issue Type: Improvement
> Reporter: Maksim Myskov
> Priority: Major
>
> Apache Ozone’s datanodes may suffer from slow read and write performance,
> particularly when dealing with high-density nodes and frequently accessed
> data. This leads to increased latency and decreased overall system
> efficiency, negatively impacting service quality. To address this issue, we
> propose implementing write-back caching, which can significantly enhance read
> and write performance while reducing expensive disk operations.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
