[
https://issues.apache.org/jira/browse/TUBEMQ-225?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17141250#comment-17141250
]
Guocheng Zhang edited comment on TUBEMQ-225 at 6/21/20, 4:09 AM:
-----------------------------------------------------------------
I used the Kafka 2.5.0 version to re-verify, the test results are not much
different from the Kafka 0.10.2.0 version, or even a little worse; The main
features of tubemq are higher performance and lower delay difference compared
with other MQs under single machine, disk storage, a large number of topics or
partitions , this is a comparison of actual use, and the data is real, so it is
more appropriate to maintain the previous datas Since the VS data has been
published before.
*{color:#DE350B}Attention:{color}{color}* in the VS test scheme data, in order
to improve the throughput of Kafka, the use case has done a delay packing, or
the Kafka's TPS will not reach the data value we said, correspondingly, the
message delay will increase, which we need to pay attention to: I know Kafka
can achieve low latency, but in high-performance scenarios, when it has to
adjust the characteristics of low latency to achieve high throughput, so we
can't say it is still low latency in high-performance scenarios
I will bring up this PR again.
was (Author: gosonzhang):
I used the Kafka 2.5.0 version to re-verify, the test results are not much
different from the Kafka 0.10.2.0 version, or even a little worse; The main
features of tubemq are higher performance and lower delay difference compared
with other MQs under single machine, disk storage, a large number of topics or
partitions , this is a comparison of actual use, and the data is real, so it is
more appropriate to maintain the previous datas Since the VS data has been
published before.
Attention: in the VS test scheme data, in order to improve the throughput of
Kafka, the use case has done a delay packing, or the Kafka's TPS will not reach
the data value we said, correspondingly, the message delay will increase, which
we need to pay attention to: I know Kafka can achieve low latency, but in
high-performance scenarios, when it has to adjust the characteristics of low
latency to achieve high throughput, so we can't say it is still low latency in
high-performance scenarios
I will bring up this PR again.
> [website] Add performance comparison doc
> ----------------------------------------
>
> Key: TUBEMQ-225
> URL: https://issues.apache.org/jira/browse/TUBEMQ-225
> Project: Apache TubeMQ
> Issue Type: Improvement
> Reporter: Guocheng Zhang
> Priority: Major
> Labels: pull-request-available
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
