[
https://issues.apache.org/jira/browse/SLING-3398?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904055#comment-13904055
]
Ian Boston commented on SLING-3398:
-----------------------------------
The benchmarks are very RDBMS focused, with high levels of concurrency and high
levels or write on some of the tests.
{code}
The request handler will be exercised at 256, 1024, 4096, and 16,384
concurrency.
{code}
Some of the datasets expected to exist at 1 level (based on the URIs) within
the repository are large (10K entities).
Being RDBMS focused the review of any implementation may not get a "realistic"
flag if the reviewers are not aware of how Sling/JCR operates, although I
notice some published implementations have only chosen to implement 1 or 2
tests (eg jetty).
Its is encouraging to see that plain old raw servlets running on Resin with
appears in top 10 or 20 on every list, something I was not expecting given the
hype surrounding so many cooler frameworks these days.
> Implement techempower.com benchmarks for Sling
> ----------------------------------------------
>
> Key: SLING-3398
> URL: https://issues.apache.org/jira/browse/SLING-3398
> Project: Sling
> Issue Type: Wish
> Components: General
> Reporter: Felix Meschberger
> Labels: gsoc2014
>
> TechEmpore provides a benchmark framework for Web Application frameworks to
> compare their performance. See http://www.techempower.com/benchmarks/
> We should implement these benchmarks for Sling to see where Sling stands in
> this comparison. It would be great if we could contribute the Sling
> benchmarks to TechEmpower to include it in their comparison chart.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
