[8u] RFR 8259886: Improve SSL session cache performance and scalability

Martin Balao mbalao at redhat.com
Fri Apr 9 17:45:22 UTC 2021


Hi Hedongbo,

Thanks for contributing this backport to 8u.

The original change makes perfect sense for 8u; I see nothing
8u-specific since we have almost the same SunJSSE engine than in newer
releases. The change is fairly trivial, low risk and -as demonstrated-
generates a good performance benefit.

In regards to your backport, the only thing I'm not sure is the
microbenchmark removal. I see many of these tests pushed to jdk/jdk,
only a couple to 11u -plus the one added by 8259886-, and no tests in
jdk8u. I now wonder if it was correct to push some of them to 11u. I
guess we will not push it to 8u, but I'm CCing Aleksey for his input on
this matter.

Thanks,
Martin.-



On 4/7/21 7:18 AM, Andrew Haley wrote:
> On 4/6/21 7:38 AM, He Dongbo(JVM) wrote:
>> JMH uses the microbenchmark in the 11u patch. In general, the CacheBench.put averageTime of JMH increased by 99% (timeout=86400), and the tomcat TPS increased by 33%. Any thoughts?
> 
> Looks reasonable enough, but I'd like a security specialist to have a
> look too. Cc: mbalao.
> 



More information about the jdk8u-dev mailing list