You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by Marcin Giedz <ma...@arise.pl> on 2022/08/01 09:00:33 UTC
ActiveMQ Openwire NIO + SSL - performance impact
Hello, are there any metrics /stats to show potential performance downgrade
when SSL is enabled between broker and client versus non-SSL ?
Thx
Marcin
ActiveMQ classic and autofs
Posted by "Nordstrom, Karl" <kx...@psu.edu>.
Hello,
We lose our kahadb NFS mounts when our scaled out nas is upgraded. They upgrade one node at a time and the mount migrates to a different node. The mount can be unavailable for several seconds. ActiveMQ sometimes fails when this happens. We are considering using autofs, to reacquire the mount.
Has anyone used autofs with ActiveMQ classic?
---
Karl Nordström
Systems Administrator
Penn State IT | Application Platforms
Re: ActiveMQ Openwire NIO + SSL - performance impact
Posted by Marcin Giedz <ma...@arise.pl>.
Thx Matt, that's what I was looking for ;)
M.
On Mon, 1 Aug 2022 at 16:42, Matt Pavlovich <ma...@gmail.com> wrote:
> Hi Marcin-
>
> I don’t know of any recent (and reliable) benchmarks. In general, we
> expect 1-2% impact on CPU utilization and a minor lag in latency — if not
> completely negligible impact in practical real-world scenarios where
> consumer code is performing tasks.
>
> SSL adds a bit of CPU overhead, but today’s processors are magnitudes
> faster than the disk I/O and network I/O which tend to be the performance
> factors for messaging and event-driven systems.
>
> Are there specific concerns or features of your use case where SSL
> performance is noticable?
>
> -Matt Pavlovich
>
> > On Aug 1, 2022, at 4:00 AM, Marcin Giedz <ma...@arise.pl> wrote:
> >
> > Hello, are there any metrics /stats to show potential performance
> downgrade
> > when SSL is enabled between broker and client versus non-SSL ?
> >
> > Thx
> > Marcin
>
>
Re: ActiveMQ Openwire NIO + SSL - performance impact
Posted by Matt Pavlovich <ma...@gmail.com>.
Hi Marcin-
I don’t know of any recent (and reliable) benchmarks. In general, we expect 1-2% impact on CPU utilization and a minor lag in latency — if not completely negligible impact in practical real-world scenarios where consumer code is performing tasks.
SSL adds a bit of CPU overhead, but today’s processors are magnitudes faster than the disk I/O and network I/O which tend to be the performance factors for messaging and event-driven systems.
Are there specific concerns or features of your use case where SSL performance is noticable?
-Matt Pavlovich
> On Aug 1, 2022, at 4:00 AM, Marcin Giedz <ma...@arise.pl> wrote:
>
> Hello, are there any metrics /stats to show potential performance downgrade
> when SSL is enabled between broker and client versus non-SSL ?
>
> Thx
> Marcin