hivemind-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Knut Wannheden <knut.wannhe...@gmail.com>
Subject Re: Interceptor performance
Date Tue, 24 May 2005 07:34:07 GMT
Konstantin,

On 5/23/05, Konstantin Ignatyev <kgignatyev@yahoo.com> wrote:
> Some time ago I compared interceptors performance, HM
> is not bad:
> http://kgionline.com/articles/aop_1/aop_perf.jsp

That's an interesting comparison. However, I think that it should be
noted that the HiveMind service interceptor framework is really
agnostic about the actual implementation mechanisms used by the
interceptors. In your example the overhead imposed by HiveMind's
interceptor framework is probably low compared to the interceptor
implementation mechanism, which is JDK dynamic proxies. If for
instance the interceptor were instead implemented using Javassist I
think you might have ended up with quite different results.

I think your preformance measurements would be even more interesting
if you instead implemented the interceptor using Javassist or CGLIB.
That would, when compared to the other approaches, give a better
indication of the actual overhead imposed by the HiveMind interceptor
framework.

Regards,

--knut

---------------------------------------------------------------------
To unsubscribe, e-mail: hivemind-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: hivemind-user-help@jakarta.apache.org


Mime
View raw message