[Hessian-interest] Benchmarking Hessian and Burlap

Christian Campo christian.campo at compeople.de
Tue Jan 8 02:29:42 CST 2008 

Hi Daniel,

very interesting benchmark given the many protocols you compare. I did  
a similar thing some years ago only comparing Axis, Hessian and RMI.

I wonder if you shouldn't also include numbers about the size of the  
request and the response data. Since Network is not in all conditions  
infinite fast that will matter.

Because then it shows that Hessian is not only by factors faster but  
also by factory smaller which would really hit you when transfering  
large requests over a mobile

internet connection (GPRS, EDGE, UMTS).

Good to know the Hessian 2 is so much faster. I wasn't aware of that.

thanks
christian campo

Am 08.01.2008 um 05:35 schrieb Daniel Gredler:

> Very nice!
>
> I've added numbers for Hessian 2:
>
> daniel.gredler.net/2008/01/07/java-remoting-protocol- 
> benchmarks/
>
> On Jan 7, 2008 8:23 PM, Daniel Gredler <daniel.gredler at gmail.com>  
> wrote:
> Hi Scott,
>
> I used Spring's Hessian integration. Looking at Spring's  
> HessianClientInterceptor class, it looks like it's just using the  
> default, Hessian1.
>
> It also looks like it might possible to tweak the Spring config  
> files to use Hessian2... Is there a newer Burlap protocol I should  
> take a look at as well?
>
> Take care,
>
> Daniel
>
> On Jan 7, 2008 6:39 PM, Scott Ferguson <ferg at caucho.com> wrote:
>
> On Jan 7, 2008, at 3:19 PM, D&J Gredler wrote:
>
>> Hi Scott,
>>
>> Thanks for the info. We're pretty much done with out benchmarking  
>> here, our results are at:
>>
>> daniel.gredler.net/2008/01/07/java-remoting-protocol-benchmarks/
>
> Interesting.
>
> What method did you use to invoke Hessian?
>
> I just checked with the HessianProxy and it uses Hessian 1.0 by  
> default, so I'm curious if the results are for Hessian 1.0 or 2.0.
>
> > As far as I can tell, Caucho's Burlap is essentially Hessian-in-XML.
>
> Historically, it's the other way around.  Hessian is Burlap-in-binary.
>
> -- Scott
>
>>
>> Regards,
>>
>> Daniel
>>
>> On Jan 2, 2008 6:10 PM, Scott Ferguson <ferg at caucho.com> wrote:
>>
>> On Jan 2, 2008, at 1:36 PM, Daniel Gredler wrote:
>>
>>> Hi,
>>>
>>> Is anyone aware of any benchmarks comparing Hessian and/or Burlap  
>>> to other (Java) remoting technologies? I'm working on some  
>>> benchmarks and would like to compare results, once I'm done.
>>
>> Not recently.  The Hessian2 implementation should be faster and  
>> more compact than java.io serialization.  I'm not sure how it would  
>> compare to RMI, since RMI doesn't have the HTTP overhead.
>>
>> Burlap should be much slower than Hessian.  And the original  
>> Hessian is significantly slower, partially due to the protocol, but  
>> also because the Hessian2 implementation improved things like  
>> buffering.
>>
>> -- Scott
>>
>>
>> -- 
>> Daniel Gredler
>> daniel.gredler.net/
>> _______________________________________________
>> hessian-interest mailing list
>> hessian-interest at caucho.com
>> maillist.caucho.com/mailman/listinfo/hessian-interest
>
>
> _______________________________________________
> hessian-interest mailing list
> hessian-interest at caucho.com
> maillist.caucho.com/mailman/listinfo/hessian-interest
>
>
>
>
> -- 
> Daniel Gredler
> daniel.gredler.net/
>
>
>
> -- 
> Daniel Gredler
> daniel.gredler.net/  
> _______________________________________________
> hessian-interest mailing list
> hessian-interest at caucho.com
> maillist.caucho.com/mailman/listinfo/hessian-interest

-------------- next part --------------
An HTML attachment was scrubbed...
URL: maillist.caucho.com/pipermail/hessian-interest/attachments/20080108/11d2effe/attachment-0001.html

More information about the hessian-interest mailing list
Related searches:
hessian interest daniel gredler burlap
gipoco.com is neither affiliated with the authors of this page nor responsible for its contents. This is a safe-cache copy of the original web site.