OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] XML Performance in Client-Server Interactions

[ Lists Home | Date Index | Thread Index ]

Rich Salz wrote:

>I was assuming most clients aren't getting XML and using it to generate
>new XML.  Instead, that they have a handful of local data, and they
>are generating XML from that.  I was also assuming that any native
>datatype already had a "convert to string" funcdtionality.
This is exactly what's done with data binding, and at least in my tests 
I haven't seen that data binding output is much faster than input (even 
in the case of my own JiBX framework).

I suspect I could double the performance if I wrote custom code for an 
application that went directly to text and did it's own buffering 
(assuming I knew the data would always be free of characters requiring 
special handling). But then I could also double the input performance if 
I did the equivalent custom parser for input (basically just ignoring 
all the complexities of XML that I don't want to support, the same as 
I'd be doing on the output). Maybe that's not a fair comparison, but I 
think the basic point still holds - the inherent difference in 
processing XML as input vs. output is not large. In any case, I'd expect 
that at this point of time somewhere upwards of 90% of XML is generated 
through some sort of framework, not by printf or equivalents (though if 
the Microsoft crew is correct about RSS being the top XML application I 
may be wrong on this - judging from what little I've seen of RSS much of 
it appears to be generated in this manner).

  - Dennis

Dennis M. Sosnoski
Enterprise Java, XML, and Web Services
Training and Consulting
Redmond, WA  425.885.7197


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS