[Mondrian] XMLA response size

Julian Hyde jhyde at pentaho.com
Fri Aug 7 13:00:53 EDT 2009


 



 Matt Campbell wrote:
 
Regarding compression:  Analysis Services 2005+ uses a compressed binary
representation of their XMLA as the default mechanism, but their binary
protocol is proprietary.  Without an open standard we'd have no way to
compress the XMLA in a way that clients could understand.

Matt,
 
I tend to think the same way as you on matters of standards compliance. I
think that by defining a very verbose standard, and then making compression
a non-standard extension, Microsoft sent us all down the river and kept the
only paddle.
 
XML compression is the obvious solution here. The compression rate should be
so high, and so easy, that it will improve CPU utilization as well as
bandwidth. Absent an open standard, if we were to use a widely used, open
source XML compression algorithm, I think it would be easy to incorporate
into any client that wants it.
 
Specifically, could you incorporate it into your client? I'm thinking that
we'd have an HTTP header in the request and response describing the
compression algorithm to use, and you'd have to interject a
compressor/decompressor.
 
Assuming you could add a compressor/decompressor to your client, let's
continue to consider that as an option. Does anyone know of any candidate
compression algorithms? I suspect that it would be a compression algorithm
targeted at XML specifically -- because of the characteristic
problems/opportunities in XML data -- but we could consider others.
 
Julian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20090807/4a2b8a94/attachment.html 


More information about the Mondrian mailing list