[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [xml-server] Re: Status

Hi,

Thu, 16 Dec 1999 Laurent Bossavit wrote:
>> Somebody have real life expierence with eXcelon, Tamino, POET XML
>Repository ?

Some days ago the mail "XML Repository" from the ozone-users Mailinglist
was forwarded to this list. It seems that members of both have the same
ideas on using an XML Repository ... this is very interesting :-)

>My experience with eXcelon can be summarized as follows - it' a very
>nice product on paper but not stable enough that I could implement my
>"test case" application with it. This does not necessarily reflect on
>eXcelon, maybe it's my test case which is unrealistic.

>My test case consists of about 20,000 records from a traditional SQL
>database; they represent Web site descriptions, and the aggregate is a
>smallish Web directory. The records are organized in categories,
>categories being recursively nestable.
>
>A good XML pseudo-representation would be :
> (category name="root")
>   (category name="technical")
>     (site name="XML Server")
>     (/site)
>   (/category)
>   (category name="nontechnical")
>   (/category)
> (/category)
>
>My criteria for judging the fit of an XML server to my requirements :
> - I want to import the 20K records from SQL to XML in < 10min CPU time

I think it is important to know what kind of hardware you want to use
or you have used for this statement. While implementing the XML Repository
on top of ozone I run my test suite again and again. And I'm not very
satisfied with the performance but my hardware is not high end ...
so comparison would be a nice thing ..

> - I should be able to query all site records, all in a category, or all
>   in a category and subcategories
> - I should be able to retrieve category structure as part of the query
>   results, i.e. tell that a given site is part of categories X>Y>Z
>
>I've come to the preliminary conclusion that "XQL" (aka extended XPath)
>is far from sufficient as a query language; my ideal server needs
>something like XML-QL, or X-OQL. As far as I'm concerned I should also
>be able to do updates and inserts in the same query language.
>
>I have also encountered some performance problems with eXcelon (these
>might have been due to limited RAM on the server, though). Inserts were
>reasonably fast, but degraded sharply as files became large; queries
>were fairly slow. It seems to me that eXcelon's problem is that it
>won't scale well to large stores.

What exactly means "fast", "larger", "fairly slow" ...? :*)
Can you give me some exact numbers like repository size, measured time?

>I'm waiting to be contacted by a SAG rep to evaluate Tamino, so I can't
>comment yet. POET CMS appears to be more of a document management
>solution than a generalist XML database, but in any case an evaluation
>version does not seem to be readily available so I can't comment either.

Regards,
Lars
--
________________________________________________________________
Lars Martin                         mailto:lars@softwarebuero.de
softwarebuero m&b (SMB)              http://www.softwarebuero.de