DB2 Native XML with Xquery in Open Beta

[Kiran]

See the docs: http://news.zdnet.com/2100-9593_22-6057835.html
Download:
http://www.ibm.com/software/data/db2..._download.html
Docs:
http://publib.boulder.ibm.com/infoce...w/v9/index.jsp

Just to give a brief idea, How XML in db2 is implemented as compared
to other Databases. XML in V9 DB2 is not stored as CLOB or BLOB
internally, nor shredded and stored as relational tables. XML is stored
as XML, rather as format which is in standards with w3c to enable
XQuery access to XML data. Which means that XML can be parsed by XQuery
Language (Like SQL For Relational data) and the XML file in one or more
pages in the table space, and the file is parsed only once to access
the data (because of architecture, when compared to be internal
implementation as LOBs).

What is in for App Developer is, once you store XML file in DB2 as XML
data type, you can use XQuery to access the data, can create
relational views out of XML data. Will not need complex transformations
that need to be done in application layer for manipulating the data,
most of XML manipulation can be done in Database layer.You can store
XML which adheres to some schema and hence do the checking using
database .

Another thing I feel, new for App Developer in Viper is the Label based
access control. You can define security policy for each object, even
attributes of the object (like columns in tables) and give access to
different roles.

Do check out Viper :) it has lots of autonomic features for
Administrators ..the who member management has gone for a an overhaul,
it faster, better and less maintanence.

[Joe Kesselman]

I'm delighted to see this finally go public.

I was somewhat involved in the design of the new DB2 XML data model (not
in coding it, but as consulting expert since I've implemented a number
of other XML models for IBM and have been involved in some of the XML
standards work).

As Kiran says, it has the major advantage of storing an XML infoset in a
fully parsed form which can be efficiently manipulated via XML-based
tools, rather than the older techniques of storing XML in text form
(which would require reparsing/regeneration every time you read or
write), as a BLOB (which would require deserialization/reserialization
into other data representations -- cheaper but still some overhead), or
in shredded form (which is fine for some applications, but not for tools
like XQuery/XSLT/etc. which want to manipulate the XML as XML rather
than as fully-extracted content).

Note that those other representations are of course still available and
sometimes appropriate; this just gives you another option which is
better suited to some tasks.

I'm not sure how many of the details have been published, so I'm going
to have to decline to answer questions on that front until I've
investigated it -- and I suspect I'm out of date myself since I haven't
spoken to that team recently -- but it's definitely worth checking out.
DB2 is already a world-class relational database; I think this addition
will make it a world-class XML database as well.
--
() ASCII Ribbon Campaign | Joe Kesselman
/\ Stamp out HTML e-mail! | System architexture and kinetic poetry