InfoWorld Test Center
XML deepens the data content of e-business applications, but it also
makes data management more complex largely because most companies store
their XML documents in database management systems that aren't
optimized for XML.
This leads to either poor access performance or programming hassles for
developers, who have to reconcile data structure differences in their
code to let XML documents flow through standard databases. The result?
Longer, more costly development cycles and the need for additional
hardware resources.
Tamino (http://www.softwareag.com/tamino), from Software AG, promises
to rectify these matters. Tamino, a database management system, can
store and deliver XML documents in native format and offers a set of
tools for effectively managing XML data.
I was impressed with Tamino's strong data management capabilities;
however, its awkward user interface prevented it from scoring higher
than Good in my evaluation.
Installation was spotty; Tamino was very attentive to some elements but
less helpful on others. For example, the program warned me to install a
missing prerequisite Service Pack 5 for Windows NT but did not set up
the ODBC environment, which I had to prepare afterward.
The product documentation, thankfully, clearly described all the steps
in the process.
Cranking out XML data
Tamino's most innovative feature is its data engine, the X-Machine,
which can efficiently store and retrieve XML data structures without
altering the hierarchical structure or sequence.
You access the X-Machine via the Tamino Manager, a browser-based
interface that simplifies the process of defining XML databases by
creating links to relational data structures and managing access
rights. The Tamino Manager is the main access point to the data engine
and is similar in functionality to IBM DB2's Control Center or
Microsoft SQL Server's Enterprise Manager.
Defining an XML database with the Tamino Manager was a no-brainer. A
wizard guided my steps, suggesting default values for most options.
To improve performance the Manager's wizard lets you easily adjust the
value, location, and size of data, indexes, and log files. For example,
you can spread data or log files across different disk devices much as
you would in a traditional DBMS.
Troublesome UI
With just a few keystrokes I had created my first XML database. My only
complaint was with the interface itself. Other databases offer a
consolidated user interface that facilitates administrative tasks, but
Tamino's GUI isn't that friendly, which is the product's major
downfall.
To create a schema from an XML document and load the actual data, you
must invoke another browser-based interface, called the Inter-active
Interface (II), by launching it from the Windows Start menu.
Because Tamino's communication protocol is HTTP, II operations must be
integrated with your Web server. This involves pointing II to your
database by providing it with the URL. Unfortunately, Tamino doesn't
offer any memory for existing locations. Once you've done that, you can
define, load, delete, or query XML data structures similar, for
example, to the way SQL Server's Query Analyzer works in relational
environments.
Although Tamino can work with popular Web server offerings from
Netscape, Microsoft, and Apache, it offers no help in configuration.
You have to set up your Web server manually. This turned out to be a
fairly involved process; I had to rely on both the product
documentation and some phone support from Software AG before I could
connect II to my new database.
In addition, Tamino requires closing and restarting your Web server to
acknowledge a new XML database, which can create disturbances in
production environments. A better solution, in my opinion, would be to
let users invoke II directly from Tamino Manager. This would solve the
database addressing issue automatically, thereby simplifying
administrative tasks, especially when the number of databases grows.
Once I had settled my configuration woes it was easy to define the
schema and load the database. I could browse the file system from II
for my sample XML schema and documents.
Once my new schema was loaded I could also use II to query the database
using XQL (XML Query Language). Tamino implements standard XQL with
some additions, such as generic search capabilities, to retrieve data.
What surprised me was that I didn't have to log on to the database to
perform these tasks, although I did have to connect to the database to
work with Tamino Manager. I was obviously logged on to Windows NT, but
most data management tools keep database and OS access rights separate.
Creating separate access rights for the database is one tweak that
Software AG could make to strengthen Tamino a bit and allay some
security fears.
Tamino also provides a set of data management tools that lets you tune
the configuration of each XML document according to specific load
demands. For example, you can define indexes to boost data retrieval
speed directly into the XML document definition.
Obviously, providing specific indexes of XML fields will make for a big
performance advantage over XML documents stored in a relational
database.
Pulling sources together
Another interesting feature is Tamino's capability of describing a
composite data structure by assembling information from heterogeneous
data sources. In fact, you can describe many different kinds of data
content in XML notation, thus creating a uniform interface for various
applications.
For example, you can create an XML structure using data from existing
relational tables and XML documents. You can deepen the content of your
relational databases, including such extras as multimedia data or
transaction-specific information, without changing the existing
databases.
Your application will include only one data structure. But behind the
scenes Tamino will retrieve the information from each data source and
feed it to your code.
Another component, the X-Node, lets you map XML definitions to
relational tables. From the Schema Editor interface, which you invoke
from Tamino Manager or the Windows Program Menu, you can interactively
modify XML data definitions and, when appropriate, link each field to a
corresponding column in a relational table.
This process takes a bit of effort, but once done you can access all
your data under the common XML paradigm. These features improve the
overall usability of your data and make it easier and less expensive to
develop e-business applications.
Even though Tamino offers a general catalog to describe its data,
you're not limited to the catalog formats. As long as your XML
documents are well formed you can add new field definitions, thereby
giving yourself greater flexibility to manage changes.
Rough edges
I reached the end of my first experience with Tamino with mixed
feelings. The product clearly has tremendous potential for making XML a
viable solution for e-business applications, and it's an exciting new
entry into the database arena that companies should evaluate.
Still, I was slightly disappointed by some rough edges that could be
smoothed in future releases. With a better, friendlier interface,
Tamino would rank as an excellent product despite its steep $25,000-per-
CPU price tag, which can easily be offset by shortened development
cycles and improved XML data retrieval.
But until then, Software AG's Tamino can still give your diversified
data a common XML look and feel, and its XML-specific database
management features are currently unmatched.
