Imagine a world where the potentials of business intelligence are more pervasive and more accessible. Thanks to an initiative that's steadily gathering steam, that reality may not be far off.

The XML for Analysis (XML/A) specification, first announced in April 2001, is a joint initiative between Microsoft, Hyperion and SAS. Representing a substantial shift in the industry, XML/A will allow open access to multidimensional databases from many providers. While still in its infancy, some think the specification may grow to be more influential than ODBC, OLE DB and ODBO.

The collaboration between these two often-at-odds OLAP vendors is expected to accelerate the adoption of Internet business intelligence software and represents significant momentum for XML-based analytic Web services. While this Microsoft-led specification had already gained wide acceptance from vendors who supported Microsoft's OLE DB for OLAP, Hyperion was the first major Microsoft competitor to join the group. Much more than a press-release vote of confidence, Hyperion and Microsoft had been collaborating for some time before the announcement.

Shortly after the initiative was announced, more than 20 related technology vendors formed the XML/A Council (www.xmla.org). Uniting the "information providers" with the "information consumers," the council's mission is to advocate for its acceptance as a standard.

Anticipation is high. If the XML/A initiative succeeds, it will finally be possible to query multiple major OLAP servers with a standard delivery method, just as it has been possible to do with all the SQL-based relational databases for many years. "It's good news for everyone that two competing major OLAP server vendors are finally putting their differences aside to come up with an agreed, XML-based OLAP query API," said Nigel Pendse, lead author of The OLAP Report ( http://www.olapreport.com/) in a recent article.

Now for the basics: Exactly how does XML/A help business partners collaborate more efficiently? XML/A is a set of XML message interfaces that uses the industry-standard simple object access protocol (SOAP) to define data access interaction between a client application and an analytical data provider (OLAP and data mining) working over the Internet. XML/A involves defining tags embedded in files so different types of programs can read information created by other programs. Thus, analytic capabilities to any client on any device or platform can be delivered using any programming language, including C++, C#, Visual Basic and Java. XML/A version 1.0 provides XML modeling of meta data and query results, and specifies MDX as the query language. A markup form of MDX (called mdXML) is currently under development and is expected to function as an object description of MDX queries. "Use and understanding of the MDX language has greatly matured in the last few years, including its differences versus SQL," says Anthony Maresco, senior managing architect at MicroStrategy.

With its genesis in OLE DB for OLAP, XML/A has many similarities; but one key differentiator it possesses is that unlike OLE DB for OLAP, it is not tied to any particular client or server platforms.

XML/A for BI and DW Deployments

A number of business and technical forces will converge to drive XML/A as a successful interoperability standard, says MicroStrategy's Maresco ­ "First, the wide participation by the market leaders of both servers and clients; second, analytical applications have become mainstream, whereas in the past they were more niche applications."

Wherein lies the promise for users, developers, vendors, and application environments and architectures?

Freedom of choice is one advantage for user organizations. "Within 12 months, it will be obvious that XML/A will be the de facto standard for interoperability among BI tools. XML/A has the critical mass of vendors that will make this 'open approach' happen. Consumers get the advantages of mixing and matching tools from different vendors and not being locked into a single vendor," asserts Michael Luckevich, CEO of Aspirity. This development may be a boost for BI in general, thinks Pendse: "This (interoperability) will give users more freedom, encourage the development of good front-end tools and help the growth of the market."

The potential to preserve ROI looks promising because organizations will gain the ability to protect server and tools investments, thus ensuring that new analytical deployments will interoperate and work cooperatively.

XML/A should be a boon to beleaguered developers struggling to work their way out of custom application request pileups from enthusiastic business users (and therefore help to bridge that always-present business and IT disconnect). With XML/A, developers can tap into existing skill sets and open access XML-based Web services, eliminating the need to program to multiple APIs and query languages. The XML/A learning curve should be gentle, insiders say. "By virtue of its basis in OLE DB for OLAP (ODBO), much of the OLAP aspect of XML/A will be familiar to developers that already use ODBO. The same MDX language is used for queries and DDL, for example," says George Spofford, chief architect at DSS Lab, a decision-support services, training and technology company. "With XML/A, client/server interactions are quick to develop and debug, especially in a network environment compared to other technologies such as RPC, COM, RMI or CORBA," comments MicroStrategy's Maresco.

User organizations are not the only ones standing to gain from open access. XML/A portends favorably for BI vendors and VARs. By writing to a single-access interface, they can significantly reduce the complexity and cost for development and maintenance.

Let's move beyond the user and vendor sphere and consider how XML/A will affect BI deployment in different environments. Luckevich is enthusiastic about XML/A's potential for thin clients. "XML/A gives a lot of value to people that are building very thin Web applications. Previous implementations of analysis tools required very large footprints on the client side. By solving this problem, XML/A will drive BI for the masses," he says.

Web architecture is another area where XML/A should leave a significant footprint. "XML/A is a Web services architecture. For those of us that believe this will be the dominant architecture of the next decade – the way that client/server was the dominant architecture of the '90s – this is very significant," says Aspirity's Luckevich. Adds Maresco, "XML/A-driven BI implementations can succeed with a variety of client and server technologies for operating systems, Web servers, Web service implementations and browsers. The greatest limiting factor is really only the speed and breadth of Web services adoptions."

XML/A vs. Competitive APIs

Here's where the controversy heats up. Is XML/A expected to replace or compete with other APIs? Probably not. As the "Switzerland" of APIs, it allows a more open means to communicate cross-platform and between information providers, and can allow a single front end to communicate with competitive back ends.

Still, XML/A is inviting comparison with the emerging Java OLAP (JOLAP) API. Expert group members include IBM and Oracle – the leading data-management vendors, both heavily invested in Java and XML and promoting application-server-centric architectures – and SAS Institute. According to the Java Specification Request, JOLAP will "provide a standard API for creating, storing, accessing and managing all meta data and data related to OLAP systems ... independent of the underlying data resource."

"JOLAP is a fairly rich Java API oriented to interactive clients. XML/A is a lightweight API covering OBDO's rich query functionality (and will soon cover OLE DB for Data Mining) for many programming languages. Vendors are taking each of these APIs seriously, and some vendors are pursuing both," says Spofford. Adds MicroStrategy's Maresco, "XML/A is scoped to analytical interactions in comparison to JOLAP which covers the entire data warehousing space."

With every multivendor API, implementation compatibility becomes a concern, and that's one of the focal areas of the XML/A Council. For vendors, the compatibility issue is two- pronged. First, does our implementation agree with the proposed specification? Second, will vendor X's implementation agree with ours? Since the initial meeting in the summer of 2001, individual task groups within the XML/A council have been advancing work on various aspects of leveling, documentation, education and testing. To answer these questions, several interoperability testing events are underway to work out any kinks in XML/A implementations prior to product deployment.

Council members are working to integrate the functionality of OLE DB for Data Mining into a future version of XML/A. Whereas Microsoft's OLE DB for OLAP and OLE DB for Data Mining are two separate APIs, XML for Analysis as an analytic query and result transport specification will cover the functionality of both.

In November 2002, the council held its first interoperability event and achieved a level of success far greater than expected at the time. Consumers and providers alike went into the private event not knowing what to expect, yet all proved a connection using XML/A. The majority were able to "see" data using the specification, and many of the remaining successfully built reports against it.

On the Horizon

On May 13, 2003, more than 20 member companies of the XML/A Council will participate in the world's first public XML/A interoperability demonstration. Leading researchers, developers, executives and industry analysts will gather to discuss how XML/A products help answer the most pressing business and technology questions of managing the complexity resulting from the explosion of BI applications. Stay tuned.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access