to homepage
 Weekly emails: how to advanced search
 Glossary lookup:


> stories > decision point

EII helps BI adapt to SOA

by Phil Wainewright
June 27th, 2005

Users want information provided in a form that's relevant to their business needs, irrespective of where the data resides.

• print  • comment
Enterprise information integration (EII) brings real-time web services into business intelligence (BI) tools, but at a cost:
  • EII can also feed live data to portals and dashboards
  • It can combine data warehouse contents with other data sources
  • Many BI tools use EII to read data from web services feeds
  • Few EII projects take place within an SOA strategy
  • Most enterprises lack a co-ordinated policy for data services

Glossary terms: BI, EII, XBRL, data warehouse, SOA, lookup tool

There was a time when the best they could hope for was a batch download into a spreadsheet or, if they were specially favored, a regularly updated data warehouse and a set of business intelligence (BI) tools with which to analyse the data residing there.

Today, they increasingly expect access to live data too, perhaps combined with information from a variety of other sources. These demands are creating headaches for enterprise architects and vendors alike, raising question marks over the future role of front-end BI tools in presenting information to users, while posing difficult choices when deciding how best to link to back-end data sources. A new class of software called enterprise information integration (EII) aims to ease some of the headaches, but the variety of approaches adopted by EII vendors raises as many questions as it answers.

One form of EII simply uses a web front-end to deliver up-to-the-minute transaction data to users without adding the kind of analysis capabilities traditionally provided by BI tools. Up to 20,000 users at computer manufacturer HP, for example, share access to a portal where they can view real-time financial data, combined with information from a data warehouse and an archive of MS-Word reports. BI doesn't figure in this type of application because "there's not a whole lot of data analysis required," explains Tim Matthews, co-founder and VP marketing of Ipedo, whose 'dual core' SQL/XQuery engine performs the integration. "EII allows you to create that interaction that goes out and touches those data resources, without having to set up a whole lot of new infrastructure."

In a souped-up version of the same principle, a major US logistics operator has adapted Cisco's newly launched AON technology to get real-time visibility into the quality of service it's providing to its major customers. While its existing IT systems allow it to track the status of any individual parcel in its network, AON adds the ability to collect information on all of the consignments in transit for any given customer and report aggregate service levels. Coupled with a dashboard application from software partner CXO Systems, the AON solution allows managers to view consolidated performance information for each of its top 10 or top 100 customers — based on up-to-the minute data from live transactions.

Data warehouse role
The availability of real-time access to live business data — business visibility, as Cisco likes to call it — will draw a line under enterprise investment in data warehouse products, says Michael Carter, co-founder and chief marketing officer of CXO Systems: "The data warehouse is going the way of the mainframe." Customers will keep using what they have, he says, but the emphasis now will shift to new investments in network-aware dashboard products that link deep into high-speed infrastructure. BI tools aren't relevant for this type of application, he adds, because they're designed for a different purpose. "Really deep analysis and planning, the BI guys are good at — there's a place for that," whereas the CXO Systems dashboard, he explains, is optimized to present up-to-the-minute operational information for real-time performance management.

Data warehouses and BI tools will still have a role, however, and several EII vendors claim their products help them deliver more value to a wider cross-section of users. One of the principal use cases for Composite Software's EII product, for example, according to CTO and co-founder Mike Abbott, is to combine information from a data warehouse with live data to create a "hot query."

At the same time, EII can provide a mediation layer that insulates the data warehouse from the impact of unmanaged raw queries, says Rob Cardwell, CTO of EII vendor MetaMatrix. "You're really trying to make the right type of information available to the consumer irrespective of where it happens to [be]." EII can act as a "data services layer" that both enables access to information sources — for example, far-flung data marts or data warehouses in separate departments — and also allows the access to be controlled and managed. "Instead of the interface being ad-hoc SQL, [users] can have constrained services that grab information in byte-sized chunks."

Beefing up BI
Business intelligence tools can be enhanced, too, by the addition of EII capabilities — which perhaps explains the flurry of alliances that have been struck this year between the two camps. MetaMatrix has linked up with Business Objects and Hyperion while Composite Software has forged alliances with Informatica and Cognos, sealing the latter with a $4.5 million equity stake from the BI vendor. Abbott explains that products like Cognos don't have the ability to consume data from web services, so the Composite technology helps out by providing a SQL interface to the data services: "We can expose our composite views either as a relational view or as a service."

In a tie-up announced last week, Ipedo is performing a similar function to bring XBRL-formatted financial data from SEC filings into Business Objects' BI tools. Matthews says this is the typical EII application that Ipedo sees in the market today: bringing a web services feed into an existing BI environment. In the case of Edgar Online's XBRL feed, says Matthews, "The financial analysts and regulators can have the same tools as they're used to and have all this fresh new data coming in live."

Services layer
Integrating individual services into BI tools is a far cry from the vision of an enterprise-wide 'data services layer' articulated by the likes of Composite and MetaMatrix. "In the market today I would say the data services layer is the future not the present opportunity," says Matthews. Many existing users of BI and reporting tools "have no interest at all" in moving to a services architecture, he says, and indeed Ipedo earlier this month responded to BEA's AquaLogic launch by announcing a migration program aimed at users unhappy with BEA's transitioning of its EII server, Liquid Data, to an SOA platform. "While we see the value of EII within SOA there are a number of people who aren't moving to SOA," he explains. "These people aren't really happy about being forced to move to SOA on BEA's schedule rather than their own schedule."

But there's a risk for enterprises who take a more project-based approach to deploying EII, rather than putting it in the context of a broader SOA strategy. They may end up with a variety of separate 'data services layers' scattered around the organization, each operating with their own individual metadata and security frameworks. Other platforms are also adding data services capabilities, further complicating the picture. Sonic Software, for example, upgraded its enterprise service bus product in March, adding a module to access and reuse relational data sources — while Cisco's AON and similar products provide another integration alternative.

Some enterprises that start out looking at EII as a means of extending BI do quickly realize the need to put it into the context of an SOA strategy — Cardwell says that MetaMatrix customer Motorola falls into this camp — but they're the exception rather than the rule. EII may make it easier to extend the productivity of BI and unleash powerful new business visibility capabilities, but if each individual EII project makes its own rules for how services interoperate, reconciling those differences across the enterprise could be tricky.

More on this topic


The death of data warehousing
The cost, complexity, and utility of the data warehouse will become marginalized ...

Integrating information at the desktop
New services-based approaches are empowering business users ...

Don't look in the attic for fresh answers
Traditional approaches to managing customer data ...


Loosely Coupled Monthly Digest May 2005
Crucial metadata role pulls EII to the heart of SOA


Copyright © 2002-2006, Procullux Media Ltd. All Rights Reserved.