CBR - ORACLE FUSION SERIES - Right Time, Right Place?

Over the years the speed of data processing technologies has matured, giving rise to the concept of real-time business intelligence (BI) and data warehousing, and spawning new acronyms such as event-driven business activity monitoring (BAM). But more importantly, companies are now understanding the importance of having a robust middleware infrastructure in place to support these real-time initiatives.

When it comes to the business of explaining real-time BI, David Pryor, VP of Oracle's BI and data warehousing business, EMEA, turns to an unlikely source - Einstein's Theory of Relativity - which predicted that time does not flow at a fixed rate in the physical universe.

Fast forward nearly a hundred years and Pryor believes real-time BI can be explained in similarly relative terms. "I prefer to use right-time. It's not just about delivering relevant and accurate information for decision- making. It's doing it at the right time, and in the right context." Pryor says that companies planning a right-time BI environment should not focus solely on speed but also business context. He believes it is important to match technology requirements to the actual action times required by the business. "For a financial company that could be hourly daily, or weekly," he says. "But for an online retailer it could be minutes or seconds," says Pryor.

Implementing real-time BI infrastructure throws up technical integration challenges at both ends of the IT stack. At the back-end, or data management layer, real-time requires moving, transforming and loading large volumes of business data from applications into data marts, data warehouses or nimbler operational data stores (ODS) for offline analysis. At the front-end, or analysis and reporting layer, it requires BI to be aligned closely with decision touch points inherent in any business processes: a strategy that has become popularised as 'operational BI'.

Pryor likes to think of right-time BI as "a mix of operational, event type activities matched with scheduled, batch-driven analysis of information typically stored in a data warehouse." For example, a credit-scoring application may be regarded as a real-time system. But the majority of the analysis is completed offline (as a batch process) and only the very last calculation for the actual decision is done in real-time. In other words, it is an event-driven activity that is done at the behest of end users, not the convenience of IT.

"The trick is blending these two architectures together," Pryor says. And it is here where Oracle's Fusion Architecture comes into play, providing an underlying set of middleware technologies that continuously fuses event-driven business processes and applications with traditional batch-driven data warehousing analysis through its service oriented architecture (SOA). Technically, Fusion provides a unified, standards-based technology blueprint that details the linkage between enterprise applications, middleware and grid infrastructure technologies. Fusion also brings the value of middleware components to bear on real-time BI infrastructures by addressing data and process integration, which have arisen from companies deploying enterprise applications using a fragmented IT infrastructure.

Gearing up

Pryor says that the Fusion middleware family of products provides all the key ingredients for real-time, including bi-directional data and application integration, business activity monitoring technologies, a rules engine, process integration and, of course, BI.

Right now BAM is perhaps the most visible application of real-time BI and it is the component in Fusion that is getting the most attention from real-time aficionados. Fusion's monitoring of high throughput processes combines data collection with process and workflow management capabilities to monitor streaming data from operational systems to detect business events. Pryor explains this can be used in variety of real-time analyses such as spotting production-line problems, a spike in customer complaints, or dwindling stock on a retailer's shelf.

Key performance metrics and alerts driven by these events are served up in a relevant and timely manner through portals and dashboard interfaces or through handheld devices and mobile phones. "Terms appearing like BAM and portals are really a spectrum of what people are looking for in real-time BI. Both capabilities are embedded in the Fusion technologies we have now," Pryor points out. A recent Oracle survey confirms this, showing that real-time monitoring plays an important part of business process improvement by helping companies identify issues for resolution 'in flight'.

Arguably the most critical component of Fusion for enabling 'righttime' BI is the business process execution language (BPEL ) engine.

Pryor believes that embedding BI analysis into people's day-to-day business processes is a great way to get a handle on right-time intelligence. Oracle uses the BPEL Process Manager to attach BI and analytics to Fusion-driven applications processes so users can see the information as they make decisions.

Pryor expects BPEL orchestration to become much more deeply integrated with Fusion to allow for more process-driven and real-time implementations of BI. Oracle's Daily Business Intelligence software, which is also integrated into Fusion, is perhaps the best example of Oracle's vision of real-time, process-driven BI in action today. "In effect, our Daily Intelligence is a real-time reporting environment for specific functional users such as HR, finance and procurement," Pryor says.

Unified platform

A key advantage that Fusion middleware has over other enterprise application integration (EAI) and legacy middleware stacks is that it supports all these real-time capabilities in a single, common platform.

Pryor says that piecemeal middleware infrastructures only partially address the issue of connecting systems, enabling them to communicate solely with each other. "Sure it's possible to provide the same set of capabilities using separate EAI tools and traditional middleware.

But you're faced with an uphill task of stitching it all together and then dealing with ill-defined or inconsistent interfaces that constantly need to be reworked every time there is a change."

Pryor says these traditional middleware solutions come at a high cost to the organisation in the long run.

According to Oracle's research, middleware complexity represents up to 50% of corporate IT costs today. Because Fusion is a pre-integrated environment, Pryor argues, it sidesteps the need to hardwire different middleware products and technologies together. "Fusion eliminates the middleware complexity in building real-time BI systems by providing singularity across the development, deployment and management architectures."

Another advantage of Fusion is its 'hot-pluggable' design, which makes its BI component 'neutral' when it comes to integrating heterogeneous data sources, something that BI vendors have attacked Oracle for in the past. The best example of this is the Enterprise Edition of Oracle's recently announced Business Intelligence Suite, which includes analytic technologies gained through the acquisition of Siebel Systems.

This technology allows BI users to mix and match Oracle's middleware components with their existing infrastructure, including databases such as IBM DB2, Microsoft SQL Server and Teradata, and even business applications from SAP. "Our hot pluggable BI optimises not only Oracle's database and middleware stack but others' as well," Pryor says. Oracle has been strengthening Fusion's real-time data management capabilities in other areas too. Last year it bought real-time database company TimesTen, which provides infrastructure software to manage large volumes of real-time data for a growing number of performance-critical applications.

While Fusion promises to take real-time or 'right-time' to a new level, the real challenge goes beyond just getting the right middleware and BI infrastructure in place. It also forces companies to re-engineer their business processes to exploit real-time data. "Being able to take real-time action on real-time insights is key," says Pryor.

PART ONE OF THE ORACLE FUSION SERIES CAN BE FOUND HERE
PART TWO OF THE ORACLE FUSION SERIES CAN BE FOUND HERE
PART THREE OF THE ORACLE FUSION SERIES CAN BE FOUND HERE
PART FIVE OF THE ORACLE FUSION SERIES CAN BE FOUND HERE