Over the last year and a half or so, many large enterprises have been increasingly waking up to the potential for using an EDA for real-world value-add. Now, many of them have the basics of SOA in place. Data from the field is very preliminary but suggests that administration and governance is a frequent use case of EDAs -- and a high priority for organizations. EDA is also highly valued for business process management (e.g., operations, logistics). Support for "management by exception" and supply chain management seem obvious targets for EDA but are apparently not yet as interesting to users.
Why is an EDA proving to be so important to the business?
In any organization, it's practically axiomatic that if you don't get something (email, reports or whatever) off your desk immediately as it arrives, it will be buried under piles of other material and may never get done.
Corporations can achieve enormously greater effectiveness by being able to detect key events (that used to be held only in papers on your desk) and handle them immediately, rather than with a semi-manual, multi-person process with unavoidable delays at each step. Despite decades of re-engineering, data warehousing and business process management initiatives, what has held us up until now is that there has been little if any ability to create a simple, automated event-response loop at the business level. Most applications and software weren't built to do that. An EDA provides the foundation for this kind of detection/response business-level event handling. Once businesses start thinking about the list of environmental events they handle that could benefit by this kind of speedup/automation, an EDA can spawn all sorts of innovative software -- or handle existing tasks more cost-effectively.
Business activity monitoring and its relationship to event-driven architectures
One area that has attracted great attention is "business activity monitoring." This is not systems management; it is an application on top of an EDA that filters, abstracts and reports on events passing over the network that are particularly relevant to corporate upper management. A variant of business activity monitoring is the "360-degree view" concept, which applies the same techniques to content from the outside environment (particularly news feeds on the Web) and to the internal network -- for example, informing the CEO about political problems in India that may affect sales, or alerting the CFO to the fiscal problems of the enterprise's investment bank.
Event-driven architecture vendors and technology
How well are vendors prepared to deliver EDAs, and applications on top of EDAs? Many implementations are still retrofits of existing products, focused on one kind of event processing (e.g., systems management). However, it is clear that plenty of software vendors provide a platform that users can elaborate to create an application-spanning EDA.
Here is a partial list of some particularly interesting technologies:
- Oracle – Oracle EDA Suite (data) and AquaLogic ESB (from BEA)
- IBM – IBM WebSphere Business Events (business transactions) and Cognos BEM (data)
- Sun – Sun Intelligent Event Processor (network administration)
- Progress Software – Apama (business transactions, stock arbitraging)
- TIBCO – Business Events (business activity monitoring)
- Streambase – Complex Event Platform (stock arbitraging)
- Agent Logic – Complex Event Processing (security)
- Avaya – Avaya Event Processor (business activity monitoring)
- Coral8 – Coral8 Server (stock arbitrage, business intelligence/monitoring ["360-degree"])
- BMC – BMC Event Manager (network administration)
- INETCO – INETCO Insight
- Gemstone – GemFire Real-Time Events (data)
- Prism Technologies – OpenSplice DDS
- RTI – RTI Event Processing
- Solace Systems – Content Routers
- Eventzero – Event Processing Network
- Aleri – Aleri Streaming Platform
- IDS Scheer – Aris Platform
- Sybase – Data Integration Suite, Real-Time Events (data)
- NeuroDimension – NeuroSolutions, TradingSolutions (stock arbitraging)
Different communities may find different technologies better suited to their needs. For example, DB2 users may be interested in IBM Cognos 8 Business Events Management (BEM). Because Cognos is primarily in the business intelligence and performance management space, Cognos BEM focuses on reporting and scorecarding of events involving data transactions for business users. Thus, Cognos BEM added to a DB2 implementation offers continuous event reporting and processing that go well beyond existing periodic reporting and ad hoc decision-support applications.
Event-driven architectures and the bottom line for data management
An EDA is most effective when existing business rules, triggers and related stored procedures are surfaced and made available to the middleware's event processor. That will mean additional work for the database administrator in the short run. It should simplify data management in the long run, however, because (a) corporate standards and changes in corporate standards can be more easily attached to the new event handlers, and (b) the new event handlers can be more easily applied across databases.
More subtly, there is a great deal of under-appreciated business-rule enforcement going on in corporate databases. The EDA's visibility to upper management will mean that the usefulness of this code will be clearly established in the CFO's and the CEO's minds. In effect, implementing an EDA will, in the long run, make IT's data management efforts better appreciated and more clearly tied to the corporate bottom line. If effectively implemented, an EDA can be a big win for data managers.
Event-driven architectures may require a new mindset
One key aspect of EDAs remains to be noted: Taking full advantage of an EDA requires that the user (and vendor) begin to think of applications and business processes as sequences of events and reactions to those events – a very different programmer and designer mindset. Because of this potential culture shock, initial implementations of EDAs are tending to focus on areas where the notion of event-driven programming has already been accepted and absorbed – such as data management, where transactional applications can now add new event sources for better alerting and more effective responses. In the long run, however, programmers and administrators alike will need to adjust their thinking to a world in which the swiftness of the response, not the performance of the code, is the primary goal of IT.
About Infostructure Associates:
Infostructure Associates is an affiliate of Valley View Ventures that aims to provide thought leadership and sound advice to both vendors and users of information technology. This article is the result of Infostructure Associates' sponsored research. Infostructure Associates believes that its findings are objective and represent the best analysis available at the time of publication.
|Tags: SOA data services and architecture, Enterprise data architecture best practices, VIEW ALL TAGS|