A short diversion to DAMA (for Semantic CEP)

Reading Time: 3 minutes

TIBCO presented [*1] at this year’s DAMA conference on Semantic Complex Event Processing[*2]; for the rest of the time, with upwards of 10 (!) parallel tracks, it was always easy to find something of interest being presented. Some CEP perspectives came to mind.

Data vs Events

Are events (and event management) a subset of data management, albeit outside of the confines of the database? For example, you can consider an event to be a timed piece of data; there is also operational metadata used in a CEP system (such as Time To Live for an event). Or is event a totally different concept that is upstream of the concept of “data”?

Of course, this question is somewhat hypothetical. In CEP tools like TIBCO’s, we can refer to events in rules, and compare an event’s payload to a data object, or map events into objects for persistence/history. Events have a particular (and different) lifecycle, which tends to be shorter-term than data. Note that data lifecycle is covered by technologies such as MDM (like TIBCO’s CIM).

Data Management Technologies du jour

Its always insightful to see who is sponsoring and exhibiting at these events. For the most part these had some overlap with CEP use cases we have seen:

MDM / Master Data Management: this is really orthogonal to CEP (although the need for Master Event Management seems somewhat unlikely). New master data records can be viewed as events, and combinations thereof could be of interest to the business (i.e. CEP in the BAM role). And of course, event-driven rules can be useful…

ETL / Extract Transform Load and associated data preparation tools: these reminded me of various TIBCO BusinessWorks (“simple” EP) roles for integrating and transforming events (and associated XML payloads), combined with a multitude of real-time adapters for various enterprise systems. And of course rule-driven transformations, where the transformations may be more complex, using rule-driven real-time CEP in tools like TIBCO BusinessEvents. Or some combination thereof. Clearly, this comes down to a build (usiing a tool like BW) vs buy (COTS ETL tool) decision for potential users (although such users should not be surprised if the economics of the build option are surprisingly good, especially if some flexibility is required from the tool and some BW/BE expertise is applied).

Metadata repositories: these are always fascinating, as (to someone like me who is not a data professional) it would always seem to make more sense to keep metadata alongside the data in your operational data stores. Unless we are talking about design-time metadata with no run-time role, in which case we are talking about all sorts of things, like versioning, traceability, role-based access control, high performance over global distances, etc etc.

TIBCO

Good to see a fair number of TIBCO customers here (users of TIBCO SOA/messaging, BPM as well as CEP). Might see them next at TUCON in a few weeks.

Notes:

[1] Our presentation was on Semantic Complex Event Processing, with semantic specialist Sandpiper Software. There is an increasing level of interest in ontology modeling prior to software development, and the way to connect the knowledge-rich ontology world with the performance-capable software world is likely to be ODM (the OMG Ontology Definition Metamodel). It will also be interesting over the next 2-3 years to see how OWL (the W3C web ontology language) and SBVR (the OMG business vocabulary metamodel) will evolve to serve business needs (without nefarious ideology baggage).

Q&A at the session was interesting: questions like “when would TIBCO support ontology mappings for BPM and CEP” ( 🙂 ), healthcare applications, and relevant rule standards.

[2] See also blogs like James Taylor’s (BRE and analytics-focused).