WIXLIB

ODM and Rules - Semantic EnabledComplex Event ProcessingPaul Vincent, Business Optimization Group, TIBCO Software 2009 TIBCO Software Inc., Sandpiper. Software Inc. All Rights Reserved.

Presenter Paul Vincent CTO, Business Rules and Complex Event Processing Contributor to standards (OMG PRR Co-Chair, W3C RIF) Contributor to Event Processing research– EPTS Reference Architecture Working Group co-chair– EPTS Metamodelling Working Group co-chair Co-author http://tibcoblogs.com/cep/ TIBCO Software Largest independent software integration company 3,000 customers in 40 countries using SOA, BPM and Business Optimization Complex Event Processing one of the fast growing trends 2009 TIBCO Software Inc., Sandpiper Software, Inc.2

Agenda1. Complex Event ProcessingWhat is it and where does it fit in theIT and semantics worlds?2. Semantic Processing and Real-time Event ProcessingHow can semantics assist inreal-world, real-time event processing? 2009 TIBCO Software Inc., Sandpiper Software, Inc.3

Real-world EventsCustomerLogonBase RateIncreaseOrderedItemArrives inStoreCustomerChecks“Close Account”Web PageNewLiabilityAddedRentalCarCrashed 2009 TIBCO Software Inc., Sandpiper Software, Inc.Mobile Callfrom ContractReturnedthru EDI4

Where is the information?CustomerLogonBase RateIncreaseOrderedItemArrives inStoreCustomerChecksClose AccountWeb PageNewLiabilityAddedRentalCarCrashed 2009 TIBCO Software Inc., Sandpiper Software, Inc.Mobile Callfrom ContractReturnedthru EDI5

Where is the information?FraudRisk!Risk ofCustomerDefectionStock Capacitytrending ApproachedRentalContractComplete 2009 TIBCO Software Inc., Sandpiper Software, Inc.Cell phonefraud alertCustomernow ratedGoldChange inProduct SalesTrendContractValidatedContractValid6

The Event Cloud What meaning can we derive from the increasing“cloud of events”? Can we infer important business eventsby correlating events automatically earlier,regardless of source / type?HR eventsRFID eventsSupply chain eventsSLA eventsDelivery events 2009 TIBCO Software Inc., Sandpiper Software, Inc.Transport events7

Complex Event ProcessingSense andRespondTrack and TraceSituationAwareness 2009 TIBCO Software Inc., Sandpiper Software, Inc.8

What Does CEP Solve?Warnings precedethreatsPotential Business ValueBusiness EventResulting Complex Event MeasuredCEP providesquicker responseto complexeventsRoot Cause / CorrelationCorrective Decision MadeAction TakenAction time 2009 TIBCO Software Inc., Sandpiper Software, Inc.9

Implementing Complex Event ProcessingAccess andMonitorthe “Event Cloud”EventsDefine complex eventsacross events and existingdataStatesandTransitionInferencesRulesSets andQueriesEventvia JMS, RV,MQ, TCP/IP,etc 2009 TIBCO Software Inc., Sandpiper Software, Inc.InformationEventand orageContinuouslyprocess eventsusingprocedural anddeclarativeeventprocessingelements

Sample “IT Models” used in CEPEvent Model and Concept Modelfor static event and concept relationshipsUMLClassUMLEventUMLState Modelfor dynamic, time-based concept lifecyclesStateQuery Modelfor sets and windows of events andconceptsUMLRule Modelfor patterns of events and conceptsPRRDecision Modelfor managed decision tables 2009 TIBCO Software Inc., Sandpiper Software, Inc.

OMG MDA and Class/Object/Data ModelsOMGW3CSBVRComputationIndependentModels (CIM)Semantics for BusinessVocabularies and RulesOWLWeb Ontology LanguageODMOntology Definition MetamodelPlatformIndependentModels (PIM)PlatformSpecificModels (PSM)UML2Class ModelsWithplatform-specificextensionsSUN Java 2009 TIBCO Software Inc., Sandpiper Software, Inc.SQLMS .NETW3C WSDLW3C XMLW3C RDF

OMG MDA and Rule ModelsOMGW3CSBVRComputationIndependentModels (CIM)Semantics for BusinessVocabularies and RulesOWLWeb Ontology LanguageOCLPlatformIndependentModels (PIM)Object Constraint LanguagePRRProduction RuleRepresentationRIFRule Interchange FormatPlatformSpecificModels (PSM)TIBCO 2009 TIBCO Software Inc., Sandpiper Software, Inc.PegaJESSDROOLSILOGBlaze

MDA: OMG PRRX Formal UML model for production rules Defined in UML Extends UML so production rules are1st class citizens alongside objectsobjectsX Vendor-neutral UML-friendly rulerepresentation eslruRules specified via tools, not manually!X 2 rule “semantics” (types):1. Forward chaining inference rules(e.g. Rete-model)2. Sequentially processed procedural rules(e.g. scripts)X Import/export for rule modeling 2009 TIBCO Software Inc., Sandpiper Software, Inc.XMI between UML tools and BREs

PRR metamodel Ruleset collection of Rule Rule is (for RuleVariables) if Condition then Actions 2009 TIBCO Software Inc., Sandpiper Software, Inc.

Agenda1. Complex Event ProcessingWhat is it and where does it fit in theIT and semantics worlds?2. Semantic Processing and Real-time Event ProcessingHow can semantics assist inreal-world, real-time event processing? 2009 TIBCO Software Inc., Sandpiper Software, Inc.16

Assumptions (1) Most IT processing uses conventional, “fixed”IT models Knowledge mapped to structured object-oriented structures that runin JVM etc efficiently: changes require recompilation Moving to knowledge-based models (e.g. RDF data) for existingapplications is too expensive (abstraction, runtime, performance) New IT management capabilities sometimes use RDF/OWL tosupport dynamic enterprise views & reduce application change time 2009 TIBCO Software Inc., Sandpiper Software, Inc.17

Assumptions (2) Knowledge-based solutions may be most valuablewhen dealing with change / changeable entities /discovery or where flexibility is essential Business intelligence / discovery activities Complex cross-domain / cross-organizational information-basedservice delivery Software system development and maintenance 2009 TIBCO Software Inc., Sandpiper Software, Inc.18

Assumptions (3) Mitigated today in conventional IT systems throughtechniques like Declarative production rules BPM Event driven architecture (type of SOA) 2009 TIBCO Software Inc., Sandpiper Software, Inc.19

Assumptions (4) Semantics help in the “software system – person”boundaries, to augment conventional approaches,increase scalability of rule sets, or where reusepotential is high 2009 TIBCO Software Inc., Sandpiper Software, Inc.20

DefinitionsAn ontology specifies a rich, updatable and verifiable description of the Terminology, concepts, nomenclature Properties explicitly defining concepts Relations among concepts (hierarchical and lattice) Rules to distinguish concepts, refining definitions and relations (constraints, restrictions,regular expressions)relevant to a particular domain or area of interest. 2009 TIBCO Software Inc., Sandpiper Software, Inc.

Ontologies versus IT ModelsFormalOntologyIT levels ofontology supportUMLODMUMLClassUMLEventUMLState 2009 TIBCO Software Inc., Sandpiper Software, Inc.UMLPRRQueries22

Ontologies driving CEP (1)FormalObject and event,Ontologyinheritance,containment, &referenceUMLODMOWLUMLEvent Model and Concept ModelClassfor static event andconceptrelationshipsKnowledge nowledge ofUMLconstrained sets, Statefor dynamic, time-basedconcept lifecyclescollectiondefinitions;May changeQuery ModelKnowledge oftime andfiltered behaviorsfor sets and windowsoverof eventsconceptsacross sets,includingdynamicUMLclassificationsPRRRule Modelfor patterns of events and conceptsDecision Modelfor managed decision tables 2009 TIBCO Software Inc., Sandpiper Software, Inc.

Ontologies driving CEP (2)OWLFormalOntologyUMLODMUMLUMLEvent Model and Concept ModelClassEventfor static event and concept relationshipsSemantic processing ofevent information,leadingUMLtoState ModelStatefor dynamic, time-based concept lifecycles new event subtypes, new classifications,Query Model updated / new setfor sets and windows of eventsanddefinitions,concepts updated / newproduction rules,UMLRule Model updated / PRRnewfor patterns of eventsand conceptsdecisionsDecision Modelfor managed decision tables 2009 TIBCO Software Inc., Sandpiper Software, Inc.

Semantic CEP Architecture exampleBusiness EventMeta-PatternsKBEventSourcesSemantic AgentEvent Reclassification& Re-aggregationTrend KBEvent BusTrend Analysis& Machine LearningEventBusUpdate LogicEventHistoryCEPEventConsumers 2009 TIBCO Software Inc., Sandpiper Software, Inc.State EngineState ModelInferenceRule EngineRulebaseQuery EngineQueries25

Example Semantic CEP roles Update object model and associated metadata(time to live, history depth, etc) Update rule parameters(new / revised classes and subclasses to look for,attribute ranges that are significant, etc) Update state model(transition rule values,wait times for missing events,new conditions,eliminate invalidated states, etc) 2009 TIBCO Software Inc., Sandpiper Software, Inc.26

Example Semantic CEP Use Cases Call Center / CRM Operations to identify conflicting Client Advisories Intelligence Analysis supporting research operations Semantically enhanced Fraud Detection and Financial Regulation IP Content Publication & Management for Media ontologyClass procedureSemantically-Enhanced Search / Retrieval- Siderean, SchemaLogic, SDI CorporationPublish / Subscribe, Agent-Driven User AccessPreference / Role-based Custom Delivery ontologyClass Algorithm ontologyClass Congestion Control Algorithm ontologyClass Forwarding Algorithm ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass TCP New Reno TCP Westwood Flow Control AlgorithmEqual Cost Multipath RandomWeighted RandomBinary Increase Congestion (BIC)High Speed TCP (HSTCP) ontologyClass Load Balancing Algorithm ontologyClass ontologyClass Least Connection SchedulingShortest Expected Delay Scheduling ontologyClass Quality of Service (QoS) Algorithm ontologyClass ontologyClass ontologyClass Class Based QueueingHierarchical Token Bucket FilteringRandom Early Drop ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass ontologyClass Differentiated ServicesExplicit Congestion Notification TCP HyblaTCP VegasStandard TCP Transactional TCP Interface Round RobinRound Robin Weighted Round RobinLocality Based Least Connection SchedulingWeighted Least Connection Scheduling ontologyClass Token Bucket Filtering ontologyClass ontologyClass Priority Based QueueingStochastic Fair Queueing ontologyClass TCP Westwood PlusDocument Mining & Extraction ServiceReference Vocabulary Drives Extraction- UIMA-based services- Verity, Inxight, other commercialDocument Content Mention /Cross-ReferencePolicy KB 2009 TIBCO Software Inc., Sandpiper Software, Inc.Valid Reference Knowledge BaseDomain Vocabulary (Ontology Components) Reference DataDocument RepositorySchema & IndexArchiveDocumentRepository27

Summary Complex Event Processing a “new kid” on the IT block using high-performance IT capabilities to providea continuous event/data aggregation architecture Semantic Extensions new approaches to bridging the semantic / KR and conventional IT /model-driven worlds convergence with modern IT solutions like CEP 2009 TIBCO Software Inc., Sandpiper Software, Inc.28

Vocabularies and Rules. Platform Specific Models (PSM) OMG. W3C. PRR . Production Rule Representation. RIF . Rule Interchange Format. OWL . Web Ontology Language. Pega. JESS. DROOLS. ILOG. Blaze. OCL . Object Constraint Language. TIBCO. Note RIF is a superset of rule types: mapping OWL vocab rules to RIF and thence to PRR does not make much .