Table of Contents
- Inference Rules
- State Machines
- Decision Tables
- Business Rules
- Process Orchestration
- Event Pattern Matching
TIBCO BusinessEvents, unique and core capabilities are listed below
Inference rules provide a mechanism for relating events that represent point-intime data with other events or concepts that represent historical data residing in diverse sources/databases. Inference rules can be written declaratively and forward-chained during execution by a high performance inference engine to generate meaningful actions. Inference rules are used, for example, to relate a customer?s purchase history with perishable inventory and provide real-time offers that are relevant to the customer and yet profitable for the organization.
State machines provide an event-driven and rules-based mechanism for tracking and tracing entity lifecycles. Missing events can be detected through timeouts. For example, they can be used to track and trace orders that go through a number of steps before being fulfilled, to better monitor SLAs at every step by alerting responsible staff in case of a time-out, and to provide real-time status updates to customers.
Decision tables are business-user defined rules that provide a spreadsheet-like interface. They are ideal for creating rules with conditions that are combinations of possible values or ranges of values for a set of parameters, for example: ?If the applicant is female, in the age group 30-40, has 10 years of driving history, resides in the state of California, has never had an accident, has a Master?s degree, owns a regular sedan in the price range of $15K-25K, then the premium is USD 400 per annum.? The actions for each combination can be specified inside cells of the spreadsheet.
Business rules are parameterized rules implemented by business users based on templates defined by technical users. Within TIBCO BusinessEvents, business users have a web-based interface for implementing rule templates. Business rules can be used to, for example, compute loyalty-related discounts based on customer category and total purchases over a period of time, where the discounts change frequently based on the time of the year, what competitors are offering, and other factors. Rules are controlled by the business through a business friendly interface and a flexible hot deployment framework that supports modifying rules as needed.
Event-driven complex straight-through processing can be achieved using BPMN 2.0-compliant processes that provide inference and decision paradigms within the processes. Event-driven process orchestration can be used to, for example, orchestrate the order fulfillment process for selling telecom bundles online where the processes involve computation of a number of rules or decision tables based on the availability of products and services in the area where the customer lives, the product and service combinations selected by the customer, and regulatory requirements in the area where the customer lives.
Temporal relations across event streams can often be expressed through patterns such as sequencing, duplicate detection, store and forward, and others. A simple pattern-matching language can be used to define and enforce patterns. Positive or negative callback functions can be triggered when the pattern is met or not met. Event pattern matching can be used, for example, by a shipping company to collect a minimum number of containers for a particular destination port or wait for a maximum period of time, whichever comes earlier, before ordering shipment.
Both snapshot and continuous queries are supported by TIBCO BusinessEvents. Snapshot queries provide information on the state of objects stored in the inmemory data grid, while continuous queries execute continuous aggregations on event streams and trigger actions through callbacks. Snapshot queries can be used to, for example, reschedule flights that were supposed to land at a particular airport that is being closed due to bad weather. Continuous queries can be used to aggregate faulty power line signals by location code, determine if a fault is likely to affect households, dispatch a technician, and proactively notify customers before they start calling to report the problem.