Complex Event Processing

1
Complex Event Processing
2
Remember This?

• 50 million people affected
• Nuclear power plants in New York and Ohio shut
  down
• Air traffic was slowed as flights were halted

3
Why?

• Electric Power Research Institute (EPRI) White
  Paper Findings

• Lack of understanding of VAR reserves in the
  region and the adjacent regions, coupled with the
  possibility of inadequate dynamic VAR support
  available from generators. (VAR is shorthand for
  reactive power, which is the additional power
  required for maintaining voltage stability when
  serving certain kinds of load, such as motors,
  air conditioning, and fluorescent lights.)
• Insufficient visibility of power flow
  conditions over the entire region, coupled with
  inadequate coordination, control and
  communication of the power system on a regional
  basis. That information is flowing from 27
  distribution feeds, eight re-closure or safety
  switches and 4,192 transformers.
• Insufficient understanding of the potential
  impact on August 14, 2003 of new power flow
  patterns caused by increased wholesale power
  transfers resulting from industry restructuring.
• Lack of real-time regional and interconnection-wid
  e power flow models for anticipating changing
  flow patterns and the formation of new
  bottlenecks.

4
Enterprise 2.0 (80s 2000)
Enterprise 1.0 (60s 80s)
Enterprise 3.0 (2000 2020)
Data Processing
Predictive
Client Server
Mainframe
Database
ESB
Building Block
3-Tier
N-Tier
2-Tier
Software
Online
Event Driven
Batch
Velocity
000,000s
000,000,000,000s
000s
Interactions
5
Enterprise 2.0 (80s 2000)
Enterprise 1.0 (60s 80s)
Enterprise 3.0 (2000 2020)
Half Life of Data
Amount of Data
Time to React
Data Processing
Predictive
Client Server
Mainframe
Database
ESB
Building Block
3-Tier
N-Tier
2-Tier
Software
Online
Event Driven
Batch
Velocity
000,000,000s
000,000,000,000s
000,000s
Interactions
6
The Data Advantage
Threats Opportunities
Events
Enterprise 3.0
Transactions
Enterprise 2.0
7
Understanding The Event Cloud
Enterprise Events
8

• Deriving Value from Events

Value
Enterprise Events
9
CEP enables More Effective Decisions
Business Automation ? Enables Fast Response
Effective Real-time Decisions
Business Event
Potential Business Value
Action time
10
Using Real Time Events A Better Way to Fly

• Agenda
• Overview and vision for Complex Event Processing
• Real-time customer problems and solutions
• TIBCO BusinessEvents capabilities
• Answer your questions

11
Complex Event Processing (CEP) Premises
Problem There is always a Plan. It Never Goes
According to the Plan. A Pattern of Events is
an indicator of opportunities (inventory
re-allocations). A Pattern of Events is an
indicator of threats (fraud, inventory
shortages). A Pattern of Events can be reactive
situations. A Pattern of Events can be
predictive (early warning) situations. Solution CE
P is the foundation technology for detecting
patterns in real-time event streams correlated
with historical data.
12
Velocity drives need for Event Processing
Business Cycle-Time Improvements
Perform Trading Analytics
30 minutes
20 minutes
Airline Operations
Handle Call Center Inquiries
8 hours
Track Financial/Risk Position
1 day
Supply Chain Updates
1 day
Document Transfer
3 days
Phone/Service Activation
3 days
Refresh Data Warehouse
1 month
Settle Trade/Transaction
5 days
Build-to-Order Product
6 weeks
Source Gartner
13
Operational Intelligence in Transport Logistics
13
14
Context-based Decisions in Operations
Aircraft Operations
Dispatch, Weather
Crew Management
15
Events driving Airline Scenarios

• March Madness Example
• An abnormally large number of no shows in one
  day could mean a mass of extras tomorrow
• Station alerting for passenger patterns, monitor
  check-in patterns and no-show rate
• Flow Rates in/out of airport
• Only X aircraft getting out of airport, but YgtX
  aircraft are arriving there
• Prevent/avoid gridlock, compare gate and off-gate
  parking capacities numbers of passengers
  transferring to which flights

16
The Goal Total Situational Awareness

• Require
• Better visibility into aircraft operations across
  all stations
• Apply rules to all inputs and report situations
  that are out of tolerance
• Display alerts in a meaningful way to inform not
  confuse
• Motivation

17
Complex Event Processing Solution
Dispatch, Weather
TIBCO BusinessEvents Match events rules
18
Results
Aircraft loading 6.25 flights/day/aircraft 544 aircraft
Flight Passengers gt86M passengers
Aircraft Event Volume gt60K Msgs/Day
Customer satisfaction Lowest ratio of complaints per passengers boarded of all Major U.S. carriers Diamond Award, at the Air Cargo World awards ceremony, with the highest overall airline score leading all airlines in the Performance and Value categories. Favorite Domestic Airline and ranked 1 in Best Customer Service, Best Airfare Prices, Best On-Time Service, Best Baggage Service, and Best Value Frequent Flier program, among others, in the Readers Choice Awards by Smarter Travel. ...
18
19
Union Pacific Railways - Challenges

• Pre-existing Business Challenges
• Less than half of all network schedules achieved.
• Track velocity 17 mph. Goal is 20 mph every
  increase in 1 mph 10 million saving to the
  bottom line (each month!).
• Customers over-schedule and use crews unwisely.
• Appear at capacity but a system of over-schedule.

20
Union Pacific Railways - Benefits

• Maintenance Crew Management
• Outsourcer instantly knows where to deliver new
  crews
• Automatically exchange locomotive maintenance
  order, fulfilment and invoicing data with service
  providers and business partners
• Trains run 1/8th mile per hour faster!
• Way Side Detection
• Real-time Alerting of wayside diagnostic alerts
• Service Enabled Maintenance System
• Pro-actively fix wheels with faulty bearings

21
TIBCO BusinessEvents v5.0 The 5th Generation
Event Platform
22
BusinessEvents - Overview
BE is a framework for messaging based
intelligent, distributed agents
2. REASON Define rules of behavior to reason on a
pattern of objects, events and time
Rules
Patterns
4. SITUATION Visualize
Predictive Enterprise
Actions
3. INFER Actions
1. SENSE Ability to listen to events and gather
data on demand.
Events
Objects
23
BusinessEvents State-Oriented CEP Behavior
State Model
State 1
State 2
24
Concept Model

• Stateful object modelling.
• Easily visualize object relationships.

25
State Model / Process Flow

• UML-Compliant state model.
• State diagram / flow diagram is simple to follow
  and maintain.

26
BusinessEvents Rule-Oriented CEP Behavior
Rule
Complex Patterns
Reaction
27
Inference Rule Features

• High Performance Pattern Matching
• Large catalogue of functions built in.

28
Decision Manager Features

• Rules managed in a decision-table interface.
• Export/import to and from Excel.

29
BusinessEvents Query-Oriented CEP Behavior
Query
Complex Query
Results Event
30
Query CEP Agent Features

• Common query language
• Based on SQL/OQL
• Familiar to SQL users.
• Defined dynamically or statically.
• API similar in structure to JDBC.

select city, count() from LoanApp policy
maintain last 7 days where amount gt 350000
group by city
31
BE Views

• Provides visibility into real-time business
  events
• Empowers users to make effective real-time
  decisions and appropriate actions on critical
  opportunities and issues.

32
TIBCO BusinessEvents Thematic Summary
events
TIBCO BusinessEvents
Complex Event Processing(CEP, ESP)
Decision Management(business rules)
Real-timeDashboards(real-time Inter-active BI)
Straight Thru Processing(real-time process
eventing)
33
Application to SITS

34
Enterprise 2.0 in Action
35
Enterprise 3.0

• It's so smart that it knew there was a power
  outage in one neighborhood 34 minutes before the
  first resident called the utility.
• It's so smart that the number of customer-voltage
  complaints about either surges or drops went
  from 70 to zero.
• It's so smart that it identified a transformer
  that was overloaded and needed to be replaced
  before it got fried.
• In the past, the utility knew to replace
  transformers when they blew and lights went out.
• The next step to collect and share information
  with about 25,000 homes and businesses that have
  installed or will install "smart meters.

36
Sample Event Processing Usage Summary

• Adaptive Marketing
• Pattern Capture opportunity with customer while
  the window is open.
• Telco
• SLA (Service Assurance)
• Real Time Service Offers and Analytics
• Finance
• Fraud Detection
• Track and Trace Trades/Deals/Settlements
• Pre/Post trade exceptions
• Logistics - Track Trace
• Track Packages against a Plan. Infer package
  delays in a proactive manner. Alert customers.
• Government
• Track and Analyze patterns that were otherwise
  very difficult to detect
• Dynamic Resource Scheduling
• Real Time Optimization of Resources against a
  Plan.

37
Next Steps

• Email us to receive High Consequence Architecture
  White Paper
• For more information  or to arrange a meeting to
  discuss your business problem  contact
• Atif Chaughtai
• achaught_at_tibco.com
• [phone]

38
Questions?