1. A Framework for Object-Based Event Composition in Distributed Systems
Peter Pietzuch and Brian Shand
June 2002 – 12th PhDOOS Workshop, Malaga, Spain

2. Motivation Large-scale systems require new middleware paradigms
need: scalability, loose coupling of components, reliability
O-O middleware (request/reply) limitations:
synchronous, tight coupling, one-to-one communication
New m/w paradigm: Event-Based Communication
require additional, expressive middleware services
Framework for Composite Event Detection
handle complexity of many events
distributed composite event detection (scalability)
large-scale systems require new communication paradigms
scalability, loose coupling, reliability
o-o request/reply limitations: synchronous, tight-coupling, only one-to-one communication
need for new powerful middleware abstractions
event-based communication
composite event detection
distributed composite event detection

3. Overview Introduction to Event-Based Systems
Event-based Middleware Systems
Composite Event Detection
Our Composite Event (CE) Framework
Composite Event Language
Composite Event Objects
Mobile Detection Objects
Distributed Composite Event Detection
Future Work
Introduction to Event-based Systems
Event-based Middleware
Composite Event Detection
Our Framework
Distributed Composite Event Detection

4. Event-based Systems Event Publisher Event Subscriber
notify
Event Publisher
Pub/Sub System
Event Subscriber
subscribe
Notion of an Event:
Examples: PrinterFinishedEvent, SalesOrderEvent
Event Publishers publish events; Event Subscriber subscribe to events with a subscription (filter expression)
asynchronous occurrence in time
contains data that describes the occurrence
can be implemented as an object
notion of an event
event sinks; event sources
publish/subscribe
good properties: asynchronous, loosely-coupled, many-to-many communication

5. Event-Based Middleware
Middleware using the Pub/Sub paradigm
built on top of an existing O-O/M-O middleware
Scalable and fault-tolerant event dissemination algorithms
Properties:
Must be able to:
integrate with O-O app programming language
handle large number of events
Publisher/Subscriber are decoupled
asynchronous notification
many-to-many communication
Middleware using the pub/sub paradigm
Integration with o-o app programming language
Scalable & fault tolerant event dissemination algorithms
must come with traditional m/w functionality +
must come with new pub/sub functionality like composite event detection
 Composite Event Detection

6. Composite Event Detection
Composite Events denote patterns of events
Composite Event Detectors
Example: “Client A submits new order” ; “Sales Dept. authorises order” ; “Plant A rejects order”
Scalability  Distributed detection
Management
Plant A
Plant B
Sales
Department
Plant C
Supplier A
CED
Supplier B
Client A
composite events
composite event detection (finding patterns)
example
helps to deal with scale and complexity
more efficient system (argument from paper)
why is it important to do it distributed?

7. Our CE Framework Design Guidelines Components:
Integration with an O-O middleware
Facilitate distributed detection
Little assumptions about pub&sub mechanisms
Components:
Composite Event Language
Composite Event Objects
Mobile Detection Objects
overview of our framework
nicely integrates with o-o m/w services
composite event objects
mobile detection objects
design guidelines

8. Composite Event Language
Regular-expression based
can be detected by (extended) finite state automata
bounded resource usage
minimal set of operators
Examples
A followed by B, or C (A;B)|C
Any number of A in parallel with B <A*,B>
Distribution:
Decomposable along syntax tree
explicit input alphabet e.g. [A|B in A,B,C]
Operators
Atom A
Concatenation
A;B
Duplication A*
Alternation
A|B
Parallelization
<A,B>
Timing
(A,B)!1s
regular expression based (why?)
show operators
examples
useful properties: decomposable, understood, bound resource usage,

9. Composite Event Objects
Represent composite events
Published by detectors
Example: (A|B);(C;D)
Interface to access primitive event instances
represent composite events
can be manipulated by the o-o programming language
example

10. Mobile Detection Objects
Detect composite events
Are Event Subscriber + Publisher
Encapsulate extended FSM with explicit input alphabet
Example:
Distribution: Agent-like behaviour
migrate to different location to optimise detection
can cooperate by decomposing expressions
contains detector
agent-like behaviour
show interface

11. Distribution Example I
Example: new expression (A|B);(C;D)
notify
A|B
C;D
MDO
MDO
A|B
C;D
Composite Event A B C D
Object (CEO)
- distribution example using mdo from paper
notify
notify
Mobile Detection
MDO
Object (MDO)

12. Distribution Example II
notify
(A|B);(C;D)
MDO
(A|B);(C;D)
notify
A|B
C;D
MDO
MDO
A|B
C;D
Composite Event A B C D
Object (CEO)
- distribution example using mdo from paper
notify
notify
Mobile Detection
MDO
Object (MDO)

13. Distribution Example III
- distribution example using mdo from paper

14. Distributed Detection
Distribution Policy handles
MDO placement and migration (optimise for bandwidth, latency, …)
Degree of CE expression decomposition
Degree of replication (reliability and load-balancing)
Distribution Issues (Network delay)
Out of order arrival of events due to network delays
Best-Effort Detection vs. Guaranteed Detection
Event Publisher failure?
distribution issues
distribution policies
heuristics
take static and dynamic properties of system into account (mention properties)
speculate

15. Future Work Implementation in Hermes Pub/Sub Simulator
investigate different Distribution Policies
performance/cost of the framework?
temporal issues…
Higher-level, domain-specific CE specification languages
user-friendly language
language-binding
GUI
simulation environment
implementation on top of Hermes
distribution results

16. Conclusions Event-Based Middleware Composite Event Detection
good for building large-scale systems
must provide additional services
Composite Event Detection
handle large number of events
should be part of the middleware
Framework for Distributed CE Detection
object-based (CEOs & MDOs)
distributed (CE detectors as agents)
composite event language (extended FSMs)
distribution policies
- summary