Sense and Respond Systems for
Crisis Management:
The Event Web
  Mani Chandy
  Computer Science
  California Institute of ...
Outline
 Requirements for crisis management.
 Abstraction: The Event Web.
 Architecture of the Event Web
 Theory
 Prototyp...
The Hypotheses

1.   Information technology for crisis management in
     different application spaces (financial, healthc...
The Hypotheses

1.   Information technology for crisis management in
     different application spaces (financial, healthc...
The Hypotheses

1.   Information technology for crisis management in
     different application spaces (financial, healthc...
The Hypotheses

1.   Information technology for crisis management in
     different application spaces (financial, healthc...
We Cannot Plan for Every Contingency


                                                 •Crisis unfolds in
               ...
Example


                                            •Pilot radios from a plane that a
                                  ...
Dynamic Coalitions of Multiple Groups
Manage Crises

                                              Groups that manage cris...
The Role of Time and Place

                                     The response of an
                                     o...
Example

 Time: Moving-point average of typhoid cases for 2 day
 window exceeds moving-point average over 1-month
 window,...
Crisis Managers Need Attention Multipliers




   Respond                                                              Sen...
Architecture
            Source of raw data:
                e.g., sensor




Event processor:                            ...
Architecture
                                                     Nassau County
                   Navy                   ...
Central idea of architecture of Event Web

 Compositional.
 Extensible component library.
 Components can be added, delete...
Outline

 Requirements for crisis management.
 Abstraction: The Event Web.
 Architecture of the Event Web
 Theory
 Prototy...
Theory (time permitting)

Semantics of events
  Error analysis
  Abstraction:
     The programming paradigm: when-then rul...
Types of events


                                                     Temperature in Minneapolis



An agent B generates ...
Events that changes estimates




                                                                    Actual

 An agent ge...
Events that change estimates

Agent generates an event to
  change “knowledge” or
“estimates” of other agents.
           ...
Events that change estimates

     In this case, an event is
encoded as a timestamp and a
bit indicating whether the band
...
Types of events: periodic events


                                                      Temperature in Minneapolis



An ...
Errors due to delay



                                                             delay
   Receiving agents’            ...
Errors due to delay


                                                                     now
                           ...
Errors

 Errors due to:
     loss of events
     delays
      clock drift
    erroneous monitoring of environment
    erro...
WHEN f(global_state) THEN generate event
•World is a multidimensional system with a huge number
of dimensions.
• Informati...
The Event Web implements When-Then
                                                Source of raw data,
                   ...
Stepwise Refinement

 Crisis manager specifies global when-then rules.
 Stepwise refinement to:
   Sensors that generate r...
Delta dataflow for executing when-then
Rajit Manohar, Cornell: building asynchronous VLSI
              chip for delta dat...
Notations for When – Then Rules

 Alternatives

 Same as query for a database: a rule is a persistent
 query.
    SELECT s...
Incremental Compilation: Map Tree to
Existing Node Structure on Some Machine

                                            ...
Incremental Computation




                      f(x1, …, xk, … xn)

                                                f


...
Incremental Computation




New value: f(x1, …, xk’, … xn)

                        f(x1, …, xk, … xn)

                  ...
Incremental Computation

                                                              Ideally, cost of computing
        ...
Outline

 Requirements for crisis management.
 Abstraction: The Event Web.
 Architecture of the Event Web
 Theory
 Prototy...
Prototype and product

 Goal: build a platform for implementing when-then rules.
 Show how global when-then rules can be r...
Generic asynchronous node in Event Web


                                                          Timer


     From event...
Prototype and product

 Event schema:
   Caltech Infospheres prototype: XML document
   iSpheres Corp product: Serialized ...
Processing an Incoming Message



Incoming message


                Current state
                (XML doc)




     Sens...
Processing an Incoming Message



Incoming message
                                                    Choose XSLT transfo...
Processing an Incoming Message



Incoming message
                                                    Choose XSLT transfo...
Architecture of event processor
                                                          Incoming events


              ...
Related Work

 Events: David Luckham, iSpheres patent apps
 Content-based subscription: Gryphon, Sienna
 Real Time Busines...
Outline

 Requirements for crisis management.
 Abstraction: The Event Web.
 Architecture of the Event Web
 Theory
 Prototy...
Upcoming SlideShare
Loading in …5
×

Sense and Respond Systems for Crisis Management: The Event Web

1,122 views

Published on

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,122
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
10
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Sense and Respond Systems for Crisis Management: The Event Web

  1. 1. Sense and Respond Systems for Crisis Management: The Event Web Mani Chandy Computer Science California Institute of Technology mani@cs.caltech.edu http://www.infospheres.caltech.edu/crisis_web/executive-summary.html Infospheres Group: Dan Zimmerman, Brian Aydemir, Elliott Karpilovsky, Andre Khorlin, Lu Tian, Jonathan White iSpheres Corp.: Eric Bax, Greg Billock, Lee Rosen, Ian Swett 1
  2. 2. Outline Requirements for crisis management. Abstraction: The Event Web. Architecture of the Event Web Theory Prototype. Dagstuhl Workshop on Crisis Management: Feb 2004 Work at Caltech and at iSpheres Corporation http://www.ispheres.com http://www.infospheres.caltech.edu/crisis_web/executive- summary.html Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 2
  3. 3. The Hypotheses 1. Information technology for crisis management in different application spaces (financial, healthcare, natural disasters, terrorism) have a common set of problems. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 3
  4. 4. The Hypotheses 1. Information technology for crisis management in different application spaces (financial, healthcare, natural disasters, terrorism) have a common set of problems. 2. A layer can be abstracted that represents these common problems. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 4
  5. 5. The Hypotheses 1. Information technology for crisis management in different application spaces (financial, healthcare, natural disasters, terrorism) have a common set of problems. 2. A layer can be abstracted that represents these common problems. 3. Computer science theory and implementation practices can be applied to this layer. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 5
  6. 6. The Hypotheses 1. Information technology for crisis management in different application spaces (financial, healthcare, natural disasters, terrorism) have a common set of problems. 2. A layer can be abstracted that represents these common problems. 3. Computer science theory and implementation practices can be applied to this layer. 4. Software platforms for this abstraction can be tailored quickly to map to different crisis situations. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 6
  7. 7. We Cannot Plan for Every Contingency •Crisis unfolds in unexpected ways. Requirement: Configure sense and respond platform when crisis strikes; continuously reconfigure platform as crisis evolves. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 7
  8. 8. Example •Pilot radios from a plane that a passenger seems to have SARS symptoms. • The plane crashes. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 8
  9. 9. Dynamic Coalitions of Multiple Groups Manage Crises Groups that manage crises are changing collections of institutions and individuals. The public, connected through the Internet and wireless, plays an increasingly important role. Requirement: Flexible dissemination network without central data but with dynamic security policies. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 9
  10. 10. The Role of Time and Place The response of an organization often depends on the history of a crisis, and not merely on its instantaneous state. Location plays an important role. Relevance: respond to state trajectories and not merely to current state. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 10
  11. 11. Example Time: Moving-point average of typhoid cases for 2 day window exceeds moving-point average over 1-month window, and, supply of antibiotics below threshold T. Location – Geographical information systems: Alert all vehicles moving towards a bridge. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 11
  12. 12. Crisis Managers Need Attention Multipliers Respond Sense Invoke distributed services Detect events across in real-time extended environment in real-time Analyze Aggregate events across multiple sources Friendly forces in area. I am under attack Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 12
  13. 13. Architecture Source of raw data: e.g., sensor Event processor: Event dissemination Influence environment: Actuators Event stream directory Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 13
  14. 14. Architecture Nassau County Navy Hospital TWA Event Nassau County CIA dissemination Police Travel agent John Q. Public Event stream directory Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 14
  15. 15. Central idea of architecture of Event Web Compositional. Extensible component library. Components can be added, deleted, modified during a crisis. Components are loosely coupled. Communication between components: Events. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 15
  16. 16. Outline Requirements for crisis management. Abstraction: The Event Web. Architecture of the Event Web Theory Prototype. Dagstuhl workshop on Crisis Management Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 16
  17. 17. Theory (time permitting) Semantics of events Error analysis Abstraction: The programming paradigm: when-then rules. Stepwise refinement global when-then rules to sets of sensors, actuators, and event processors with local when-then rules. Event notations Delta dataflow algorithm and chip design Incremental computation to process event streams. Optimization: intelligent sensors, capacity limitations,… Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 17
  18. 18. Types of events Temperature in Minneapolis An agent B generates an event to change other agents’ estimates of the state of the environment. Time Model used by agent B of daily variation on Nov. 3 Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 18
  19. 19. Events that changes estimates Actual An agent generates an Model band event to change other agents’ estimates of the state of the environment. Time Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 19
  20. 20. Events that change estimates Agent generates an event to change “knowledge” or “estimates” of other agents. event Time Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 20
  21. 21. Events that change estimates In this case, an event is encoded as a timestamp and a bit indicating whether the band event was crossed above or below Time Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 21
  22. 22. Types of events: periodic events Temperature in Minneapolis An agent samples values according to a specified sampling strategy Time The model deals with the sampling strategy Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 22
  23. 23. Errors due to delay delay Receiving agents’ now estimation of state when no signal is received. Time Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 23
  24. 24. Errors due to delay now delay Receiving agents’ Estimated estimation of state when distribution no signal is received. of current temperature Time Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 24
  25. 25. Errors Errors due to: loss of events delays clock drift erroneous monitoring of environment erroneous computation Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 25
  26. 26. WHEN f(global_state) THEN generate event •World is a multidimensional system with a huge number of dimensions. • Information about different dimensions represented by discrete events generated at different times. Generate event when Danger point system enters this space t=3 t=5 t=7 t=9 Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 26
  27. 27. The Event Web implements When-Then Source of raw data, Actuators e.g., sensor Event Dissemination Apps Network Sets of When -Then Rules Event processors Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 27
  28. 28. Stepwise Refinement Crisis manager specifies global when-then rules. Stepwise refinement to: Sensors that generate raw events for the “when” predicate. Event processors that execute local when-then rules. Actuators that implement the final “then” action. So that collectively the distributed system implements the global when-then rules. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 28
  29. 29. Delta dataflow for executing when-then Rajit Manohar, Cornell: building asynchronous VLSI chip for delta dataflow At Least 2 Moving – point average f Avg > 2-min avg > Threshold 60 min avg Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 29
  30. 30. Notations for When – Then Rules Alternatives Same as query for a database: a rule is a persistent query. SELECT s WHERE b General structure dealing with histories, geography, action sequences (e.g., time series, location sensitive, workflow representation) Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 30
  31. 31. Incremental Compilation: Map Tree to Existing Node Structure on Some Machine c c c c c c c T>0 P>1 P T Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 31
  32. 32. Incremental Computation f(x1, …, xk, … xn) f x1 xk xn Change in value xk’ Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 32
  33. 33. Incremental Computation New value: f(x1, …, xk’, … xn) f(x1, …, xk, … xn) f x1 xk xn Change in value xk’ Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 33
  34. 34. Incremental Computation Ideally, cost of computing f(x1, …, xk’, … xn) from f(x1, …, xk, … xn) should be New value: f(x1, …, xk’, … xn) independent of n. f(x1, …, xk, … xn) Examples: •Threshold functions •Moving-point computations f •Boolean operators x1 xk xn Change in value xk’ Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 34
  35. 35. Outline Requirements for crisis management. Abstraction: The Event Web. Architecture of the Event Web Theory Prototype. Dagstuhl workshop on Crisis Management Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 35
  36. 36. Prototype and product Goal: build a platform for implementing when-then rules. Show how global when-then rules can be refined automatically, or systematically by engineers, to obtain implementations. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 36
  37. 37. Generic asynchronous node in Event Web Timer From event dissemination network Message- triggered program Event inboxes To event dissemination network Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 37
  38. 38. Prototype and product Event schema: Caltech Infospheres prototype: XML document iSpheres Corp product: Serialized Java object. Event dissemination Infospheres: explicit subscription to event generator by event consumer iSpheres prototype: subscription deduced from high-level notation. Event notation Infospheres: only provides compositional structure iSpheres: explicit notation for parametric and time series analysis. Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 38
  39. 39. Processing an Incoming Message Incoming message Current state (XML doc) Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 39
  40. 40. Processing an Incoming Message Incoming message Choose XSLT transformation based on incoming message and current state Current state (XML doc) New state (XML doc) Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 40
  41. 41. Processing an Incoming Message Incoming message Choose XSLT transformation based on incoming message and current state Current state (XML doc) Sequence of outgoing Messages (XML docs) New state (XML doc) Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 41
  42. 42. Architecture of event processor Incoming events queues Data structures shared by event queues and when- then rules engine Rules execution Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 42
  43. 43. Related Work Events: David Luckham, iSpheres patent apps Content-based subscription: Gryphon, Sienna Real Time Business Intelligence Real Time Business Rules Active databases Gartner: Roy Schulte on Event-Driven Architecture Download the iSpheres Sense and Respond platform from http://www.ispheres.com Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 43
  44. 44. Outline Requirements for crisis management. Abstraction: The Event Web. Architecture of the Event Web Theory Prototype. Dagstuhl workshop on Crisis Management Work at Caltech and at iSpheres Corporation Download product: http://www.ispheres.com http://www.infospheres.caltech.edu/crisis_web/executive- summary.html Sense and Respond Systems for Crisis Management 3 November 2003, Agent Based Modeling and Simulation 44

×