enabling value added services in the event-based internet
DESCRIPTION
specific presentation gives an overview of the FP7 project PLAY, along with its main objectives, latest achievements and remaining work. We discuss thoroughly on how to enable value added services in the event-based internet. Specifically, we give details about the development of two dedicated open-source software components that can exploit distributed and heterogeneous event sources for enabling: i) efficiently and dynamically new event subscriptions (i.e. Event Subscription Recommender) and ii) the recommendation and implementation of workflow adaptations (i.e. Service Adaptation Recommender)TRANSCRIPT
1
Enabling Value Added Services in the Event-based Internet
Towards the Social Event-based Internet of Services, 17 October 2012
Dr. Yiannis VerginadisInsitute of Communication and Computer Systems
National Technical University of Athens
2
Agenda
Introduction- Complex Event Processing- PLAY: An Event Marketplace
Value Added Services- Event Subscriptions Recommender (ESR) - Service Adaptation Recommender (SAR)
Conclusion
3
Real-time
Real-time has become one of the crucial characteristics of modern applications and is completely changing the game in the data processing- Data is on the move
o Find results immediately or never one should be informed as soon as his/her flight has a delay
- Information searches for the relevant consumerso instead of searching for information, it should find us
Inform me immediately if my luggage is not onboard and we are about to start (and not after landing)
Google search vs. Twitter followers
Attribution of Slide, Thanks to: Opher Etzion, IBM
4
What time is Real-time Twitter world record, 29. Aug. 2011
- Beyoncé's pregnancy announcement during the MTV VMA show resulted in 8,868 tweets per second.
- The previous record was during the final of FIFA Women's World Cup, between Japan and the United States. That resulted in 7,196 tweets per second
Financial market- Nanoseconds trading
eHealth: Remote patient monitoring- One semantic signal in 5 sec
Energy: Smart meters- One reading in 15 min
Real-time is the business real-time or near real-time
5
CEP - Complex Event Processing An event is something that has happened, or is contemplated as
having happened (Etzion & Niblett 2010) Event processing is a form of computing that performs
operations on events CEP is an enabling technology that supports on the fly,
(business-) real-time processing of huge event streams CEP is about a timely recognition of the situations of interest and
corresponding reaction- A complex event pattern describes a situation of interest
Pattern detection is one of the notable functions of event processing
EventPatterns
Attribution of Slide,
Thanks to: Opher Etzion, IBM
6
Agenda
Introduction- Complex Event Processing- PLAY: An Event Marketplace
Value Added Services- Event Subscriptions Recommender (ESR) - Service Adaptation Recommender (SAR)
Conclusion
7
Motivating example
Paul is a businessman who has been flying from Paris to New York. He used the entertainment service on board, but hasn’t finished watching the movie before the landing. Two hours later he is entering his room in the downtown hotel he booked earlier and wow: the room entertainment service is ready to PLAY the movie Paul was watching in the plane – of course only the unfinished part.
PLAY
8
Vision
To develop and validate an elastic and reliable architecture for dynamic and complex, event-driven interaction in large highly distributed and heterogeneous service systems.
Such an architecture will enable ubiquitous exchange of information between heterogeneous services, providing the possibilities to adapt and personalize their execution, resulting in situational-driven adaptivity
9
Big picture
service system A
service B
1 2
Event space 3 (service domain B) Event space 3 (service domain B)
newevent
complex events
AFFECTING
Event space 1 (service domain A)Event space 1 (service domain A)
localcontext
Event space 2 (domain A)Event space 2 (domain A)
localcontext
events
events
SENSING
globalcontext
COMBINING
10
Use casesThe first use case is related to the Crisis
Management domain (nuclear accident), which shows the advantage of our platform by increasing situational awareness in a complex crisis situation.
The second use case Socially-aware Smartphone is related to enabling relevant social contacts through mobile devices, by efficient combining real-time and past social-network related events
11
Conceptual Architecture
QoS for Event Processing (ELA - Event Level Agreement)
Massively distributed PubSub
„Endless“ P2P storage
Highly elastic EP
Management of Complex situations
12
Current status
Play Portal: demo.play-project.eu- RDF-based events- Semantic Event Storage- EP-SPARQL2.0 as the query language
o Allows combining real-time and historical events- Many sources connected to the Portal
o Sensor events Cosm (Pachube events from Solar Panels)
o Social media events Twitter („microsoft“), Facebook („jeans“)
o Personal events SmartPhone events
13
14
Agenda
Introduction- Complex Event Processing- PLAY: An Event Marketplace
Value Added Services- Event Subscriptions Recommender (ESR) - Service Adaptation Recommender (SAR)
Conclusion
15
Motivation
Not everything can be captured at design time regarding system’s behavior - but we can predefine the Goals to be pursued and achieved
Unexpected or non-ordinary situations must be identified- from streams of events exchanged in SBA environment- they are usually not considered at design-time
Identified situations; they update and enrich application context Selection of (re)Actions to the unexpected situations at runtime
- exploiting the current application context information
A Mechanism for Dynamic Event Subscriptions is needed- Recommender services for dynamic subscriptions will enable how a service in
the distributed EDA will "decide” which (complex) event to subscribe to.
16
Situation Action Networks (SANs) SAN is a Goal-directed hierarchical modeling approach that
targets automatic search for new goals when specific circumstances, i.e. situational and contextual settings, arise and adjust active capability in order to satisfy the currently active goals.
In goals-directed modeling, goals are typically refined into sub-goals that elaborate how the goal is achieved, creating in such way hierarchical structure
Hierarchical Structure
Step by step “activation” of Goals
17
Event Subscriptions Recommender (ESR)
Event Subscription Recommender (ESR):- undertakes the task of providing event subscription
recommendations based on service context and event semantics.
- recognizes significant situations in order to recommend event subscriptions to new event sources and complex events.
ESR provides added value to services by enabling them to subscribe: - to the “right events”- at the “right time” and- for the appropriate duration, reducing the cost and network
traffic.
“significant situations”: interesting situations referring to
the states of a system along with its environment that create improvement opportunities
critical situations referring to the need for mitigating a risk or hazard.
18
ESR Architecture
19
Agenda
Introduction- Complex Event Processing- PLAY: An Event Marketplace
Value Added Services- Event Subscriptions Recommender (ESR)
o Marine Scenarioo Socially-aware SmartPhone Scenario
- Service Adaptation Recommender (SAR)Conclusion
20
Marine Scenario – AIS Hub (1/2) The International Maritime Organization (IMO) requires all vessels to
carry an AIS (Automatic Identification System) transponder on board,- which transmits their position, speed and course, among some other static information,
such as vessel’s name, dimensions and voyage details.
A vast amount of real time events are available from portals (e.g. http://www.aishub.net/ ) connected to automatic identification systems (AIS) that contain important vessel information worldwide.
21
AIS Adapter
22
AISHub Scenario
Small Boats in Proximity
Speeding in Windy Conditions
23
Safety in high Speed Crafts SAN
24
AISHub Scenario
25
Agenda
Introduction- Complex Event Processing- PLAY: An Event Marketplace
Value Added Services- Event Subscriptions Recommender (ESR)
o Marine Scenarioo Socially-aware SmartPhone Scenario
- Service Adaptation Recommender (SAR)Conclusion
26
“Communicate with a Friend” SAN
SIAFU Simulator
27
“Communicate with a Friend” SAN
This Recommendation will be accepted or rejected by the user (UcTelcoAnswer Event)
This Recommendation (UcTelcoEsrRecom) can be
exploited by another smartphone app
OR it can be handled by ESR (on user’s behalf). This will result into a number of notifications
interesting to the user
28
Socially-aware SmartPhoneContextualized
Latitude„Inform me WHEN there is in proximity a friend of mine, who twitted in the last two days about something that is interesting for me „
Or any other relevant activity in social networks is related to that friend (e.g. Facebook wall update)
29
Agenda
Introduction- Complex Event Processing- PLAY: An Event Marketplace
Value Added Services- Event Subscriptions Recommender (ESR) - Service Adaptation Recommender (SAR)
Conclusion
30
Approach
We aim to suggest workflow or services’ configuration adaptations in order to overcome problems or achieve superior performance- based on significant detected situations
Specifically, for detected situations, we need to be able to define- adaptation points, that is points in a service flow that need to
be adapted as a reaction to a certain situation, and - adaptation actions, indicating what to adapt and how, such as
removal of a problematic service, replacement of an underperforming service, addition of a new service
31
Adaptation Approach
Every time the execution reaches an activity or event node, just before its execution, the BPMN2.0 engine can exchange notifications with the situation reasoning engine - through a bridging component in order to check whether the
specific node is a join-pointThe bridging component should be able to suspend
the execution of that particular process instance and - find the advice that should be weaved in the business
process, using the reasoning serviceAfter the successful adaptation the normal process
execution should resume just after the join-point
32
A Simple Process Adaptation Scenario
Bridging component queries situation reasoning
33
Adaptation Types
BEFORE adaptation type
AFTER adaptation type
AROUND adaptation type
REPLACE
adaptation type
BYPASS adaptation type
REPLACE – AROUND composite adaptation
AFTER – BYPASS composite adaptation
34
Framework Architecture Situation reasoning engine that
handles the execution of SANs- SAN repository- SAN execution component- Event management component and- Context component. (Patiniotakis et
al. 2012)
Adaptation Manager- Monitors the execution of process instances, detects whether execution comes to
an active joint point, checks the associated pointcuts and activates the corresponding advices
Aspects Repository- Stores the definitions of aspects along with their comprising parts
BPMN 2.0 engine- Worktoken Open Source Engine
35
Agenda
Introduction- Complex Event Processing- PLAY: An Event Marketplace
Value Added Services- Event Subscriptions Recommender (ESR) - Service Adaptation Recommender (SAR)
o Nuclear Crisis Management Scenario
Conclusion
36
Nuclear Crisis Management
37
Illustrative Scenario This process is composed of several activities necessary during
emergency situations in nuclear power plants Main actor is the “Representative of the national authority”
- responsible for assessing the current situation and deciding emergency actions e.g., evacuate area, distribute iodine capsules etc.
38
Adaptations in SMP
If ‘study advice’ task takes more than 30 min then weave
Advice 2
If new alerts received before the actor has asked for any advice from the experts yet
then analyze new alerts.
Advice 3 is implemented if the actor has already received the
expert’s advice but has not started to study it yet, while
new alerts have been detected
39
Agenda
Introduction- Complex Event Processing- PLAY: An Event Marketplace
Value Added Services- Event Subscriptions Recommender (ESR) - Service Adaptation Recommender (SAR)
Conclusion
40
Conclusion
PLAY is a global events agreegator and processor, based on three technologies:- Semantics for interopeability- Cloud computing for scalability- Event Processing for real-time/situational
awarenessCurrent development is focused on making the
platform stable and more performant
41
Follow Us
Twitter: @play_fp7RSS: http://news.play-project.eu/
Web: http://www.play-project.eu/
42
Enabling Value Added Services in the Event-based Internet
Towards the Social Event-based Internet of Services, 17 October 2012
Questions?