Get documents about "EDBPM Haspa
Document Sample
Event-Driven Business Process Management
taking the example of Hamburger Sparkasse
Rainer von Ammon (CITT)
Andreas Hehmann (Haspa)
Rainer von Ammon / Andreas Hehmann 1
Agenda
The forecast of ED-BPM for the next decades
EDBPM – a combination of two disciplines: Business Process
Management (BPM) and Complex Event Processing (CEP)
A reference model for ED-BPM – how its components work together
The business modeller and the event modeller – different
qualifications not in personal union
Event Processing Languages and the problem of standardising
Special challenges from the point of view of database-technologies
First experiences from the Haspa-Projekt:
11 theses about the quality of BPA-models, BPEL-Import in BPMS
and about designing BAM-dashboards
Rainer von Ammon / Andreas Hehmann 2
The forecast of ED-BPM for the next decades
The forecast of Prof. David Luckham…
… we need skilled people at least up to 2050…
… we are only at the end of the period of Simple CEP
Rainer von Ammon / Andreas Hehmann 3
The forecast of ED-BPM for the next decades
The warning of Roy Schulte (VP of Gartner)
since 2006…
…we won‘t have enough skilled people who would be
able to do all the jobs and projects
The statement of Prof. Mani Chandy/California
Techical University at the Gartner Event
Processing Summit 2007…
…The work of IT during the next twenty years will be to
complete the evolution of business processes from
sequences of slow-moving, disjointed applications to more
responsive end-to-end, event-based straight-through flows of
action.
Rainer von Ammon / Andreas Hehmann 4
Managing and monitoring of processes mean ...
representation of the monitoring of time-critical
effects of system bottleneck factors
availabilities and and transaction control
-disturbances
Business
Workflow
Process
Management
Monitoring
process and system- transparency over
linked emergency and integrated represented
disturbance management processes
monitoring of
SLA-compliance
Rainer von Ammon / Andreas Hehmann 5
The whole picture and what it really means:
Monitoring Business Processes and Activities
event cloud with thousands of events per sec…
passwdchange account login
event patterns and complex event processing…
new auto pay
account login
deposit withdrawal new auto pay
transfer
logout account login
enquiry deposit
activity history
account balance
logout
enquiry
passwdchange account login
new auto pay enquiry
account login
deposit withdrawal new auto pay
transfer account login
logout
enquiry deposit
activity history
logout
…e.g. above a bank
The important steps:
1. Redesign the business processes for SOA and BPM
2. Make a SOA, identify services, build WSDL-interfaces…
3. Precise description of patterns of events
4. Detecting patterns in the event cloud
5. Abstraction of complex event pattern instances to higher level events
Rainer von Ammon / Andreas Hehmann 6
The Pain Point: The Event Cloud, the IT-Blindness
and the „Event Tornado“
Often even additional events are needed for BAM and a better Business Insight
startedService_6 startedService_4
startedService_7
Today‘s existing event cloud exitService_4
and thestartedService_5
IT-blindness
exitService_5
exitService_6
startedService_8
exitService_8
exitService_7 startedService_1
Low level events exitService_1
without semantics startedService_2
exitService_2
BusinessProcess1 …
startedService_3
exitService_3
Visualization of the
…
processed/correlated
events via Business BusinessProcessn …
Activity Monitoring
Rainer von Ammon / Andreas Hehmann 7
The Challenge and the Principle of BPM/BAM/CEP
Enterprise cockpit
realize scenario
Domain specific process instances
reference models
Workflow for event patterns set parameters
Modeler
Event
Modeler Monitor / Analyze / Act
workflows Workflow Engine
based on BPEL
Model ^=
Scenario Rules Engine
„special“ SQL
analyse Event AppServer resp. other languages
history…
Store Normalized events, IF …
build business level events AND …
Adapters FOLLOWED BY…
WITHIN…
e.g. RFID, topics of
ACTION
Pub/Sub, …
e.g. payments calculate
Low Level e.g. credit offers (completed) pott,
Event Clouds / offer rates,
profit…
Streams Rainer von Ammon / Andreas Hehmann 8
The business modeller and the event modeller –
different qualifications not in personal union
A proposal for a curriculum of a new Master course of study
“Event-Driven Business Process Management”
Draft V0.1
1. Description: Curriculum together with required credits
and examinations
Certificate: Master of Science
Programme Duration: Four Semesters (120 credits/cr)
Mastercourse-EDBPM-v01.doc
Rainer von Ammon / Andreas Hehmann 9
The business modeller and the event modeller –
different qualifications not in personal union
Concept:
The course consists of the fields of study Business
Process Management, Complex Event Processing,
Business Activity Montoring included Business
Intelligence and Data Warehousing, Computer
Networks, Messaging as well as several application
disciplines like Algorithmic Trading, Supply Chain
Management in the retail domain, fraud detection in
the banking and insurance domain etc. All courses
are completed with course-related tests and Credits
(cr) according to European Credit Transfer System
(ECTS) are awarded. All courses are given in
English and are designed as distance
learning/eLearning courses.
Rainer von Ammon / Andreas Hehmann 10
First lessons learnt from the Haspa project
11 theses about the quality of BPA-models,
BPEL-import into BPMS and about designing
BAM-dashboards
Rainer von Ammon / Andreas Hehmann 11
First lessons learnt from the Haspa project
The long way of standards and the decisions of Haspa
2008
BPDM
BPMN 2.0
Source: Martin Bartonitz/Saperion
Rainer von Ammon / Andreas Hehmann 12
First lessons learnt from the Haspa project
Thesis 1:
1st Thesis
Already the process models - modeled by the 2nd Thesis
operating departments - are imprecise, out of 3rd Thesis
4th Thesis
use or even incorrect: 5th Thesis
6th hesis
(X) true ( ) false ( ) depends 7th Thesis
8th Thesis
9th Thesis
- Haspa process: 10th Thesis
11th Thesis
originally: remodeled:
(eEPK – Notation)
Rainer von Ammon / Andreas Hehmann 13
First lessons learnt from the Haspa project
Thesis 2: 1st Thesis
The process models modeled by the operating 2nd Thesis
3rd Thesis
departments are too coarse, even they are 4th Thesis
correct from the point of view of operating 5th Thesis
6th Thesis
departments: 7th Thesis
8th Thesis
(X) true ( ) false ( ) depends 9th Thesis
10th Thesis
11th Thesis
- Haspa process:
originally: remodelled:
Rainer von Ammon / Andreas Hehmann 14
First lessons learnt from the Haspa project
Thesis 3: 1st Thesis
Already the decision for a modelling tool is 2nd Thesis
3rd Thesis
responsible for different modelling results of 4th Thesis
business processes: 5th Thesis
6th Thesis
7th Thesis
8th Thesis
(X) true ( ) false ( ) depends 9th Thesis
10th Thesis
11th Thesis
Details presented by the hands-on demonstrations
Rainer von Ammon / Andreas Hehmann 15
First lessons learnt from the Haspa project
Thesis 4: 1st Thesis
Human interactions, Peoplelinks, Partnerlinks, 2nd Thesis
3rd Thesis
WSDL‘s, Compensations, Exceptions etc. are 4th Thesis
5th Thesis
modelled and will be updated directly in the 6th Thesis
BPEL – platform: 7th Thesis
8th Thesis
9th Thesis
(X) true ( ) false ( ) depends 10th Thesis
11th Thesis
If true, there will be no way back into the BP-
modelling tool (of a third party vendor).
But it will work for the BP-modeller inside the
same BPM-platform.
- Example WID jPass, objectiF
- but example WID WebSphere Modeler
Rainer von Ammon / Andreas Hehmann 16
First lessons learnt from the Haspa project
Thesis 5: 1st Thesis
If theses 1 and 2 are true, BPEL-export/import 2nd Thesis
3rd Thesis
doesn‘t make sense, because the effort for 4th Thesis
reworking the process in the BPEL-platform 5th Thesis
6th Thesis
would be unacceptably high: 7th Thesis
8th Thesis
9th Thesis
(X) true ( ) false ( ) depends 10th Thesis
11th Thesis
Example WID BPEL-import and
reworking (see hands on demonstration)
Rainer von Ammon / Andreas Hehmann 17
First lessons learnt from the Haspa project
Thesis 6: 1st Thesis
A generic BPEL-export is not possible, 2nd Thesis
3rd Thesis
because of deficits of BPEL2.x (e.g. human 4th Thesis
interaction) and because e.g. Haspa says that 5th Thesis
6th Thesis
there are no processes without human 7th Thesis
8th Thesis
interactions: 9th Thesis
(X) true ( ) false ( ) depends 10th Thesis
11th Thesis
- Example Haspa process already the first
process step is a human interaction
- (see Demonstration of BPEL-exports from
Websphere Modeler, jPass, objectiF…)
Rainer von Ammon / Andreas Hehmann 18
First lessons learnt from the Haspa project
Thesis 7: 1st Thesis
If e.g. theses 1, 2 and 4 are true, a process 2nd Thesis
3rd Thesis
modelling without the IT-department doesn‘t 4th Thesis
5th Thesis
make sense, if the process shall become 6th Thesis
executable: 7th Thesis
8th Thesis
9th Thesis
So, we need a new procedure for modeling 10th Thesis
11th Thesis
processes in the future.
(X) true
( ) false
( ) depends
Rainer von Ammon / Andreas Hehmann 19
First lessons learnt from the Haspa project
Thesis 7 comments: 1st Thesis
2nd Thesis
3rd Thesis
hands-on demonstration: 4th Thesis
5th Thesis
6th Thesis
- For a combined modelling between operating 7th Thesis
8th Thesis
and IT-departments directly in the BPM- 9th Thesis
platform, a very skilled BPM specialist is 10th Thesis
11th Thesis
needed, who is able to hide the complexity of
the platform
- an experienced moderator is needed, who is
able to avoid discussions about unnecessary
parameters and details offered in the UI of the
platform
Rainer von Ammon / Andreas Hehmann 20
First lessons learnt from the Haspa project
Thesis 8: 1st Thesis
If all processes would be „executable“, 2nd Thesis
3rd Thesis
organisation manuals of process models would 4th Thesis
be redundant because executable processes 5th Thesis
6th Thesis
are self-explanatory by the BPM-platform: 7th Thesis
8th Thesis
no „Schrankware“ anymore 9th Thesis
10th Thesis
11th Thesis
( ) true ( ) false (X) depends
Rainer von Ammon / Andreas Hehmann 21
First lessons learnt from the Haspa project
Thesis 9: 1st Thesis
Each process must be measured. Therefore 2nd Thesis
3rd Thesis
each process has (a kind of) a KPI which has 4th Thesis
to be monitored (e.g. in a dashboard): 5th Thesis
6th Thesis
7th Thesis
8th Thesis
9th Thesis
(X) true ( ) false ( ) depends 10th Thesis
11th Thesis
- hands-on demonstration:
WebSphere Business Monitor
(KPI: Key Performance Indicator)
Rainer von Ammon / Andreas Hehmann 22
First lessons learnt from the Haspa project
Thesis 10: 1st Thesis
Each process owner needs a dashboard for 2nd Thesis
3rd Thesis
monitoring his processes. Because an 4th Thesis
enterprise has thousands of processes with 5th Thesis
6th Thesis
millions of process instances, an enterprise 7th Thesis
8th Thesis
will have a lot of dashboards: 9th Thesis
10th Thesis
11th Thesis
(X) true ( ) false ( ) depends
Rainer von Ammon / Andreas Hehmann 23
First lessons learnt from the Haspa project
Thesis 11: 1st Thesis
The C-level management needs an all- 2nd Thesis
3rd Thesis
encompassing Enterprise Cockpit with 4th Thesis
sophisticated drill down features: 5th Thesis
6th Thesis
7th Thesis
8th Thesis
(X) true ( ) false ( ) depends 9th Thesis
10th Thesis
11th Thesis
- Enterprise Cockpit: shows aggregated operating
figures of the whole enterprise, worldwide.
- Drill down features: zoom in geographical areas and / or in
individual processes straight down to a bottleneck
(e.g. a specific role or a employee – but prohibited by law)
Rainer von Ammon / Andreas Hehmann 24
Thanks to the students of the University of Applied
Sciences of Regensburg for preparing the slides and
presentations:
http://www.citt-online.com/index.php?id=veranstaltungen&id3=industrieaufgaben&id4=more
Rainer von Ammon / Andreas Hehmann 25
Thanks for your attention!
Rainer von Ammon / Andreas Hehmann 26
Related docs
Other docs by HC12080802124
12 To issue a limited certificate of operation for any device or equipment subject
Views: 0 | Downloads: 0
DANGEROUS GOODS DECLARATION AND CONTAINER PACKING CERTIFICATE - Download Now DOC - DOC
Views: 24 | Downloads: 0
