Configurable Services in the Cloud

Supporting variability while enabling cross-organizational process mining

Wil van der Aalst

Acknowledgements

• Marcello La Rosa• Florian Gottschalk• CoSeLoG: Joos Buijs, Jan Vogelaar, Boudewijn van

Dongen, Eric Verbeek, Hajo Reijers.• Marlon Dumas, Arthur ter Hofstede, Niels Lohmann,

Michael Rosemann, Jan Mendling, …• ProM team (www.processmining.org)• YAWL team (www.yawlfoundation.org)

PAGE 1

The need for configurable process models: CoSeLoG project

PAGE 2+/- 430 Dutch municipalities

The need for configurable process models: Suncorp case

PAGE 3

End to end process has between 250-1000 process steps

Sources: Guidewire reference models, GIO CISSS Project, CI US&S P4PI Project

Product Dev

• 25+ steps

Sales

• 50+ steps

Service

• 75+ steps

Claims

• 100+ steps

30variations

500steps

Home

Motor

Commercial

Liability

CTP / WC

Two variants of the same process …

PAGE 4

Variation points

PAGE 5

… in the cloud

Cloud computing

PAGE 6

Traditional Situation

PAGE 7

Processes Organization 1

M1E1

IS1

Processes Organization 2

M2E2

IS2

Processes Organization n

MnEn

ISn

...

IS = Information SystemE = Event logM = Models

ExampleAcknowledgement of an Unborn Child

• Same but different …• “Couleur Locale”• Different from NVVB

models.• Configurable process

models!

PAGE 8

Using SaaS Technology

PAGE 9

Processes Municipality 1

C1

EIS-SaaS

Processes Municipality 2

C2 Processes Municipality n

Cn

...

CM

IS-SaaS = Information System (using a SaaS-based BPMS)E = Event logCM = Configurable ModelsC = Configuration

Process Mining: Before

PAGE 10

Processes Municipality 1

M1E1

IS1

Processes Municipality 2

M2E2

IS2

Processes Municipality n

MnEn

ISn

...

Process Mining: After

PAGE 11cross-organizational process mining

Configuration

Some quotes from Michelangelo• “Every block of stone has a statue

inside it and it is the task of the sculptor to discover it.”

• “I saw the angel in the marble and carved until I set him free.”

• “Carving is easy, you just go down to the skin and stop.”

Michelangelo's David

Positioning of Configuration

Life is about making choices …

PAGE 14

Time and artifacts

• Design time (generic model, i.e., is not released for instantiation)

• Configuration time (specific model, i.e., can be instantiated)

• Instantiation time (specific model + instance)

• Run time (specific model + instance + state/partial trace)

• Auditing time (specific model + instance + full trace)

PAGE 15

Continuum

• In The Netherlands, …• In Brisbane, …• When the sun shines, …• On Sunday, …• When very busy, …• For these customers, …• …

Branching structurePAGE 16

Hiding and blocking

PAGE 17

Action

Blocking

Activating

HidingBlocking

Activate

Hide/skip

Block

Configuration = limiting behavior !

Configurable Process Models

Purchase order created

Service is accepted

Goods receipt posted

Invoice received

V

V

Process Invoice

XOR

Goods receipts

to be settled automatically

Evaluated Receipt

Settlement (ERS)

Invoicetransmitted for vendor’s

records

Material is released

Invoice postedand blocked for release

Invoicing plans require

settlement

Invoicing Plan Settlement

V

Release Invoice

manually

Payment mustbe effectes

V

GUIDELINEERS = ON, if - long term

contract- goods and

conditions are specified

REQUIRED:IPS = ON

⇒ ERS = ON

Consignment/pipeline

liability is created

Consignment/pipeline

liabilities are to be settled

Consignment/ Pipeline

Settlement

V

XOR

V

Consignment/pipeline

settlement document

transmitted

XOR

XOR

Invoice postedand not

blocked for release

Release Invoice

automatically

Purchase order created

Service is accepted

Goods receipt posted

Invoice received

VV

Process Invoice

Material is released

Invoice postedand blocked for release

Release Invoice

manually

Payment mustbe effectes

V

XOR

XOR

Invoice postedand not

blocked for release

Release Invoice

automatically

Blocking HidingConfiguration

C-EPC

C-Petri Net

C-YAWL

C-BPEL

EPC

Petri Net

YAWL

BPEL

a bc

d ef g

h i

jk m

no p

l

a bc

egi

mn

o p

l

C-LTS

LTS

PAGE 18

Variant BSuperclassSuperclass Reference ModelVariant A

a

c

d e

f g

h i

o

a b

c

e

g

i

j

k m

n

o p

l

a b

c

d e

f g

h i

j

k m

n

o p

l

a

c

d e

f g

h i

o

a b

c

d e

f g

h i

j

k m

n

o p

l

a b

c

e

g

i

j

k m

n

o p

l

a b

c

e

g

i

j

k m

n

o p

l

a

c

d e

f g

h i

o

Inheritance Inheritance

Subclass

Configuration Configuration

Inheritance of dynamic behavior

PAGE 19

a

c

e

i

g

i

Configuration Techniques

b

p

d

f

h

l

m

n

j

k

l

o

• Blocking(removing an option)

• Hiding(skipping activities)

a b

c

d e

f g

h i

j

k m

n

o p

l

τ

ττ

Blocking and hiding are the essential concepts of configuration.

“Every block of stone has a statue inside it and it is the task of the sculptor to discover it.”

PAGE 20

PAGE 21

Cross-organizationalmining

PAGE 22Data Mining

Smoker

Drinker

Weight

Short(91/10)

YesNo

Long(30/1)

NoYes

Long(150/20)

Short(321/25)

<81.5 ≥81.5

Process Mining =

Process Analysis

start register initial conditions

check_Aneeded?

check_A

modify conditions

check_Bneeded?

check_B

check_Cneeded?

check_C

assesrisk

declinec1

c2

c3

c4

c5

c6

c7

c8

c9

c10

c11

c12

c13

makeoffer

handleresponse

handlepayment

send insurance

documents

timeout1 timeout2 withdraw offer

c14 c15 c16

c17

(RM,RD)(RM,RD)(E,SD) (E,RD)

(SM,SD) (E,SD)(E,FD)

(E,SD)

(E,SD)

(YE,RD)

(YE,RD)

(FE,FD)

(RM,RD)

+

PAGE 23

Process mining: Linking events to models

software system

(process)model

eventlogs

modelsanalyzes

discovery

records events, e.g., messages,

transactions, etc.

specifies configures implements

analyzes

supports/controls

extension

conformance

“world”

people machines

organizationscomponents

business processes

Example: WMO Harderwijk

• Process related to the execution of “Wet Maatschappelijke Ondersteuning” (WMO) Harderwijk

• Handling WMO applications• WMO: supporting citizens of municipalities (illness,

handicaps, elderly, etc.).• Examples:

• wheelchair, scootmobiel, ...• adaptation of house (elevator), ...• household help, ...

PAGE 24

Event log (796 applications, 5187 events)

PAGE 25

Helicopter view of 1.5 years

PAGE 26

Process discovered using Genetic Miner

PAGE 27

Various representations

PAGE 28

Seamless abstraction

PAGE 29more detailed more abstract

Fuzzy Replay

PAGE 30

Conformance checking using Replay

PAGE 31

= should not have happened but did

= should have happened but did not

Performance analysis using Replay

PAGE 32

Prediction based on Replay

PAGE 33

“Your application is expected to be finalized in 65 days”

From one to many organizations

• More than 80,000 organizations are using Salesforce• More than 1 million organizations are using Google Apps• All 430 Dutch municipalities are implementing the same set of

processes• All 94 U.S. District Courts in the United States share the same

set of workflows• All car-rental offices of Hertz, Avis, …• …

PAGE 34

Consider n organizations

PAGE 35

process 1 eventlog 1

processmodel 1

process 2 eventlog 2

processmodel 2

process n eventlog n

processmodel n

... ... ...

Cross-organizational process mining

PAGE 36

process 1 eventlog 1

processmodel 1

process 2 eventlog 2

processmodel 2

process n eventlog n

processmodel n

... ... ...

(configurable)process model

event log

C

C

C

Pure model-based

PAGE 37

process 1 eventlog 1

processmodel 1

process 2 eventlog 2

processmodel 2

process n eventlog n

processmodel n

... ... ...

(configurable)process model

C

C

C

PM1 + PM2 + … + PMn = CM

Pure log-based

PAGE 38

process 1 eventlog 1

processmodel 1

process 2 eventlog 2

processmodel 2

process n eventlog n

processmodel n

... ... ...

(configurable)process model

event log

C

C

C

α(EL1 + EL2 + … + ELn) = CM

Questions

PAGE 39

process 1 eventlog 1

processmodel 1

process 2 eventlog 2

processmodel 2

process n eventlog n

processmodel n

... ... ...

(configurable)process model

event log

C

C

C

How to merge process models into a single configurable model?

How to discover a configurable model from a collection of event logs?

How to derive the configuration for a process given a configurable model?

How to find and characterize differences among processes using event logs?

How to find and characterize differences using models / configurations?

What are the effects of these differences on the performance of a process?

Evidence-based “best practices”

• Organizations can learn from each other.• Configuration support and diagnostics.• Software vendors/service providers can improve

their products/services.PAGE 40

PAGE 41

Correctness

Remember …

PAGE 42

Action

Blocking

Activating

HidingBlocking

Activate

Hide/skip

Block

Configuration = limiting behavior !

Correctness of configurations

• Question 1: Is a particular configuration correct?

• Question 2:Is there a correct configuration?

• Question 3: How to characterize the set of all correct configurations?

• Question 4:How to auto-complete a configuration?

PAGE 43

Configurable Model + Configuration = Configured model

+ =

Can t3 be blocked?

PAGE 44

pI

p5

t7

p6

t8

pO

t2

p3

t1

p2

t4t5

p4

t6

p7

pI

p5

t7

p6

t8

pO

t2

p3

t1

p2

t4t5

p4

t6t3

p7

XOR-join

XOR-split

AND-join

AND-split

Flow

Place

Transition

Token

Block t1 and hide t3 ?

PAGE 45

p2

p3

p5

p8

t3

t1 t2

t4

t6

Prepare Travel Form(Secretary)

Prepare Travel Form(Employee)

ApproveTravel Form

(Admin)

RejectTravel Form

(Admin)

Submit Travel Form for Approval (Employee)

Request for change(Admin)

t7Check & Update Travel Form(Employee)

Arrange travel insurance (Employee)

p4

t5 τ

p6

p7

t9t8

p1

XOR-join

XOR-split

AND-join

AND-split

Flow

Place

Transition

Token

p2

p3

p5

p8

t3

t2

t4

t6

ApproveTravel Form

(Admin)

RejectTravel Form

(Admin)

Submit Travel Form for Approval (Employee)

Request for change(Admin)

t7Check & Update Travel Form(Employee)

p4

t5 τ

p6

p7

t9t8

τ

p1Prepare Travel Form(Employee

Block t4 also?

PAGE 46

p2

p3

p5

p8

t3

t1 t2

t4

t6

Prepare Travel Form(Secretary)

Prepare Travel Form(Employee)

ApproveTravel Form

(Admin)

RejectTravel Form

(Admin)

Submit Travel Form for Approval (Employee)

Request for change(Admin)

t7Check & Update Travel Form(Employee)

Arrange travel insurance (Employee)

p4

t5 τ

p6

p7

t9t8

p1

XOR-join

XOR-split

AND-join

AND-split

Flow

Place

Transition

Token

p2

p3

p5

p8

t3

t2

t6

Prepare Travel Form(Employee

ApproveTravel Form

(Admin)

RejectTravel Form

(Admin)

Submit Travel Form for Approval (Employee)

Request for change(Admin)

t7

p4

t5 τ

p6

p7

t9t8

τ

p1

Existing approaches

• Most approaches only consider the syntactical issues or simply create the configured model and analyze it (i.e., trail and error).

• Naïve approach: enumerate all possibilities and check, or trail-and-error at configuration time.

• Better approach based on partner synthesis: construct a configuration guideline at design time!

PAGE 47

W.M.P. van der Aalst, N. Lohmann, M. La Rosa, and J. Xu. Correctness Ensuring Process Configuration: An Approach Based on Partner Synthesis. In BPM 2010, volume 6336 of Lecture Notes in Computer Science, pages 95-111. Springer-Verlag, Berlin, 2010.

Partner synthesis

PAGE 48

t11 t12 xy

τ t14t13 z

p11

p12

p13

Ω ={[p13]}

p14 =desirable propertye.g. weakly terminatingpartner??

Given a service:• Is this a partner?• Is there a partner?• How to describe all partners?

Partner synthesis

• Existing approaches can do the following:− Check whether there exists a partner.− Construct partners having desirable properties.− Characterize all such partners.

cf. [Karsten Wolf. Does My Service Have Partners?. T. Petri Nets and Other Models of Concurrency 2: 152-171 (2009)]

• These are implemented in Wendy. cf. [Niels Lohmann, Daniela Weinberg. Wendy: A Tool to Synthesize Partners for Services. Petri Nets 2010: 297-307.]

• How can this be used for ensuring the correctness of a configurable model and its configurations?

PAGE 49

Configurable interface (allow by default)

PAGE 50

Configuration guideline (allow by default)

PAGE 51

CG

Configuration guideline (allow by default)

PAGE 52

CG

Configuration guideline (allow by default)

PAGE 53

CG

PAGE 54

Configuration guideline (allow by default)

CG

Configuration guideline (allow by default)

PAGE 55

CG

Tool support

• YAWL editor− Creating C-YAWL models− Configuring C-YAWL models− Verification and auto-completion (using SAT solver

and/or Wendy)• YAWL engine• YAWL services

PAGE 56

Mapping(.cmap)

C-EPC model(.epml)

ConfiguredC-EPC (.epml)

EPC(.epml)

Questionnaire Model (.qml)

ConfiguredC-YAWL (.yawl)

YAWL Engine spec. (.yawl)

C-YAWL model(.yawl)

Quaestio

Mapper ProcessConfigurator

ProcessIndividualizer

Configurationmodel (.cml)

QuestionnaireDesigner

C-YAWL Editor

C-EPC Designer Process Merger

cf. www.processconfiguration.com

p2

p3

p5

p8

t3

t1 t2

t4

t6

Prepare Travel Form(Secretary)

Prepare Travel Form(Employee)

ApproveTravel Form

(Admin)

RejectTravel Form

(Admin)

Submit Travel Form for Approval (Employee)

Request for change(Admin)

t7Check & Update Travel Form(Employee)

Arrange travel insurance (Employee)

p4

t5 τ

p6

p7

t9t8

p1

C-YAWL

PAGE 57

C-YAWL

PAGE 58

Conclusion

• BPM in the cloud triggers the need for configurable process models.

• Configuration is important, however, existing reference models are crap!

• Challenges:− Design of configurable models (language+approach)− Analysis of configurable models, e.g., ensuring

correctness− Discovering configurable models − Cross-organizational process mining

PAGE 59

More information

• www.processconfiguration.com(various references to configuration literature and a comprehensive toolset)

• www.win.tue.nl/coselog(webpage of the CoSeLoG project)

• www.processmining.org(webpage for process mining)

• www.yawlfoundation.org(C-YAWL)

• service-technology.org(analysis of services)

PAGE 60

References (Configurable Process Models)• M. Rosemann and W.M.P. van der Aalst. A Configurable Reference

Modelling Language. Information Systems, 32(1):1-23, 2007.• F. Gottschalk, W.M.P. van der Aalst, and M.H. Jansen-Vullers. SAP

WebFlow Made Configurable: Unifying Workflow Templates into a Configurable Model. In BPM 2007, volume 4714 of Lecture Notes in Computer Science, pages 262-270. Springer-Verlag, Berlin, 2007.

• F. Gottschalk, W.M.P. van der Aalst, M.H Jansen-Vullers, and M. La Rosa. Configurable Workflow Models. International Journal of Cooperative Information Systems, 17(2):177-221, 2008.

• M. La Rosa, W.M.P. van der Aalst, M. Dumas, and A.H.M. ter Hofstede. Questionnaire-based Variability Modeling for System Configuration. Software and Systems Modeling, 8(2):251-274, 2009.

• F. Gottschalk, W.M.P. van der Aalst, and H.M. Jansen-Vullers. Configurable Process Models: A Foundational Approach. In F. Lehner, H. Nosekabel, and P. Kleinschmidt, editors, Proceedings of the Multikonferenz Wirtschaftsinformatik 2006 (MKWI '06). GITO-Verlag, Berlin, 2006.

• …PAGE 61

References(Correctness Issues)• W.M.P. van der Aalst, M. Dumas, F. Gottschalk, A.H.M. ter Hofstede, M.

La Rosa, and J. Mendling. Correctness-Preserving Configuration of Business Process Models. In FASE 2008, volume 4961 of Lecture Notes in Computer Science, pages 46-61. Springer-Verlag, Berlin, 2008.

• W.M.P. van der Aalst, M. Dumas, F. Gottschalk, A.H.M. ter Hofstede, M. La Rosa, and J. Mendling. Preserving Correctness During Business Process Model Configuration. Formal Aspects of Computing, 22(3):459-482, 2010.

• W.M.P. van der Aalst, N. Lohmann, M. La Rosa, and J. Xu. Correctness Ensuring Process Configuration: An Approach Based on Partner Synthesis. In BPM 2010, volume 6336 of Lecture Notes in Computer Science, pages 95-111. Springer-Verlag, Berlin, 2010.

• M.T. Wynn, H.M.W. Verbeek, W.M.P. van der Aalst, A.H.M. ter Hofstede, and D. Edmond. Business Process Verification: Finally a Reality! Business Process Management Journal, 15(1):74-92, 2009.

• H.M.W. Verbeek, T. Basten, and W.M.P. van der Aalst. Diagnosing Workflow Processes using Woflan. The Computer Journal, 44(4):246-279, 2001.

PAGE 62

References(Process Mining)

• M.H. Jansen-Vullers, W.M.P. van der Aalst, and M. Rosemann. Mining Configurable Enterprise Information Systems. Data and Knowledge Engineering, 56(3):195-244, 2006.

• F. Gottschalk, T. Wagemakers, M.H. Jansen-Vullers, W.M.P. van der Aalst, and M. La Rosa. Configurable Process Models: Experiences From a Municipality Case Study. In CAiSE'09, volume 5565 of Lecture Notes in Computer Science, pages 486-500. Springer-Verlag, Berlin, 2009.

• W.M.P. van der Aalst, A.J.M.M. Weijters, and L. Maruster. Workflow Mining: Discovering Process Models from Event Logs. IEEE Transactions on Knowledge and Data Engineering, 16(9):1128-1142, 2004.

• W.M.P. van der Aalst, H.A. Reijers, A.J.M.M. Weijters, B.F. van Dongen, A.K. Alves de Medeiros, M. Song, and H.M.W. Verbeek. Business Process Mining: An Industrial Application. Information Systems, 32(5):713-732, 2007.

• …PAGE 63