search

 software engineering

0 comments

file time: 2008-02-16

filetype:pptx

Click Here To Download...

>  

Business Process Mashups??? Process Integration and Web 2.0 Growing Together 

Fabio Casati

University of Trento

Wetice 2007

A very short history of BPM00/font> Mostly, a history of failures but we00l see a few success stories as well Then, we will see BPM in a new, promising  context and study some more failures, so we do not get excited with the success stories or with the hype New trends, and the Web How Web 2.0 can influence and learn from BPM and integration technologies in general A two-way communication00with lots of opportunities for us, if we start doing useful research00/font>  
 
 
 

Over the next 50 minutes00/b>

 


Business processes are at the heart of what companies do Obvious goal: better/cheaper/faster processes How? Automation!! (1990s) Rise of workflow technology  
 
 

Business Process Automation via Workflow Technology 

DBMS applications 

sendmail application 

Online account access 

SAP accounting system 

XYZ 

Automate 00usiness logic00 information flow

 


Workflow paradigms aim at integrating coarse-grained components (applications) Programming in the large Have a single place where business logic is defined (?) Differences from conventional programming Native support for parallelism and asynch interactions Transactional and persistent (recoverable) Document and data management Organization model and resource assignment model Adaptation to invoke different kinds of components/apps  

Programming versus Megaprogramming

 


Many Benefits (in theory00/b>.) 

Reduced need for human labor No repeated data entry cheaper Faster processing Information flows automatically, 24x7x365 More reliable No case left behind Paperless and secure  
 


Workflow system architecture 

workflow engine 

development tools 

Workflow model repository 

execution logs 

analytics engine 

SAP adapter 

Workflow model, + possibly org model (or go to enterprise directories) 

SAP 

Email adapter 

Email app 

Custom adapter 

Account mgmt

 


Hundreds of WF systems developed in the 90s Results and adoption have been depressing Reasons for lack of success Complexity of WF systems License cost Heterogeneous components Lack of support for application development lifecycle Organizations not ready: You must have repeatable and understood processes, and be selective in what to automate  
 

A Success Story?

 


internal infrastructure 

supplier 

customer 

warehouse 

internal infrastructure 

internal infrastructure 

internal procurement

requests 

Web service 

Web service 

Web service 

Copyright Springer Verlag Berlin Heidelberg 2004  

Enter the (web) service 

services standards decentralized protocols 00nd even verticals  


SOA is catching on fast! 

The Global 2000 68% were engaged in SOA projects in 2005 98% will be engaged in SOA projects in 2007 Development costs will be reduced by 25%, or $53B over 5 years Aberdeen Group, 2006 
 
A survey of 306 US organizations found that 100% had started or planned to start SOA initiatives within 2 years Yankee Group, Nov 2005 
 
By 2008, 80% of enterprise development projects will be based on SOA Gartner, Mar 2005  
Once implemented, their SOAs will be used for: 44% Integrating applications internally 28% Providing services directly to customers or consumers 21% Connecting with external apps provided by partners 53% All of the above The top 5 factors influencing companies adopting an SOA: 1 Greater degree of business flexibility 2 Lower cost of integrating existing applications 3 Lower cost of developing new applications 4 Less time required to develop new applications 5 Less time required to maintain existing applications Source: Computerworld / CIO Survey, Feb 2006  

10 

Workflow mgmt failed to keep its promises But service composition is different (or, the business and IT environment is different) standardization (components and composers) maturity reduced costs (small layer on top of other middleware) mindset already there (even too much00 lots of products, open source efforts,00/font>  

Failures of the past. And of the present?

mmmmm, I don00 know about that Maybe we need to rethink our basic assumptions behind SOC Reuse is considered at the heart of SOC, and the basic assumption of much of the work done here Seems logical. Makes sense. But: has anybody seen it work? For real? I am talking about SOC, not SOAP00/font> CBSE, workflows, also made sense, but had more hype than success Design for change, same thing Is our research based on the right assumptions?  

Now: a success story? (again)

 

12 

Building a house

 

Looking for the right components Similar arguments for REST vs SOAP and WS-* But: we have to find the right level of separation between programming and megaprogramming the right features (KISS) the right components and the killer apps! (yes, beyond travel reservation)  

Not for everybody00

Everything here and now Complete visibility, no delays In fact, we are annoyed if we do not have a 00eal time monitoring00of the future But mom, email is for old people!! You see this in many aspects of IT, and of life Planning gives way to reaction-driven behavior: let00 go, and then we adjust Trip planning, autonomic computing, evolutionary approach to SW development00/font> Everything is in beta forever00 If not in beta, it00 old!00  

More expectations from the Web00

 

Our lesson?  

Visibility, manageability, adapt as you go Monitoring, testing, prototyping, analysis solutions become really important  
Again: in the right context!!!!!  

Gartner CIOs survey, 2006 

25 

Top 10 Business Priorities  Top 10 Technology Priorities  Business process improvement Business Intelligence applications Controlling enterprise operating costs Security technologies Attracting and growing customer relationships Mobile workforce enablement Improving competitive advantage Collaboration technologies Improving competitiveness Customer sales and service Using intelligence in products and services Service Oriented Architectures (SOA) Security breaches and disruptions Workflow management Revenue growth Networking, voice and data communications Faster innovation Virtualization Faster innovation and cycle times Legacy application modernization  

More 00ntegration00lessons from the web 

Managing heterogeneity, collaborative development, content integration, collaborative evaluation00/font> Wikipedia? Nice! Flickr? Cool!! Desktop search? Great!  
We did not get it, eh? We do not write collaborative, evolving textbooks Think about how we would do an online encyclopedia00 We write papers like centuries ago (waterfall model) We did not notice that there is the (social) web, we do not really collaborate We write about reuse, but our level of reuse is null  

Some thoughts are percolating.. 

Software engineering processes Process Management Also because of outsourcing and compliance DataSpaces ProcessSpaces  

28 

The Two Sides of Business Process Management 

Execution/Automation WF, megaprogramming Flexibility Management Visibility, Monitoring Analysis, Optimization Modeling, Re-engineering  
You can have one without the other (and in many cases, you should)  

0/0/00 

29 

Abstract Process Modeling

Understand your basic process behaviors Detect bottlenecks Real time visibility of process KPIs Understand business impact of IT failures All this without changing your process  

0/0/00 

30 

Process Insight: Results

 

0/0/00 

31 

Process Mining Process 

Preparation and Labeling 

Mining 

NumExec 


GetApproverDecision 

<=2 

>2 And <=6 

T         V 


11.9%  11.4% 


88.1%  88.6% 


2850    2030 

{Res1,..} 


Resn 

,...} 

>6 

Resource_Init_ 

GetApproverDecision 

T         V 


0%  

1.87% 


100%  98.13% 


0       

150 


9115     

6076 

T         V 


10.1%   8.9% 


89.9%  91.1% 

1  

1217      802 


10945    7233 

T         V 


60.6%    60% 


39.4%  

40% 


1633    1078 


1089   

718 

T         V 


4.11%  

7.6% 


95.89%  92.4% 


17      

18   


396    

220 

T         V 


70%   68.0% 


30%   32.0% 


1616      1060 


693       

498 


21150  15790 

{Sat,..Thu} 

{Friday} 

T         V 


5%       8% 


95%     92% 


553     

618 


10522    7111 

T         V 


61.1%  60.0% 


33.9%  40.0% 


652       184 


435       122 

StartDay 

NumExec 


GetApproverDecision 

<=2 

>2 And <=6 

T         V 


11.9%  11.4% 


88.1%  88.6% 


2850    2030 

{Res1,..} 


Resn 

,...} 

>6 

Resource_Init_ 

GetApproverDecision 

T         V 


0%  

1.87% 


100%  98.13% 


0       

150 


9115     

6076 

T         V 


10.1%   8.9% 


89.9%  91.1% 

1  

1217      802 


10945    7233 

T         V 


60.6%    60% 


39.4%  

40% 


1633    1078 


1089   

718 

T         V 


4.11%  

7.6% 


95.89%  92.4% 


17      

18   


396    

220 

T         V 


70%   68.0% 


30%   32.0% 


1616      1060 


693       

498 


21150  15790 

{Sat,..Thu} 

{Friday} 

T         V 


5%       8% 


95%     92% 


553     

618 


10522    7111 

T         V 


61.1%  60.0% 


33.9%  40.0% 


652       184 


435       122 

StartDay 

NumExec 


GetApproverDecision 

<=2 

>2 And <=6 

T         V 


11.9%  11.4% 


88.1%  88.6% 


2850    2030 

{Res1,..} 


Resn 

,...} 

>6 

Resource_Init_ 

GetApproverDecision 

T         V 


0%  

1.87% 


100%  98.13% 


0       

150 


9115     

6076 

T         V 


10.1%   8.9% 


89.9%  91.1% 

1  

1217      802 


10945    7233 

T         V 


60.6%    60% 


39.4%  

40% 


1633    1078 


1089   

718 

T         V 


4.11%  

7.6% 


95.89%  92.4% 


17      

18   


396    

220 

T         V 


70%   68.0% 


30%   32.0% 


1616      1060 


693       

498 


21150  15790 

{Sat,..Thu} 

{Friday} 

T         V 


5%       8% 


95%     92% 


553     

618 


10522    7111 

T         V 


61.1%  60.0% 


33.9%  40.0% 


652       184 


435       122 

StartDay 

NumExec 


GetApproverDecision 

<=2 

>2 And <=6 

T         V 


11.9%  11.4% 


88.1%  88.6% 


2850    2030 

{Res1,..} 


Resn 

,...} 

>6 

Resource_Init_ 

GetApproverDecision 

T         V 


0%  

1.87% 


100%  98.13% 


0       

150 


9115     

6076 

T         V 


10.1%   8.9% 


89.9%  91.1% 

1  

1217      802 


10945    7233 

T         V 


60.6%    60% 


39.4%  

40% 


1633    1078 


1089   

718 

T         V 


4.11%  

7.6% 


95.89%  92.4% 


17      

18   


396    

220 

T         V 


70%   68.0% 


30%   32.0% 


1616      1060 


693       

498 


21150  15790 

{Sat,..Thu} 

{Friday} 

T         V 


5%       8% 


95%     92% 


553     

618 


10522    7111 

T         V 


61.1%  60.0% 


33.9%  40.0% 


652       184 


435       122 

StartDay 

Interpretation 

Classification rules 

abstract process logs 

business data 

metric logs 

00/font> 

Training and Validation sets

 

32 

Factors affecting duration 

Factors affecting the outcomes 

Critical Factor Identification

 

ProcessSpaces: Following the breadcrumbs 

Web server 

App server 

Message broker 

Workflow system 

ERP system 

sharepoint

 

Process Space Management System 

Recognize that process executions are characterized by: Support from a variety of IT systems Semi-dynamic nature Heterogeneous information formats and repositories Provide a facility to query your enterprise based on processes, not row tuples and tables Abstractions are processes and instances and process data and00/font>  

Examples of questions to be answered 

Where is order #352? What happened to it? Is it blocked? Get me all the docs and info related to the processing of order 352 What is the typical path of orders. What is the process flow? Where is time spent in processing orders? Where do delays occur? At which stage do orders get rejected? How many, why? How often is it a business decision and how often is it a technical problem? And which problem? How much time and resources are spent in processing orders that are eventually rejected?   

Current solutions 

Mostly, do it yourself Main forms of tool support Workflow analysis Data warehousing  

Information space management system: functionality 

Explore: Identify and query relevant information sources Each with its own 00uery support00/font> Join (correlate): Query across sources Trace: follow the crumbs for each process Discover: extract the process flow (typical and exceptional), discover dependencies Analyze (OLAP, mine): derive process level statistics, bottleneck analysis, failure analysis, compliance analysis00/font>  

Ingredients of the solution 

Metadata on information sources 

Noise reduction 

Discovery and refinement 

Search and query 

Process analysis models 

Monitoring and Analysis 

Correlation

 

What we can teach to the social web00/b> 

UI integration Flows, esp data-related flows Discovery Tracing, confidence, security/privacy  

Convergence of approaches to  integration 

Data 

App logic 

Presentation 

Data 

App logic 

Presentation 

Data 

App logic 

Presentation

 

Data integration 

Data 

App logic 

Presentation 

Data 

Data 

Integration layer

 

Application integration 

Data 

App logic 

Presentation 

Data 

App logic 

Data 

App logic 

Integration layer

 

UI/presentation integration 

Data 

App logic 

Presentation 

Data 

App logic 

Presentation 

Data 

App logic 

Presentation 

Integration layer 

component

 

UI services00/b> 

Event driven Implicit notion of states Pub/sub integration more than flow based Layout aspects  

Convergence of UI, application, and information integration 

ESB to access data, DBMSs can invoke services EAI company selling as data integration, and vice versa Workflow vs pipes Integration of presentation content vs integration of UI Pure UI integration vs hybrid UI/app integration  

So what? 

Simplicity is the key Not completeness Separate simple from complex 20/80 always a good rule General/modular principles not always good In composition, need to find right component and right composition language 00ell established00 principles do not seem to work well Do now, think later seems what people like assess, advice, act Bring your lessons to the Web We BP and integration people should have a saying in mashups Avoid sudoku  

1

 


Presentation Title 

3

 


Presentation Title 

4

 


Presentation Title 

5

 

Presentation Title 

6

 


Presentation Title 

7

 


Presentation Title 

8

 


Presentation Title 

9

 

10 

Presentation Title 

10

 

11

 

Presentation Title 

12

 

14

 

15

 

Presentation Title 

20

 

Presentation Title 

21

 

28 

Presentation Title 

28

 

Presentation Title 

29

 

Presentation Title 

31

 

Presentation Title 

32

 

33 

What happens

What we want to find out

 

40

 

41

 

42

 

43

 

46

   download software engineering

Responses to software engineering

It's no comment...

 

Your Name:
Your Email:
Your Talk: