Data Encapsulating Services:A New Paradigm for

Developing Enterprise Systems

Jianwen SuUniversity of California, Santa Barbara

浙江大学 2

Hangzhou Housing Management Bureau

2012/11/16

500 workflow models300,000 cases/year

Obtaining a Permit (商品房预售申请 )

2012/11/16浙江大学 3

Permit for Selling an Unbuilt Appartment

Application申请 preliminaryreview初审

secondaryreview复审 Approval审批

lic. fee payment缴费Certificate制证Delivary递交

浙江大学 4

Ad hoc design, developed over time, patches, multiple technologies, … a typical legacy system

Problems:Embedded business logic, hard to learnhard to maintain, costly to add new

functionalityhard to change/evolve

A Typical Housing Management System

2012/11/16

浙江大学 5

OpenAPI approach towards Business Process as a Service (BPaaS) Enterprise may run virtual IT systems

Virtual Enterprise Systems

2012/11/16

Enterprise System

Tax Calculation

Reassessment Title ChangeInheritance

Determine tax base

Appraisal

TaxTitle

HR Accounting

Assessor

Good idea but needto manageenterprise data

浙江大学 6

Data and services are separately modeled, designed, managedAdds difficulties in design, execution,

maintenance, and making changes Furthermore, many issues can’t be addressed

Workflow transaction remains an artData consistency is a concern of DBMS even

though violations are caused by service execution

Business analytics is an after thoughtLong tail phenomenon is a “holy grail”Big data

Current Practice

2012/11/16

Mckinsey Global Institute, June 2011: Big data: The next frontier for innovation, competition, and productivity

Availability of “big data” brings opportunities for improving productivity

浙江大学 72012/11/16

Big Data—A Gowing Torrent

浙江大学 8

Big Data + Biz Processes Big Potential

2012/11/16

Source:MGI Analysis

Two observations A significant portion of big data

generated by biz processes Productivity growth only

obtainable via more efficient/effective biz processes

浙江大学 9

OpenAPI approach towards Business Process as a Service (BPaaS) Enterprise may run virtual IT systems

Virtual ES’s: Data Management Problem

2012/11/16

Enterprise System

Tax Calculation

ReassessmentTitle Change

Inheritance

Determine tax base

AppraisalPAL

TaxPALTitlePAL

HR AccountingPAL

AssessorPAL

?

浙江大学 10

No major players Enterprise makes its own data

management plan Bring Your Own Data (BYOD)

An Open Market Model

2012/11/16

Enterprise system

Data needed for services

+ services states

BPaaS withOpenAPI

浙江大学 11

Business Artifacts (业务流程工单 ) A business artifact is a key conceptual business

entity that is used in guiding the operation of the businessfedex package delivery, patient visit,

application form, insurance claim, order, financial deal, registration, …

both “information carrier” and “road-maps” Technically, it includes two parts:

Information model:data needed to move through workflow

Lifecycle:possible ways to evolve

Very natural to business managers and BP modelers

2012/11/16

浙江大学 12

DisagreedReceipts

PaidReceipts

PendingReceipts

ArchivedReceipts

ArchivedKOs

ReadyKOs

PendingKOs

Add Item

Prepare &Test Quality

Deliver

Payment

RecalculateReceipt

PrepareReceipt

CreateGuest Check

UpdateCash Balance

ArchivedGCs

ClosedGCs

OpenGCs

Guest Check

Artifacts

Kitchen Order

Receipt

Cash Balance

Example: Restaurant Processes

CashBalance

Activity

repository

2012/11/16

GC

KO

RC

CB

浙江大学 13

SeGA separates data from execution engine Serves as a mediator

SeGA: A Service Wrapper/Mediator

2012/11/16

[Sun-Xu-S.-Yang CoopIS 12]

SeGA ...

Dispatcher Repository

Event Queue

Barcelona

Engine 1

EZ-FlowEngine n

Barcelona

Engine n

Incomingevent

to SeGA

Fetch correlatedinstances ofthe event

Send the event,the process instances,& their schemasto the engine

Outgoingevent

Process the event,update the instances, & emit outgoing events

Dispatcher retrievesthe updated instances &stores into the repository

. . .

浙江大学 14

Data package between SeGA & services:Business data, enactment data, resource data,

correlation data Data encapsulating services: Stateful services

but the engine need not maintain state Independence of

data and service management

Data Encapsulating Services

2012/11/16

Enterprisesystem SeGA

浙江大学 15

Design: What are appropriate service designs? Choreography vs orchestration (Part II)? Design aid (analysis/model checking tool), interoperation

Runtime: Enforcement of process/data constraints, KPI/monitoring techniques, resource planning and management

Transactions: What is the notion of workflow transaction?

Change/evolution: Process vs instance changes, long lasting vs temporary, longtail

Big data: monitoring to analytics to change

Research Challenges

2012/11/16

浙江大学 16

Traditional and new service sectors and businessesHousing management (房管 )Social security (社保 )RetailAuditing (审计 )…

May spawn new types of service businesses such as cloud platforms for BPaaS providersconsulting businesses to analyze and improve

business processes (big data)

Potential Applications

2012/11/16

浙江大学 17

Inclusion of persistent data is critical to capture business logics into services

Data encapsulating services enable separation of data and service management and supportIndependence of data and service

management Many research challenges

Conclusions

2012/11/16