WARIA course Introduction - PowerPoint PPT Presentation

1 / 53
About This Presentation
Title:

WARIA course Introduction

Description:

Amit Sheth Keynote Address HEALTHCOM99 Sydney, AU Jan.11, 1999 Large Scale Distributed Information Systems Lab at University of Georgia http://lsdis.cs.uga.edu – PowerPoint PPT presentation

Number of Views:181
Avg rating:3.0/5.0
Slides: 54
Provided by: StefJo
Category:

less

Transcript and Presenter's Notes

Title: WARIA course Introduction


1
Workflow Process Management and
Enterprise Application Integration in Healthcare
Requirements, Applications, Technology and
Research
Amit Sheth Keynote Address HEALTHCOM99 Sydney,
AU Jan.11, 1999
Large Scale Distributed Information Systems
Lab at University of Georgia http//lsdis.cs.uga.e
du
Thanks gt Krys Kochut, John Miller and METEOR
Teamgt NIST-ATP, IONA Technologies, Informix, ...
2
Introduction
  • Comprehensive Healthcare IT solutions require
  • broad variety of process management
    applications, and
  • enterprise application integration
  • in a rapidly changing, distributed, and
    heterogeneous computing environments.

3
Reprsentative Healthcare Applicationsfor which
the METEOR EAppS has been used
  • Application
  • NeoNatal Clinical PathWays
  • Genome Sequencing
  • Eligibility-Referral
  • Immunization Tracking
  • Key Characteristics
  • Clinical, single site/server
  • Non-medical, multiple user organizations, two
    servers
  • Administrative, two participating organizations
  • Clinical and Administrative, multiple servers,
    multiple organizations

4
Neonatal Clinical Pathways
Nurse Coordinator
Neonatologist
Nurse
Server
5
Neonatal Clinical Pathways
  • Multiple inter-related clinical pathways or
    treatment plans for premature, underweight babies
    (head ultrasound, metabolic screen, nutrition,
    immunizations, etc.)
  • Alert on test to be ordered, enter tests and
    review results, update database, print reports
  • Users nurse coordinator, nurse, neonatologist
  • single server, Web-based clients

6
Head Ultrasound Workflow
7
(No Transcript)
8
Multi-institutional Genome Sequencing
Users in different labs worldwide
Georgia
Solaris Server
SGI Server
Legacy App
Legacy App
9
Data Analysis in a Multi-institution Genome
Sequencing Enterprise
Task 1
Task 2
Task i
Raw output from DNA sequencing machines from
multiple institutions
Fully sequenced and annotated genome
...
Automatic and human-computer data analysis tasks
running on heterogeneous servers
10
Characteristics of Data Analysis Tasks
  • Many existing legacy data analysis applications
    and new applications running
  • different servers (SGI and Solaris)
  • Each human-computer task performed by an expert
    who specializes in that task experts are located
    at different institutions throughout the world

11
Eligibility and Referral
12
Eligibility and Referral(i.e., patient dumping)
  • Sending Organization -- Receiving Organization --
    (optionally insurance company and ambulance
    services)
  • Two independent workflows with interoperability,
    optionally EDI with other organizations

13
Sending Organization
Receiving Organization
14
An Example of Mission-criticalEnterprise
Application Developed using METEOR EAppS
15
Schematic of Immunization Tracking
16
Implementation Testbed Immunization Tracking
Demoask us for a demo
Admit Clerk
Triage Nurse
Doctor/ NP
Maternity Ward
Administrator
Case Worker
(Solaris)
DBMS
(Windows NT)
CORBA
Hospital Info System
Web Server
MPI
MEI
Web Server
(Solaris)
Immunization Database
DBMS
(Solaris)
Encounter Database
Hospital
CHREF
EDI
Clinic
CHREF/SDOH
Office Practice Mgmt System
(Solaris)
DBMS
Admit Clerk
Triage Nurse
Doctor/ NP
Web Server
Encounter Files/Databases
Insurance Eligibility Database
17
Web-based Implementation
Triage Nurse Task
Admit Clerk Task
Update Personal Data
Update Local Database
Start
. . . .
Start
Stop
Collect Vitals
Generate Alerts
Enter Patient Info.
Check Eligibility
Workflow Design
Updated Results
Implementation
Generate Triage Nurses Page
Update Personal Data
N
Generate Alerts
CGI
CGI
Display Worklist
Enter Patient
Update Local Database
CGI
CGI
CGI
CGI
PULL
N
Add to Worklist
DBMS
HTTP
WORKLIST HANDLER
N
N
CGI
Machine Boundary
Alert Results
Eligibility Results
Check Eligibility
Worklist File
N
Web Page
Submit Button
18
Web-based Implementation
Triage Nurse Task
Admit Clerk Task
Update Personal Data
Update Local Database
Start
. . . .
Start
Stop
Collect Vitals
Generate Alerts
Enter Patient Info.
Check Eligibility
Workflow Design
Updated Results
Implementation
Generate Triage Nurses Page
Update Personal Data
N
Generate Alerts
CGI
CGI
Display Worklist
Enter Patient
Update Local Database
CGI
CGI
CGI
CGI
PULL
Add to Worklist
DBMS
HTTP
N
N
CGI
Machine Boundary
Alert Results
Eligibility Results
Check Eligibility
N
Web Page
Submit Button
19
Provider InterfaceImmunization Recos
List of overdue vaccinations
Link to contraindication info obtained from the
Internet
Clinical update to administer vaccination
20
CORBA-based Implementation
21
Requirements of mission-critical healthcare
applications that METEOR supports
  • Some of the important requirements for this
    application include
  • Support for a distributed architecture in a
    heterogeneous computing environment. At the level
    of any user of the system, this distribution
    should be transparent
  • Support for inter- and intra-enterprise wide
    coordination of tasks
  • Provision of a standard user-friendly interface
    to all users of the system
  • Support for a variety of tasks transactional and
    non-transactional, user and application

22
Requirements of mission-critical healthcare
applications that METEOR supports
  • Capability of using existing DBMS infrastructure
    and databases across organizations
  • Low cost of system for the providers and user
    organizations
  • Ease of modification (re-design), scalability,
    extensibility and fast design-to-implementation
  • Open architecture, use of standards (e.g., CORBA,
    CORBAmed, OMG WF, WfMC, SWAP) and middleware
    (JDBC)
  • Security authorization for users and secure
    communication (required as patient data is
    typically confidential)

23
Adaptive Workflow Example 1
  • If initial exam identifies needs for additional
    procedure in the blood test, the task/subprocess
    component is obtained from the repository and
    added to the subprocess

standard test
get blood
write report
This example is from Han et al.
24
Adaptive Workflow Example 2
  • Comorbidity new lab result identifies previously
    unknown illness or health condition that requires
    additional health management in addition to the
    current condition.
  • New related workflow (or subworkflow) is kicked
    off.

25
State of the art technology, products and research
26
Market Approaches to Workflow
Internet
ORB
e-mail
workgroup/ collaborative
document mgmt.
imaging
database
TP-monitoring
27
Key conclusion from DocuLabs workflow product
comparison study
  • Products that allow development of complex
    applications, provide flexibility and support
    integration are very hard to use, take too long
    and require experience programmers (and are still
    not quite comprehensive)
  • Easy to use products are too restrictive --and
    are useful for small or standard applications
    only
  • METEORs objective is to avoid significant
    compromise.

28
What is lacking from state of the art in
Workflow Technology
  • Primarily centralized, client/server
    architectures
  • Difficulty in integrating with existing
    applications and databases, not meant for
    heterogeneous, multi-server environments
  • Static workflows that can be fully defined before
    enactment starts
  • Focus on repetitive processes
  • Most products are, however, suitable for many
    office automation and other human oriented
    processes.

29
Research Disciplines Contributing to to Workflow
Management
Database Management/ Information Systems
Distributed Computing/Systems
Software Engineering, HCI, etc.
MIS/ Reengineering/ Methodology
Software Process
CSCW
Quantitative and Formal Methods/Modeling
Organizational Science
30
Enterprise Application Integration EAI
  • The average Fortune 2000 company relies on 49
    enterprise-level applications to run its business
    and spends 25 to 33 percent of its IT budget just
    to get tem to talk to each other Dan Sholler,
    Meta Group
  • Systems integration costs in many cases amount of
    40 percent of IT costs

31
Message/Data Tools/Architecture of current EAI
products/solutions
Business Object/Process Tool/Architecture of the
METEOR Enterprise Application Development
Integration
METEOR Component
METEOR Component
User Task
HOST 1
HOST 4
Adapter
Adapter
Broker
METEOR Component
METEOR Component
METEOR Component
Adapter
Adapter
Application
EDI
Database Transaction
HOST 2
HOST 3
Services and Business Objects
Publish/Subscribe event-based messaging
Distributed Object Management (ORB)
Enterprise Java Beans
DCOM
Messaging/PersistentQueue Middleware
32
METEOR EAppS Enterprise Application Suite of
Services and Tools
Enterprise Application Integration
CORBA
Adaptive and Scalable Workflow
Java
Internet/Web
E-Commerce
database middleware
33
METEOR Architecture
34
METEOR2 EAppS Enterprise Application Suite of
Tools and Services
End Users
35
METEOR Model
  • Integral support for human and application tasks
  • transactional and non-transactional tasks
  • human tasks forms, collaborative tasks
  • automated tasks database access, legacy
    applications
  • Complex control and data dependencies
  • much more

36
(No Transcript)
37
METEOR Workflow Builder Service
  • Comprehensive set of graphical tools for building
    workflow process applications
  • Map Designer
  • Data Designer
  • Task Designer
  • Web Form Designer
  • Exception Designer

38
(No Transcript)
39
(No Transcript)
40
WebWork Application Development
.wil
.spec
Workflow Designer
Workflow Generator
Application Prototype
source
Task Customizer
Page Customizer
Specification Customizer
Workflow Application
.spec
.html
.cc
41
(No Transcript)
42
Web Enabled and Web Based Engines
Web Based
Web Enabled
Centralized
Distributed
Front End
Enactment, monitoring, administration
Vendor1
Vendor2
Vendor3
OzWeb
Endeavors
WebWork
Error handling Recovery (Transactional Support)
Distributed Scheduling
EJB
CGI
Tcl, Java, XML
43
(No Transcript)
44
(No Transcript)
45
OVERVIEW
Host A
Host B
Workflow Monitor
GRM
Cleanup Task
Workflow Administrator
GPS
Host C
Object Request Broker (CORBA)
Host E
LRM
LRM
Web
TTM
Database Task
LPS
DBMS
LRM
LPS
Web Server
Activate Message
Data Object
GPS Global Persistence Store GRM Global
Recovery Manager LPS Local Persistence
Store LRM Local Recovery Manager NtTM
Non-transactional Task Manager TM Task
Manager TTM Transactional Task Manger UTM
User Task Manager
UTM
NtTM
Application Task
User Task
Watchdog
Host D
46
ORBWork Enactment Service
  • Designed to address a variety of shortcomings
    found in todays component middleware, enterprise
    integration, and workflow process management
    systems, it supports the following capabilities
  • Integration of application components and
    databases on disparate distributed and
    heterogeneous computing environments across
    enterprise intranet and extranet boundaries
  • A fully distributed and highly scalable
    enterprise application process enactment system
    capable of supporting dynamic business processes

47
ORBWork (continued)
  • Utilization of open and industry standards such
    as CORBA/IIOP, Java, JDBC, HTTP, and OMG Workflow
    Facility for developing open, distributed,
    object-based, interoperable enterprise
    applications
  • Support for cross-platform portability of tasks
    in business processes (by exclusively developing
    them in Java)
  • Provision for a browser-based Web user interface
    for the end-users and the administrators.

48
Schedulers and Task Managers
  • Activation (scheduling) start, termination
  • Error Handling
  • Recovery
  • Task (Application) Invocation
  • Note that automatic code generation means all
    task managers are automatically generated! At
    run-time, they are started when needed.

49
Task Schedulers
  • A Task Scheduler performs the scheduling of all
    workflow instances flowing through it.
  • Task schedulers differ, depending on the type of
    task they manage.
  • Task schedulers form a hierarchy.

50
Task Schedulers(continued)
  • Upon detecting an enabled task
  • an AutoTaskScheduler activates an associated Task
    Manager to oversee the execution of the task
  • a UserTaskManager prepares a form with the
    instance-specific data and places the task on the
    worklist.
  • The worklist is managed by the scheduler.
  • The next task (instance) is then picked up by the
    end-user from a Web browser.

51
Task Scheduler Details
Monitor
Recovery System
Transition1
Input Dependency
Output Dependency
Data Objects
AND
Task
Transition2
Scheduler
OR
Transition3
Task invocation
Task Manager
Task
Data Objects
52
Task Scheduler Managing Multiple Workflow
Instances
Monitor
Recovery System
Transition1
Data Objects
Task
AND
Transition2
Scheduler
Task Manager
Task Manager
Task Manager
Task Manager
...
Task Manager
...
Task Manager
Pending
Done
Task invocation
Task
Task
Running
53
Task Managers
  • A Task Manager oversees the execution of an
    automatic task in the workflow.
  • Task Managers differ, depending on the type of
    task they manage.
  • Task Managers form a hierarchy.

54
Handling of Automated Tasks
  • Newly developed tasks
  • typically implemented as CORBA objects
  • Existing (legacy) tasks
  • wrapped as CORBA objects
  • invoked directly by the task manager which
    prepares the input data and handles the results

55
Handling of User Tasks
  • User tasks use a Web browser and an HTML form.
  • Communication between a CORBA task scheduler and
    the browser
  • access data objects to retrieve necessary
    attribute values
  • generate a form instance (for the current
    workflow instance)
  • place the form on the worklist.

56
Handling of User Tasks
  • The scheduler provides a rudimentary
    implementation of the HTTP protocol.
  • The end-users browser connects directly to the
    task scheduler.
  • The scheduler receives the GET method from the
    browser.
  • The task scheduler sends to the browser the
    worklist with the enabled task instances and/or
    the individual task forms.

57
ORBWorks fully distributed scheduling and
flexible placement
58
Achieving Scalability
  • Use of a fully distributed scheduler.
  • Component task schedulers mirror the workflow
    map design.
  • Natural distribution of scheduling and task
    activation functions among participating hosts.
  • Task schedulers can be replicated to increase the
    load handling potential.
  • Task managers can be distributed among many hosts
    to balance the load.

59
Dynamic Changes
  • Dynamic changes both to the enterprise
    application or workflow process schema and to the
    individual workflow process instances.
  • Administrator can easily modify the schema at
    runtime by acquiring the new information from the
    repository, or even by modifying the
    specification by HTTP-based direct interaction
    with the scheduler.

60
ORBWork implementationUtilizing ORBIXs
features
61
Why METEOR?
  • Generation shift (from first generation to
    second) and high-end market needs
  • client-server to distributed and network-centric
    (trend, scalability, flexibility, robustness
    requirement)
  • new generation of distributed computing
    infrastructure (WWW, CORBA and Java) and new
    standards some companies are going for CORBA
    standard (E.g. various CORBAmed and defense
    companies)
  • new standards OMG Workflow Facility is being
    proposed with METEOR as the prime example

62
Why METEOR?
  • Workflow management system can become an
    enterprise-wide and inter-enterprise application
    integration platform if it is open and
    standards-based most products do nothing to deal
    with heterogeneity commonly found in most
    enterprises
  • METEOR is the first and currently only
    implementation using distributed Web and CORBA
    infrastructures, and Java

63
Why METEOR?
  • Many products are seen as consultant-ware with
    complex toolkits
  • METEOR offers advanced code generation capability
    which can provide 60 to 80 saving in generating
    distributed workflow application in heterogeneous
    environments
  • METEOR has a significant edge in ease of use,
    deployment and management

64
Features, capabilities and benefits
65
Emergent New Market for Enterprise Application
Development
Workflow
EAI
E-Commerce
66
Summary
  • Enterprise Healthcare applications need more than
    what most workflow products and consultantware
    systems have to offer.
  • METEOR EAppS uses CORBA, Web, and Java offers
    unique solutions to realize the promise of recent
    advances in distributed computing infrastructure,
    middleware and Web, by quickly integrating
    applications and information systems, to support
    complex and dynamic business process management.

67
Additional information
METEOR EAppS (Technology/ Product Licensing)
  • http//lsdis.cs.uga.edu
  • (METEOR Research)

Infocosm, Inc. http//infocosm.com
Write a Comment
User Comments (0)
About PowerShow.com