Discussion Document (Walter Harris Briefing

1
CTMS Work Space Whats Coming Down the
Road?Moving from Isolation through Integration
to Agility
2
Overview

• Definitions and Examples
• The Vision of caBIG
• Scoped to CTMS WS
• The CTMS WS today focus on static data
  (query/response)
• The Grid v1.0
• The CTMS WS day-after-tomorrow focus on dynamic
  behavior
• Coordination/Cooperation in a complex
  organization
• The Steps Ahead (tomorrow)
• Introduction to Process Integration
  Integration vs Agility
• Multi-layered process and project management
• The role of caBIG/CTMS Governance
• The role of Domain Analysis models for
  representing static and dynamic data
• (next session)
• The CTMS Integration Project (v0.1 vs v1.0)
• Other Integration projects (intra-CTMS vs
  inter-WS)
• caTRIP, other TBPT projects, etc.
• Q A

3
Definitions and Examples

• Isolation software built without knowledge or
  awareness of other application development
  efforts. Absence of standards in data,
  functionality, or SEP
• Pre-caBIG
• Interaction ability of applications to perform
  specific, pre-defined exchanges of data or
  coordination of functionality
• Static
• normalization of data element/value set semantics
• Application of standards
• Silver-level caBIG compatibility
• Dynamic
• Scenario-specific (i.e. fine-grained/context-speci
  fic) coupling of application behavior
  (Integration layer)
• Non-scalable, brittle/resistant to change
• Agility ability to adopt to change or
  unexpected usage requirements
• Static similar to interaction (i.e. semantic
  normalization, application of standards)
• Dynamic process integration
• Course-grained business-level services
• Exchange of context as well as data enables
  run-time flexibility
• Tools for ongoing creation of scenarios
  (workflow processing)

4
The (CSI) Problem for Clinical ResearchThe
Biomedical Information Tsunami

• An overwhelming volume of data from
• A multitude of sources with
• A plethora of perspectives and (often
  conflicting) agendas

Ken Buetow, NCI
5
The (CSI) Problem for Clinical TrialsThe
Information Tower of Babel

• Each cancer research community speaks its own
  scientific dialect
• Different dialects exist between clinical and
  research areas of interest
• Information integration is critical to achieve
  the promise of molecular medicine ?
  individualized patient care

Ken Buetow, NCI
6
Protocol a common term
Symbol Protocol
Source John Speakman
7
Terminology ServicesIntra-terminology Version
Management
Terminology A (e.g. ICD-9)
Version 1.0 (06.04) Term 1 Code abc Term 2 Code
def Term 3 Code ghi
Version 1.1 (11.05) Term 1 Code abc Term 2 Code
ghj Term 3 Code ghk
8
Terminology ServicesInter-terminology Mapping
Management
Terminology A (e.g. ICD-9)
Version 1.0 (06.04) Term 1 Code abc Term 2 Code
def Term 3 Code ghi
HL7 has developed a Common Terminology Services
specification (i.e. ODBC for terminology servers)
9
caBIG CTMS Critical Path - 2006
Isolation vs Integration vs Agility?
10
What is the Grid (source Grid Update Aug,
2006)?

• A lot of different things to a lot of different
  people
• Evolution of distributed computing to support
  sciences and engineering
• Some common themes prevail
• Sharing of resources (computational, storage,
  data, etc)
• Secure Access (global authentication, local
  authorization, policies, trust, etc)
• Open Standards
• Virtualization
• The real and specific problem that underlies the
  Grid concept is coordinated resource sharing and
  problem solving in dynamic, multi-institutional
  virtual organizations.
• I. Foster, C. Kesselman, S. Tuecke. International
  J. Supercomputer Applications, 15(3), 2001
• A good general overview can be found here
  http//gridcafe.web.cern.ch/gridcafe/

11
caGrid Trajectory
11
/
15
/
05
1
/
25
/
06
5
/
12
/
06
caGrid
0
.
5
.
2
caGrid
0
.
5
.
3
caGrid
0
.
5
.
4
Release
Release
Release
Release
12
/
1
/
03
7
/
1
/
04
caGrid Concept Origin
Initial caGrid Prototype
Isolation no Integration yes (at the
query/response level) Agility Grid 1.0 does
contain standards-based workflow capabilities.
However, CTMS/caBIG does not have the same level
of process governance that it has in place for
data governance (e.e. Silver Compatibility) ?
providing Agility is the next big step and CTMS
is the lab
12
Why Agility?The Anatomy of Complexity

• Multiple vertical levels of organization combined
  with horizontal cross-organization processes to
  produce Products of Value
• Problems occur at the crossing points
• Understanding syntax and semantics
• Coordination dynamic behavior and context
  management
• Politics responsibility and accountability
• Cumulative experience in industry, art, and
  (cognitive) science has repeatedly shown that the
  best way to deal with complexity is through
  abstraction, layering, and the use of standards.
• Disparate/unstructured approaches to lead to
  unorganized/inefficient outcomes
• The world of CTMS/caBIG (we do discovery) is a
  complex system

Cross-organization Processes
13
CTMS/caBIG We do discovery

14
Other Factors Contributing to Complexity

• Number of involved parties
• Diversity of
• Roles
• Skills
• Perspectives
• Responsibilities
• Expectations
• Interactions
• Persons with People
• People with Machines/Systems
• Machines/Systems with Machines/Systems
• Complexity of processes to be automated
• Static components
• Dynamic behavior
• Agility of automation (contextual
  extensibility)
• Breadth and depth of data/data types involved
• Physical and temporal distribution of parties,
  data, and process start/end points

15
Problem Solving in a Complex EnvironmentAbstract
ion and Layering (chunking)

• Experience in a number of contexts science,
  architecture, industry, (even) art has shown
  that a universal approach to managing complexity
  is separation of static and dynamic aspects of a
  system in combination with layering of multiple
  levels of abstraction
• Cyclical (iterative/incremental) application of
  core process of definition, discovery,
  intervention, (re)evaluation (re-definition)

• The Nursing Process as a model of complex problem
  solving

16
The Unified Process FrameworkMulti-Disciplinary,
Iterative-Incremental, Risk-Focused,
Architecture-Centric
Phases
Governance Project Management Stakeholder
Management Business Requirements Requirements
Analysis Analysis Design Implementation Business
Testing Performance Testing Technical
Architecture Data Conversion Transition Operations
and Support
Iterations
Milestones
17
Applying the Unified Process Framework at
multiple levels/layers yields maximum impact
18
The Importance of Understanding the Business
Architecture

• The union of structure and function
• Both Static and Dynamic descriptions are
  essential
• Multiple nearly independent views of a single
  model at varying levels of abstraction
• the wiring diagram of the bathroom on the 15th
  floor
• the superstructure of the roof
• Recognition of the process and change
  implications of
• Structural walls
• Partition walls
• Furniture
• Degrees of Complexity Users and Uses
• Dog house
• Shack
• Single-family home
• Skyscraper

The application of Architecture is a discipline
for solving complex problems. It is not limited
to software or building. Different problems
require different architectures. However, all
architectures having the characteristics of
abstraction, layering, and application of
standards (often called patterns).
19
Architecture

• The most important characteristics of
  architectural complexity are
• Number and Type (perspective) of Users
  (stakeholders)
• Number and Type of Uses
• Conceptually, architecture can be thought of in
  terms of
• Structural Walls
• Partition Walls
• Furniture
• Expectation setting involves coming to agreement
  on (minimally) the structural walls
• Agility follows
• The degree of architectural complexity determines
  tooling and process requirements (which are, in
  turn, dependent on organizational maturity)
• The practice of traditional architecture has
  led to a number of Best Practices and
  Architectural Patterns
• The same is now true for Software Engineering and
  Business Re-Engineering

20
Achieving CTMS/caBIG Agility Management
Architecture

• Management
• Compliance with caBIG compatibility standards
• Application of UPF-compliant SEP for all projects
• Multi-layered project management of integration
  projects (short-term)
• Architecture
• Development and Implementation of a multi-layered
  governance framework that enables process
  integration in addition to data integration ?
  computable semantic interoperability at both a
  static and dynamic level
• Service Oriented Architecture
• Both of these big topics are defined and
  themselves managed within a mulit-layered
  Governance Framework

21
Governance A Multi-Factorial/Multi-layered
Framework
Establish chains of responsibility authority and
communication to empower people (decision
rights) And establish measurement and policy
control mechanisms to enable people to carry out
their roles and responsibilities
Governance is about getting people to do the
right thing at the right time in the right way
22
Marks of Good IT Governance With SOA

• Differentiated Business (Discovery) Strategies
  Enabled by SOA
• Business (Clinical Trial) Objectives for
  Evaluating SOA Investments
• Executives (Clinical Trial Stakeholders)
  Engaged (not just technologists)
• Stable with Few Changes Year to Year
• Well-Defined Formal IT Requirements and Exception
  Processes
• Multiple Formal Communications Methods to Engage
  Business Leaders (Clinical Trial Stakeholders)

23
Where Does CTMS/caBIG Stand?
Are we where we want/need to be?
24
Governance is Key to Delivering on SOA by Design
25
Making SOA by Design Happen
Policies (What)
Decisions (Who)
Processes (How)

• GOVERNANCE with SOA ADDRESSES
• What decisions must be made for effective
  management
• Who should make those decisions and who has input
  rights?
• How will the decisions be formed and enacted

26
6 Steps to Successful SOA Governance
6
1. Define Goals and Strategies
2. Define Standards, Policies, Procedures
Around Financial, Portfolio, Project, Service, etc
6. Refine and Go to the Next Level of SOA
Maturity
These 6 steps allow an organization to
incrementally develop and mature their overall
SOA and achieve the vision of operational agility
3. Define Metrics
5. Analyze and Improve Existing Processes
4. Put Governance Mechanisms in Place
27
Goals Strategies

• Align Organization business goals with IT
  Goals
• Define an SOA Strategy
• Assess Existing Capabilities
• Prepare an SOA Roadmap
• Recognize the importance of Journey Management
  (aka change management)

1
28
Create Standards, Policies Processes

• Communicate

2
29
Define Metrics for Success

• Why Measure ?
• Ensure Business Goals
• Deliver SOA Strategy
• What to Measure ?
• standards, compliance, of projects adhering to
  processes, of reference architectures,
  usability of reference architectures, of
  exceptions, of services created, of reusable
  services, service reuse metrics, etc.
• How to Measure ?
• What can be automated?
• What can be easily captured?

3
30
Put Governance Mechanisms in Place1. Decision,
Policies, Processes
Financial
Portfolio
People
Service Funding Model
Service Usage Fees
Projects
Roles Responsibilities
Platform Funding
Business Services
EA Group
4
Applications
Service Process Owners
Service Ownership
Capacity Planning
DRIVEN BY EXECUTIVES
Operations
Projects
Service Lifecycle Gov
Enforce Service Levels
Shared Artifacts
Enforce Policies
Strategic SOA Platform
Reference Architectures
Enforce Platform Decisions
Architectural Standards
Shared Foundation Srvcs
Blueprints Patterns
Technology
Architecture
31
Put Governance Mechanisms in Place2. Mechanisms

• Vision for Governance Endorsed by Executives
• Force Behavioral Change
• Ensure Participation of Appropriate People
• Awareness
• Communication Collaboration
• Center of Excellence
• Roles and Responsibilities
• Financial, Portfolio, People, Architecture,
  Projects, Technology, etc
• Education Strategies
• Processes (Automated)
• Capture and Report on Metrics
• Administration, Monitoring and Enforcement
• Requirements Management (CM/CM)
• Exceptions Handling Mechanisms

4
32
Analyze and Improve Evolution of Organizational
Maturity

• Metrics on Governance Process Itself
• Metrics on Progress of Goals and Roadmap
• Relax Restrictive Policies
• Take Corrective Action
• What do you do with the Information?
• Make Decisions
• Create Feedback Loop

5
33
Refine and Go to the Next Maturity Level

• SOA Strategies, Goals, Objectives Met for this
  SOA Maturity Level level
• Refine SOA Strategies, Goals, Objectives for
  Current Maturity Level
• Create New SOA Strategies, Goals, Objectives for
  next SOA Maturity Level

6
34
Designing a SOA based Composite Application
8. Performance, Reliability, Scalability
Reusability
7. Security Policy Management
6. Business Activity Monitoring
5. User Interface Collaborative Portal,
integrated content services
4. Incorporate Human interactions Business Rules
3. Process Assembly Service Orchestration
2. Enterprise Service Bus wiring services
1. Portfolio of Services publish organize
0. Business requirements Conceptual Design
35
The (first) CTMS Integration Project

• GOAL Produce a demonstration of intra-CTMS
  interoperability
• REQUIREMENTS
• Not just another demo
• First and foremost, scenarios driven by domain
  experts/actual users
• Production-quality code from GA application
• Realization of scenarios that include both the
  exchange of static data (query/response) and the
  coordination of dynamic functionality
• Static exchange has historically been the focus
  of caBIG
• Increasing focus on dynamic behavior will be a
  focus of the Grid going forward
• Implementation on scalable architecture/infrastruc
  ture (as much as possible)
• ? Deployable on the Grid ?
• ? Utilize CTMS lingua franca (HL7 V3
  specifications) ?

36
The (first) CTMS Integration Project v0.1

• CONSTRAINTS
• Produce first version by Annual Meeting (Feb 1,
  2007)
• Limited ability to change/enhance existing
  application functionality/data
• ORGANIZATION
• Running as a UPF-compliant project
• Initially two one-month iterations preceded by a
  3-week start-up iteration
• Post Annual Meeting ???
• Multi-disciplinary project team
• Project Manager with experience in Agile
  Development Process
• 3 SMEs with feet on the street experience
• Relevance Monitors
• Scenario authoring
• Acceptance testing
• Analyst (requirements ? analysis models)
• Design/Implementation teams
• C3D (NCI)
• C3PR (Duke)
• caAERS (Semantic Bits)
• caXchange (ScenPro)

37
The (first) CTMS Integration Project v1.0

• Enhanced functionality/data expressiveness from
  component applications
• caAERS and C3PR currently in very early
  development cycles
• Additional applications will become available
• Protocol Lifecycle Tracking
• More robust infrastructure for managing
  integration projects
• Better coordination/integration with BRIDG, VCDE
  efforts
• Lessons learned

38
caBIG CTMS Critical Path - 2006
Other
Input
Patient
Registry
(
C
3
PR
)
Lifecycle
Adverse
Management
Events
(
caAERS
)
Patient
Study
Calendar
NCI
Repository
Lab
System
Lab
Lab
Data
System
Hub
Agility stringing the same beads in an infinite
number of ways We do discovery
C
3
D
Lab
System
Other
CTMS
Systems
Lab
Repository
39
Additional CTMS Integration Projects

• Cross-domain projects
• caTRIP
• Other TBPT applications/data
• Other applications in other work spaces

40
QUESTIONS ANSWERS