1667 making z rules work session
TRANSCRIPT
© 2014 IBM Corporation
Making zRules Work with Healthcare at
Humana
Session # ASZ-1667
Helen Havener Rob Hilgert
Technical Architect Technical Architect
Humana Humana
Agenda
1
• Humana at a Glance
• Addressing Application Transformation
• Why Operational Decision Manager, why “z”
• Humana Operational Decision Manager (ODM) Background
• Rule Adoption within Humana
• Reflections/Summary
• Special Thanks
• Reference Material
Leading health care company that offers a wide range of insurance products
and health and wellness services; founded in 1961; headquartered in
Louisville, KY
2014 revenues of $48.5 billion
Total assets of approximately $23.5 billion
Over 30 years of experience in the Medicare program
One of the nation’s top providers of Medicare Advantage benefits with
approximately 2.9 million members
Approximately 13.8 million medical members nationwide
Approximately 7.7 million members in specialty products
Operates more than 500 medical centers and approximately 200 worksite
medical facilities
About Humana
2
Humana has two zEC12 CEC(s) supporting workloads
•ODM environmentWebSphere Application Server on z/OS (webservice calls)
Test, QA and Production (Decision Server)zRES – standalone execution server (COBOL and CICS calls)
Test, QAProduction - 3 LPARS
Decision Center – DB2 Repository on z/OSRule Execution Server – DB2 Repository on z/OSOne Administration for multiple LPARS
• CICS environment
Running CICS v5.1 (since March 2014)
168 CICS Regions
18,000 CICS COBOL programs
90+ Million transactions / day
200+ CICS Web Services
CICS related project growth = 5%
Reuse is low
• Initiatives
Transparency
Ease of Use
Agile
Performance
Modernization of Applications
Reduction of general CP MIPS
Increase Re-Use of Resources
Humana’s IT Environment
Agenda
4
• Humana at a Glance
• Addressing Application Transformation
• Why Operational Decision Manager, why “z”
• Humana Operational Decision Manager (ODM) Background
• Rule Adoption within Humana
• Reflections/Summary
• Special Thanks
• Reference Material
Transform Humana’s IT Assets to Preserve Investments
Legacy Applications – Monolithic, Siloed, Module-Based Approach that mixes program Code for Flow, Data Access, Integration, Presentation and Business Logic
VERSUS
Modernized Applications - Layered Service-Based Approach with Separation of Process Flow, Data Access, Integration, Presentation, Business Rules, Events, Monitoring and Business Services
Creating Business Components/Services from harvested/salvaged legacy modules…
Benefits of Application Renovation are:
– Simplify - Separate business functions from legacy applications and re-use of existing application services in new business processes
– Flexible - Process Assembly approach verses Complex Custom Code approach
– Respond - Increase IT responsiveness to business needs
– Align - IT Process Models with Business Process Models
– Agility - Increase revenue by enabling business to more quickly and dynamically adapt and capture new market opportunities
– Reuse – Extend the useful life of valuable, proven IT Assets
Monolithic Siloed
Applications Consist of
Intertwined:
• User Interface Logic
• Process or Flow Logic
• Security Logic
• Integration Logic
• Data Access Logic
• Error / Recovery Logic
• Plumbing Logic – threads, memory,
sockets, transactional semantics,
etc…
• Application Rule Logic and finally…..
• Business Logic
Modern Layered Applications
Consist of:
• UI Layer – Channels: Mobile, Web, Portal,
ESB, Client, Cloud API, etc…
• Process Layer – Process Engine
• Security Layer – Application / Security
Server and Directories
• Integration Layer – ESB
• Data Access Layer – Database Manager
and Related Capabilities
• Error / Recovery – Application Server
• Plumbing Logic – Application Server
• Business Rule Layer – Rules Engine
• Business Logic Layer
VS
Monolithic Versus Layered Approach
Agenda
7
• Humana at a Glance
• Addressing Application Transformation
• Why Operational Decision Manager, why “z”
• Humana Operational Decision Manager (ODM) Background
• Rule Adoption within Humana
• Reflections/Summary
• Special Thanks
• Reference Material
“ I look at technology as a way for us to engage with and help our customers. If we don’t do it, someone else is going to.” Humana CEO Bruce Broussard
• Aligning ODM with Humana’s Core Values• Pioneer Simplicity, one resource for Business Rules• Thrive Together, initiates collaboration in rule creation• Rethink Routine, Reuse of corporate assets across LOB’s
• Enforce Consistency to ensure Compliance
• Leverage Information to Manage Risk
• Cost Savings
• Reduce Time to Market
• Better Visibility and Control
Why ODM on z?
• Components of WebSphere Operational Decision Manager
• Rules and Events Designer
• Decision Center
– Decision Enterprise Center Console
• Decision Server
• Humana’s Deployment of Decision Center
• Rule Execution Server Deployment Options on z/OS
1. WebSphere Application Sever Rules Deployment Option
2. Standalone Rules Deployment Option
3. CICS Rules Deployment Option
4. Batch Integration
• Integration Patterns and Benefits of Rule Execution Server Deployment Options on z/OS
Rule and Event Designer
• Design
• Rules and events business objects
• Vocabularies
• Projects structure and organization
• Rule Templates
• Test
– Step by step debugging
– Value inspectors
– Test and simulation suites
– Completeness reports
• Configure
• Business environment (Decision Center)
• Deploy
• Rules and events projects to their respective execution environments
Comprehensive technical environment
Event DesignerRule Designer
Decision Center Enterprise Console
• Access rule artifacts concurrently without conflict or
delay
• Represent complex policies using rule overrides and
hierarchies
• Take control of very large rule bases
with Smart Views, easy search
and reporting
• Get automatic notification of potential
rule conflicts, redundancies
• See where rules are used across
projects using queries
• Hot-deploy rule and event changes in minutes
• Secure, integrated with enterprise security facility
including single sign-on
• Multiple release management supporting diff and merge
Web-based Event and Rule
Maintenance
Decision Center Console
Decision Center
Decision Server
Synchronize
Deploy
Deploy
Decision Center Enterprise Console
Decision Center Repository
Rules and Events Execution Server
Rules and Events Designer
Ruledocs
Rule Solutions for Office
RuntimeDesign
time
Decision Center Business Console
z/OS
Rule
Execution
Server
z/Linux
Decision
Center
DB2 DS Repository
Deploy Rules
Decision Center – Humana’s Deployment on z/Linux
• Advantages:
Takes advantage of IFL so no additional costs
“z” hardware qualities of service
Use of hipersockets to communicate with z/OS
Improved Security Configuration
The platform of choice for server consolidation including simplification of support model
DB2 DC Repository
z/OS
WebSphere Application Server
for z/OS*
Events ExecutionServer
Rule Execution Server
CICS TS v5.1
*Limited Use
zRule Execution Server
zRule Execution Server
• Deployment runtime options on z/OS:
Deployed on WebSphere Application Server for z/OS
Deployed standalone to z/OS for COBOL
Deployed in CICS TS 5.1 in JVMServer environment
ODM Deployment Options on z/OS
3
2
1
1
2
3
Deployment Option
Rule Execution Server for WAS
on z/OS
Web Services, Plain Old Java Object (POJO), Message Driven
Beans (MDB), Enterprise Java Beans (EJB).
Decisions can also be invoked from COBOL applications
using the WebSphere Optimized Local Adapter (WOLA).
WOLA is a functional component that provides an efficient
cross-memory mechanism for calls both inbound and
outbound calls to WAS z/OS
Options for Rules Invocation, Integration Patterns and Benefits
1
Integration Patterns
Unique Cross memory communication, z/OS WLM and a secure infrastructure.
Ease sharing of Rules: Decision services can be invoked from distributed platforms using the patterns shown above
Unique Automated Workload Balancing, High Availability and Scalability with WAS on z/OS topology
Rule execution is offloaded to zAAP reducing mainframe software costs and freeing up CPU cycles.
Rules shared across Logical Partitions in z/OS for unparallel high availability requirements (SYSPLEX)
Unique in industry Events Execution Server is integrated within the deployment option.
Ability to grow Rules and Events usage based on system capacity and application requirements versus a “per”
project basis.
Benefits:
Deployment Option
zRule Execution Server (Standalone) Simple COBOL API
Options for Rules Invocation, Integration Patterns and Benefits
2
Integration Pattern
COBOL applications use copybooks as the interface to invoke decisions making development simple.
z/OS permits the copying of message buffers from one virtual address space to another which makes ODM efficient,
fast and secure as it is within LPAR.
zRES can be added in an Automatic Restart Manager (ARM) sequence. This will allow zRES to be started
automatically in case of a failure.
Rule execution is offloaded to zAAP reducing mainframe software costs, freeing up CPU cycles.
Rules shared across LPARS: With DB2 data sharing, rules can be stored in DB2 and accessed from multiple LPARs
in a SYSPLEX environment
Batch/CICS and Rules Engine running on the same LPAR adds to better performance as there are no network hop.
Decision services are deployed and instantly available across z/OS without making any application changes
Ability to grow Rules and Events usage based on system capacity and application requirements versus a “per”
project basis.
Benefits:
Deployment Option
zRule Execution Server (CICS JVM) Simple COBOL API
Options for Rules Invocation, Integration Patterns and Benefits
3
Integration Pattern
Same benefits as option 2 plus additional
Running ODM inside CICS allows access to the CPSM* Qualities Of Service like workload management, reliability,
availability and security within a CICS infrastructure.
Uniquely Integrates with Business Event Emitting from CICS, which results in meeting high volume capacity
requirements.
Local execution within CICS region.
* CICSPlex System Manager(CPSM ) is a system management tool to manage multiple CICS Instances from a single
point.
Benefits:
Localize Execution of Decisions within a Batch Application
Address Space
JVM
Native
Control
Code
Rule
Server
User COBOL
Application
Decision
Service Stub Decision
Engine
• When calling zRES API call stays within same address space
• No cross address space communication or memory copies
• Reduced ‘native code’ path -> higher percentage zIIP eligible
• Configured and managed the same way as a standard zRES
CONN
RULE
DISC
Initialise
Invoke
Destroy
zRES API commands with Batch local deployments
• HBRCONN
• Starts the JVM locally and RES ready for rule execution
• Initially starts one processing thread
– Make threads configurable as part of this work
• Does not return successfully until the environment is started and waiting to execute rules
• HBRRULE
• Sends a rule request to the zRES code and points to the local data structures
• Waits for the local rule engine to return control
• HBRDISC
• Issues the call to the native code to terminate and clean up the JVM
• Frees all local storage and unloads all product code
What does ODM bring to z/OS?
Challenges forHumana
1. Consolidation, Isolation, Extension of COBOL application portfolio
2. Ability to react to increasing variety and volume of change requests regardless of platform
3. Sharing business rules across platforms & channels
4. Ensuring seamless business experience in migration/ application evolution
Benefits of the ODM Approach
Cost savings
– Shorter change cycle, without increased business risk
– Rule engine processing is zIIP eligible (approximately
90+% offload experienced)
Improved agility
– Improved Time to Market
– Manage business decisions in natural language
– Decouple development and business decision change
lifecycles
Single version of the Truth
– Consolidated and shared expression of business policy
– Maintainable with a Center of Competency model
Incremental Adoption
– Deploy decision methodology one decision at a time
– Focus on decisions that need to change often & quickly
– Expand adoption of “market validated” decisions
Agenda
21
• Humana at a Glance
• Addressing Application Transformation
• Why Operational Decision Manager, why “z”
• Humana Operational Decision Manager (ODM) Background
• Rule Adoption within Humana
• Reflections/Summary
• Special Thanks
• Reference Material
Humana had two initiatives with IBM Investigating ODM feasibility
• PBP (distributed based application) Humana CMS Software used to create Medicaid /Medicare Bids
Governing Body for Medicaid/Medicare Bids
CMS Publishes a questionnaire and we’re required to align our
700+ plans with those questions
.net Application
1500 rules ranging from simple to complex
Primarily executed January through July
Every plan, will execute business rules
SLA of 2 seconds
Critical Application within Humana
Planned Benefit Package (PBP)/Solar Application
• Solar (z/OS based application) Agent and Commission Application
COBOL Application
CICS/Batch
DB2
IDMS
WebSphere MQ Initiated (migrating to CICS Web Services)
Agenda
24
• Humana at a Glance
• Addressing Application Transformation
• Why Operational Decision Manager, why “z”
• Humana Operational Decision Manager (ODM) Background
• Rule Adoption within Humana
• Reflections/Summary
• Special Thanks
• Reference Material
•MPPR (Multiple Procedure Policy Reduction) Primarily used for determining reduction in payment of claims
COBOL CICS Application and scheduled batch daily processing
Frequent changes anticipated based upon government regulation with limited time to
implement
Approximately 422 Rules
Currently over 1 million ruleset executions daily
Sub-second response time requirement for CICS transactions
Majority of rules are not complex
Business vocabulary a plus for Line of Business
Ongoing ODM Adoption
• LL (Language Library) Enable Humana Plan Rules to be created based upon the language required.
Currently supporting English and Spanish
1500+ Rules
Supports the PBP application mentioned previously
Enabled the business users to build their own LL complex rules based on Humana
plans
.net application
• CCVT (CAS Constraint Verification Tool) Used to compare benefits and benefit types to ensure they meet requirements to pay
claims correctly.
Updated yearly to ensure logic is accurate
.net application
Agenda
26
• Humana at a Glance
• Addressing Application Transformation
• Why Operational Decision Manager, why “z”
• Humana Operational Decision Manager (ODM) Background
• Rule Adoption within Humana
• Reflections/Summary
• Special Thanks
• Reference Material
Understanding the extent and the visibility to which Operational Decision Manager, or a BRMS will impact the
Enterprise is Vital to SuccessCulture Impact
• Lines of Business are the initial point of development for business rule change• Timing for LOB changes are not aligned with application release schedules
• ODM promotes collaboration at all levels• Blurring of roles and responsibilities • New Roles will be required• Rapid technology adoption• Understanding Infrastructure Requirements to meet LOB Requirements
• Physical (sizing)• Licensing
Business Impact• LOB’s are more aware of BRMS, in theory, than generally the larger audience in IT• Business requirements for change are frequent resulting in concurrent change
activities outpacing typical application release change implementations• There is varying skill level within the business and different levels of engagement• Consulting engagements expedite adoption• Change activities involve integration and testing the backend
Reflections
Reflections Continued
28
Role Considerations (Source Redbook Governing Operational Decisions in a Enterprise
Scalable Way)
• Business Owner – Request business decision changes, provides business approval for release
• Change Control Board – Triages and selects change requests to be included in a specific release.
• Release Manager - Coordinates the contributors work toward a release, gathers appropriate approvals.
• Business Analyst – Provides business domain knowledge. Oversees change and validation activities. Is
accountable for the quality of decisions.
• Business Policy Analyst – Determines and manages the implementation of a business policy. Formalizes the
business terminology used ODM. Acts as a liaison between the business and IT at the time of the design
and implementation of decisions.
• Rule Author – Authors rules to support decisions. Manages and runs test cases to test decisions. Runs
simulations
• Operational Decision Manager (DC Admin) – Creates and configures Projects, manages permissions and
access rights on projects and artifacts
• Configuration Management - Performs configuration management tasks that are related to the release.
Acts as a liaison between business and IT at the time of deployment
• IT Integration Developer – Develops components around integration to other systems, web services, and
data services
• IT Quality Assurance (QA) – Performs integration testing
Security
• Decision Center (running on Linux on System z with DB2 repository on z/OS)
• Need to define security in relationship to responsibility and roles
• Requirement to isolate projects such that one LOB, won’t have access
to another LOB’s project
• Ability to share and enforce rule reuse when appropriate
• Decision Server (running on z/OS)
• LDAP aligning with ACF2 security
• Decision Server is used in Test for monitoring traces
• A number of rule developers aren’t familiar with z/OS
• Decision Warehouse resides on z/OS
• Marshaller XOM and Java XOM are deployed separately and both
are deployed directly on decision server
• Decision Server developers have RES monitor and RES deployer
access rights
• RES Admin is only granted to infrastructure
• Development access to JES, zRES and WebSphere Log for
problem determination
Reflections Continued
Summary
IBM’s Operational Decision Manager delivers on the promise to automate, govern and
improve operational decision making across business processes and applications for
improved business outcomes.
Communication and Collaboration across the Enterprise is key to success.
It enables the implementations of business changes in a easy, reliable and secure way to
meet market demands and policy requirements
Line of business can implement rule changes in a expedited manner reducing the time
implementing business decisions, regardless of application release cycles.
ODM on z/OS is effective as an enterprise solution for enabling a BRMS across multiple
platforms relying on the strengths of the enterprise server
• Reliability, stability, and availability
• Redundant processing architecture
• Advanced Technology
• Cost of ownership
Alignment with IBM Technical Resources is critical.
Agenda
• Humana at a Glance
• Addressing Application Transformation
• Why Operational Decision Manager, why “z”
• Humana Operational Decision Manager (ODM)
Background
• Rule Adoption within Humana
• Reflections/Summary
• Special Thanks/Reference Material
Special Thanks
• Craig Whitaker, Humana Technical Manager
• Craig Gardner, Humana Technical Architect
• Kripa Singh, Humana Technical Consultant
• John Cowel, IBM Certified IT Specialist, [email protected]
• Subhajit Maitra, zChampion Senior IT Specialist, [email protected]
• Dave Krouse, zClient Architect, [email protected]
• Bill Yager, Software Client Architect, [email protected]
• James Taylor, Master Inventor FVT Lead, Operational Decision Manager for
z/OS [email protected]
• Chris Backhouse, IBM Operational Decision Manager for z/OS
Reference Material
1. Flexible Decision Automation for Your zEnterprise
with Business Rules and Events, SG24-8014-01
Redbooks, published 7 November 2013, last updated 18
November 2013
2. Governing Operational Decisions in an Enterprise
Scalable Way, SG24-8127-00
Redbooks, published 26 April 2013, last updated 26 April
2013
3. WP102291 Why IBM Operational Decision Manager for
z/OS, last updated 05/29/2013
4. IBM ODM Information Center,
http://pic.dhe.ibm.com/infocenter/dmanager/v8r5/index.js
p?topic=/com.ibm.help.doc/zos_infocenter_homepage.ht
ml