Best practices in IBM Operational Decision Manager Standard 8.7.0 topologies

© 2015 IBM Corporation

2475 - Best practices in IBM Operational Decision Manager TopologiesPierre FEILLET – ODM Product Architect [email protected]

Agenda

• What’s new in ODM 8.7.0• ODM Standard Architecture & runtime elements• Recommended topologies• Decision Service Development Lifecycle• Business Rules everywhere• Best practices

2

What’s new in ODM 8.7.0

ODM 8.7.0 Editions• Standard

• Decision Center• Decision Server Rules

– Coming from ILOG JRules– Mostly stateless decision– Decision Runner– Decision Service projects– HTDS extended to RES/JSON exec

• Advanced = Standard +• Decision Server Event

– Continuation of WebSphere Business Event– Stateful execution – Event driven– WAS ND cluster with XS

• Decision Server Insights– Dynamic modeling to define Events, Entities & detection rules to search

for situations– Server Grid with events & data in memory– WAS Liberty + WXS

4

ODM Standard Architecture

ODM Standard - Concept of Operations

6

Decision Server

Development

Operation

Continuous Improvement

Rule ExecutionServer

JEE/J2SE

Rule Designer(Eclipse)

Decision Center -Decision Validation Services

z Rule ExecutionServer

COBOL

Decision Center Repository

IT ArchitectIT Developer

Business AnalystProcess Owner

Business UserBusiness Leader

Code Generation

Deploy

Synchronize

ODM Standard Component Organization

7

•Designer• Decision Center

• Business Console• Enterprise Console

•Decision Server Rules• Rule Execution Server

• Rule Execution Server web console

• Hosted Transparent Decision Service

• Scenario Service Provider• Decision Warehouse• Rule Sessions• eXecution Unit

ODM Standard Component List

8

Component Acronym Description DB tables SIBus

Decision Center DC Web consoles + the DM authoring repository to store rule and event projects

Decision Center tables None

Decision Server / Rule Server Console

RES Console Executable Decision Service mgt through RuleApp/ruleset assets

Rule Execution Server tables

None

Decision Server / Rule eXecution Unit

RES XU Rule server core runtime deployed as JCA or Pojo Rule Execution Server tables

None

Decision Server / Rule Sessions

Rule Sessions

Rule runtime components available as EJBs or Pojo None ODM Bus

Decision Server / Decision Services

Custom or OOTB JEE artifacts that encompass the invocation and the decision logic (RuleApp/Ruleset). Are based on Rule Sessions.

None None expected unless MDB used

Decision Server / Scenario Service Provider

SSP Test and simulation rule runtimeDecisionRunner war added in 8.7.0 for new Simulation

None None

Decision Server / HTDS HTDS Dynamic Web Services None None

Decision Server / Decision Warehouse

DW Runtime that listens and store decisions Decision Warehouse tables

None

A full fledge Enterprise Proposal

9

System z Enterprise 196 (or 114) +

LPAR, Sysplex

100% Java product on distributed plus a native secret sauce in zRES

OS Distributed including

z/Linux z/OS

Power or Intel bladesz/VM z/OS

z/Linux

WAS ND

& LWAS ND

CICS

ODM ODM

distributed

WAS ND Other AS

Works with a variety of stacks WAS ND, WAS Liberty profile (8.5.5.3) on distributed & z/OS zRES Lightweight server for z/OS, run standalone and in

CICS JSE Rule Execution Server & rule engine

ODM

ODM Technical Detailed System Requirements

Distributed

WAS

Liberty

WAS

Liberty

https://www-304.ibm.com/support/docview.wss?uid=swg27023067

https://www-304.ibm.com/support/docview.wss?uid=swg27023067

ODM Standard Consoles & Touch Points

10

Decision Center

Decision Server

ODM cell

Rule Designer

Decision Center Consoles

Desktop

WAS Administration Console

Rule Execution ServerConsole

DC Ant Tasks

DS Ant Tasks

Synchro over HTTP(S)

Executable artifact deploymentover HTTP(S)

All Consoles, synchronization, deployment going through HTTP(S)

HTTP(S)

HTTP(S)

HTTP(S)

HTDS WS endpointsHTTP(S)

Zoom on Decision Server / Rule Architecture

11

RuleApp & Ruleset archive deployment

Rule Sessions EJBPOJO

Engine JCA

CustomDecision Services

eXecution Unit

Management Console

RuleApp/Ruleset Repository

EngineEngine

Mgt Notification

JMX/IP

Management Model

Decision Warehouse DB Decision

Traces

SSPTesting & Simulation backendIncludes

DecisionRunner

JMX

11

HostedTransparent

DecisionServices

WS WS, JSE, Message, your choice

RES Console

• Workload• Only management of the rule executable artifacts (RuleApp,

ruleset, eXecutable Object Models)• Execution restricted to diagnostic and interactive testing• Notification to eXecution Units when a new artifact comes

available• Design

• In memory stateful webapp on the top of the RES DB• Contains a JMX Model

• Kind of WAS Deployment Manager for Decision Services

12

RES What If Scenarios

• What if RES Console shuts down• What do we loose

– Deployment of new ruleapps/rulesets versions through HTTP(S)– REST Mgt API become unavailable– Execution statistics are on hold

• What do we keep– Rule execution remains fully operational– Direct deployment to the RES DB remains possible

• Restart to dynamically get back these functionalities• What if RES DB goes down

• XUs will not be able to read rulesets asked but not in memory• Execution remains fully operational for all rulesets already in

memory

13

RES Console In or Out a Cluster

• Deployed in a cluster• Shared RES DB• Pros

– Included in a Decision Server cluster – no adhoc server– No SPOF

• Limitations– RuleSession Interceptor not supported– 1 only RES Console diagnostic supported at a time

• Diags run in parallel not supported– Round robbin load balancing dispatches deployed artifact deployed

over http(s) across the different RES Consoles• All deployed artifacts are stored into the shared DB• Must refresh each RES Console with the DB to see the artifacts

deployed through the other RES Consoles– Imply a local deployment flag for the SSP

14

RES Console In or Out a Cluster

• Deployed apart a cluster• Install the RES Console on a separated server• Pros

– All features supported– All deployments go to the singleton web app and underlying DB

• Cons– Needs a dedicated server– SPOF but with limited impact if it fails

15

RES Notification Protocol

• IP• RES Console hostname/IP & port number specified in each

eXecution Unit to get notified• JMX

• Works in a WAS Cell or equivalent giving a JMX MBean server federation

• RES Console defines its RES model top level MBean• Each XU defines its MBean

16

ODM Standard Topologies

How to choose your ODM topology

18

What is my Decision Management scope?

Decision servicesSituation detection

Dev phase

Dev Integration Testing ProductionHA/DR

What is the workload nature ?

Event processing Remote/Local decision execution

Business decision Validation through Simulation and Testing

What is the platform?

distributed z: z/OS, z/Linux Hybrid Cloud

User Testing PreProd

ODM Standard Topology cheat sheet• Abbreviations for product scopes

• ODM: Full platform• ODMR: Rules only• ODME: Event only• ODM/DC: Decision Center• ODM/DS: Decision Server• ODM/DSR:DS for Rules only

• WAS ND topology concepts• ODM Profile templates delivered for

clustered DC & DS

19

Cluster

DMGRNode Agent

Server

Server

Node Agent

Server

Server

Dmgr’s nodeCustom nodeCustom node

Cell

ODM CopperDC + DS combined in a standalone server

ODM Bronze1 cell with 1 DC server + 1 DS Server

ODM Silver1 cell with a single cluster for DC & DS

ODM Gold topology1 cell with a DC cluster & DS clusterSeparation between business authoring from executionCover all product usage including business authoring, execution and simulation for rule and event Best tradeoff between Performance, HA, and cost

Platinum topologies All upper topologies that provide more workload separation and usage specialization Client App cluster can be added in this cell or run in a separated cellSimulation is isolated from business authoring & pure execution

ODM Topologies at a glance

Cell

ODM Standard Gold topology: DC + DS clusters + RES Console

Node Agent

Cluster Member 2XU, SSP, HTDS, in option Custom Decision Services developed on RES Rule Sessions

Cluster Member 1Decision Center

Deployment Manager

Node Agent


DecisionCenterCluster

Decision ServerCluster


Decision Center DB

Decision Server Rule

DB

Decision Warehouse

DB

IP Sprayer

IHS

Server 1RES Console

• Functional separation/isolation of workload

• Several cells for staged dev lifecycle

• RES XU deployed at Node level

• Datasources deployed at cluster level

• SSPs are directly referenced on each member from DC

• Remote ClientApp cluster run in another cell or can be added to the ODM cell

• Failover• Similar types of

workloads grouped together

• Can scale individual clusters as needed

Cell

ODM: DC + DS + RES Console + Remote Client App clusters

Node Agent

Cluster Member 3XU, SSP, HTDS, in option Custom Decision Services developed on local RES Rule Sessions


Deployment Manager

Node Agent


DecisionCenterCluster


Cluster Member 6XU, SSP, HTDS, in option Custom Decision Services developed on local RES Rule Sessions

Decision Center DB


DB

Decision Warehouse

DB

IP Sprayer

IHS

Server 1RES Console






Cluster Member 2Client Apps invoking remote Decision Services

Cluster Member 5Client Apps invoking remote Decision Services

RemoteClient App

Cluster

Cell

ODM Standard – Decision Center cluster


Deployment Manager

Cluster Member 2Decision CenterDecision

Cluster

Decision Center DB

Node Agent Node Agent

Cell

ODM Standard - Gold Decision Server

Node Agent


Deployment Manager

Node Agent




DB

Decision Warehouse

DB

IP Sprayer

IHS

Server 1RES Console





• SSPs are directly referenced from DC

• Remote ClientApp cluster run in another cell or can be added to the DS cell

Cell

ODM Standard - Single cluster

Cluster Member 1Decision Center,XU, HTDS, SSP,Potentially customer Decision Services based on Rule Session API

Deployment Manager

Cluster Member 2Decision Center,XU, HTDS, SSP,Potentially customer Decision Services based on Rule Session API

Application &

DecisionCluster

Decision Center DB


DB

Decision Warehouse

DB




Server 1RES Console


Cell

ODM/DSR - Testing & Simulation clusterDeployment Manager


DB

IP Sprayer

IHS

Cluster Member 1XU, SSP

Cluster Member 2XU, SSP

Test & SimulationCluster

Server1RES Console




ODM Standard - Standalone server

ServerDecision Center, XU,

SSP, HTDS, potentially custom Decision Service

Apps developed based on RuleSession API

Decision Center DB


DB

Decision Warehouse

Full ODM Standard running in a single serverIncludes all runtime parts for web decision authoring, test and execution

ODM - 2 servers

ServerDecision Center

Decision Center DB

ServerRES XU, HTDS,

potentially customer Decision Service based on the RuleSession API


DB

Decision Warehouse

Decision Center Decision Server

Decision Center on single server + Decision Server single server2 servers environments:

• no failover• 2 minimal # of JVMs• easy to create (create standalone profile) • Automatic DB creation

Operational Decision Management Scope

Functional restriction

Project Phase

Recommended Topology

ODM Standardmeaning DC + DS for Rules

Dev ODM - standalone

Test ODM – from 2 servers to DC + DS cluster (Gold)

Prod ODM DC + DS clusters (Gold)

ODM / Decision Center Dev ODM/DC standalone

Test ODM/DC 1 cluster

Prod ODM/DC 1 cluster

ODM / Decision Server Dev ODM/DS – standalone

Test ODM/DS - 1 cluster

Prod ODM/DS – 1 to 4 clusters

ODM/Decision Server for Rule Rule Dev ODM/DSR – standalone

Test ODM/DSR – 1 cluster

Prod ODM/DSR – 1 or more clusters

Rule Simulation Dev ODM/DSR/TS – standalone

Rule Simulation Prod ODM/DSR /TS– 1 cluster

ODM Standard Topology Catalog

Decision Service Development Life Cycle

Decision Management Development Life Cycle

In a strict isolation required between types/purposes of environments:• DEV: application developers have access to “DEV” only , purpose is application dev and basic function testing• SIT: integration testers have access to “SIT” only, purpose is integration and function testing• User Acceptance Test: user acceptance testers have access to “UAT” only, purpose is user testing• PREPROD: function and integration proven in earlier environments, purpose of this env is largely perf testing• PROD - apps were proven in earlier environments. Barely anyone has access to this env, it is tightly regulated.

DEV SIT UAT PREPROD PROD

DEV SIT UAT PreProd PROD

Decision Management Life Cycle Features are deployed in phased environment depending on purpose Business authoring

Business Object Model and Rule Editing Business Simulation & Test

Create and run Business Test Suites and Simulation to guide and improve decision logicChampion/Challenger comparison on KPI computation

ExecutionDecision run

Business Authoring

Execution

Business Simulation & Testing


Classic Decision Management Life Cycle Development of delimited Decision Services Business Authoring stopping most likely at UAT Business Simulation used until PreProd max

Business Authoring

Execution

Business Simulation & Testing

Agile Decision Management Life Cycle

Agile ODM Provide an ODM as a PaaS with change/extension of the decision logic without

starting back from Dev environment LOB leverage business authoring, simulation and apply their corporate decision

logic on their data Business Authoring and Simulation available in PreProd and even Prod More agility coming with more risk in production

Need governance and workload isolation to handle business authoring, simulation and execution and avoid regressions


Execution

Business Simulation

Business Authoring

ODM Deployment across phases

Each life cycle phase is mapped into an environment Each environment is a WAS cell (or equivalent in other AS) Decision Server deployed in isolation by phase Decision Center can be shared or isolated per phase

1 or more cells dedicated by phasePhase environment contain 1 DC (cluster) when business authoring required or

1 unique shared Cell with Decision Center + multiple Cells for Decision Server

This configuration is the most commonly usedMultiple DC are possible for sandbox

Shared DC across environments• One Decision Center managing multiple Decision Servers• DC clustered with a HA/DR DB• User access, Branches mgt for multi team/releases activities• DC and DS can be split into 2 isolated envs• Online and offline RuleApp/Ruleset deployment supported in DS

Decision Center

Decision Server

DC env

Decision Server

DS env

DM for Dev DS for Integration

Testing

Executable Decision artifact deployment over HTTP

Decision Server

DS env

Decision Server

DS env

DS for Business

Simulation and

Testing

DS for PreProd

Decision Server

DS env

DS for Productio

n

DS env

SDLC – Shared Decision Center

Pros One source of truth for business rule authoring Deploy on all Decision Servers; Multiple DS/RES are supported by DC HA with DC and DS in cluster Leverage project branch and merge in DC and deploy executable rules on the various DS Access to projects and DS scoped by groups & users in DC Isolation is preserved for the execution runtime

Cons Doesn't allow DC customization for one phase only (Dev, Test or Prod)

Comments Applied by most customers Leverage clustering and HA/DR DB

Best practice is to minimize the number of DCs and have 1 source of truth

Decision Center

Decision Server

ODM env

ODM for Dev

Decision Center

Decision Server

ODM env

ODM for Integration

Testing

Decision Center

Decision Server

ODM env

ODM for PreProd

Decision Center

Decision Server

ODM env

ODM for Production

ODM Standard Fully isolated environments

Project Export/Import

Executable artifact deployment

Decision Center

Decision Server

ODM env

ODM for Business

Simulation & Testing

One Decision Center per environmentDC Synchronization through Rule Designer Export/Import of Rule Projects between DCsExport/Import of RuleApps between DS/RESsDC and DS/Simulation in PROD depending on the Agility level

SDLC – Full Isolated Environments

ProsIsolation. We have a full ODM Standard runtime in a single cellHA with DC and DS when clusteredIsolation of authoring and execution by phase and cell

ConsMultiplication of JVMs1 DB per DC to administrateDC Repository content has to be synchronized across cells from Dev cell to Prod cellDS Repository is exported/imported, or rebuilt from Decision Center or Rule DesignerEach Decision Center will have its local version of Rule & Decision projects

CommentsThis pattern is applied by clients who privilege a full ODM Rule environment

including possibly Rule Designer at each phase of the governance cycle.

Business Rules everywhere

zRule Execution Server Stand-alone WebSphere Application

Server for z/OS

WOLA

CICS

COBOL Application

Rule Execution Server for WAS for z/OS

COBOL <-> JavaMarshaller

COBOL Generation

Rules

GeneratedCOBOL

JVM ServerzRES

zRule Execution

Server

IMS

COBOL Application

z/OS Batch

COBOL Application

COBOL Generation

Rules

GeneratedCOBOL

COBOL Generation

Rules

GeneratedCOBOL

DS StubDS Stub DS Stub

Decision invocations flavors in z/OS

1 2

3

1

3

2

zRES

RES on WAS on z/OS

COBOL gen

3 options

ODM in the Clouds• ODM Patterns

• in PureAS (client IT)• In Pure Service (SoftLayer)

• Bluemix Business Rules• A subset of ODM ready to use in Pay as you Go approach• Deploy your decision services in a PaaS• Based on Cloud Foundry• RES Console + HTDS + Rule Designer update site

Wrap up• ODM 8.7.0

• Standard for decision services• Advanced for situation detection with Decision Server Insights

• ODM Standard• Target stateless decision services• Versatile platform running on distributed, z/OS & clouds

– Run the same business rules everywhere• Open to platform combinations:

– ex DC in distributed + DS in distributed + zRES on z/OS– ex DC & DS in Pure Service + zRES

• Decision Center & Decision Server come with their own repository and are independent– DC & RD build and deploy executable rule artifacts to DS– All execs performed in DC, or RD for dev only– Testing and simulations are triggered from DC and delegated to DS

43

Best Practices

• Prefer 1 shared DC + multiple DS dedicated per env• Use the branch/merge for handle team collaboration & multiple

releases• Additional DCs are possible but adding artifact lifecycle mgt

complexity• 1 WAS Cell per stage env with DS• Leverage AS cluster for preprod and production envs• Add DB HA/DR for DC and DS production envs• Use collocated DC+DS for standalone dev envs• Make a dedicated simulation env when this workload become

significant

44

Best Practices

• Privilege a single RES Console that would be automatically restarted if a failure occurs

• Choose your RES notif in Decision Server• JMX notification when RES Console & all eXecution Units are

in the same AS admin scope (WAS cell or equiv)• IP notif when running on JEE/JSE without JMX federation

45

Wrap up

• Tell us if you are interested in new platforms & languages: embedded, mobile, Javascript in web browser

• Help the Lab by entering RFEs to push for a full RES Console cluster support

• For more• ODM topology DevWorks article• ODM Performance Redbook

46

Notices and DisclaimersCopyright © 2015 by International Business Machines Corporation (IBM). No part of this document may be reproduced or transmitted in any form without written permission from IBM.

U.S. Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM.

Information in these presentations (including information relating to products that have not yet been announced by IBM) has been reviewed for accuracy as of the date of initial publication and could include unintentional technical or typographical errors. IBM shall have no responsibility to update this information. THIS document is distributed "AS IS" without any warranty, either express or implied. In no event shall IBM be liable for any damage arising from the use of this information, including but not limited to, loss of data, business interruption, loss of profit or loss of opportunity. IBM products and services are warranted according to the terms and conditions of the agreements under which they are provided.

Any statements regarding IBM's future direction, intent or product plans are subject to change or withdrawal without notice.

Performance data contained herein was generally obtained in a controlled, isolated environments. Customer examples are presented as illustrations of how those customers have used IBM products and the results they may have achieved. Actual performance, cost, savings or other results in other operating environments may vary.

References in this document to IBM products, programs, or services does not imply that IBM intends to make such products, programs or services available in all countries in which IBM operates or does business.

Workshops, sessions and associated materials may have been prepared by independent session speakers, and do not necessarily reflect the views of IBM. All materials and discussions are provided for informational purposes only, and are neither intended to, nor shall constitute legal or other guidance or advice to any individual participant or their specific situation.

It is the customer’s responsibility to insure its own compliance with legal requirements and to obtain advice of competent legal counsel as to the identification and interpretation of any relevant laws and regulatory requirements that may affect the customer’s business and any actions the customer may need to take to comply with such laws. IBM does not provide legal advice or represent or warrant that its services or products will ensure that the customer is in compliance with any law.

Notices and Disclaimers (con’t)

Information concerning non-IBM products was obtained from the suppliers of those products, their published announcements or other publicly available sources. IBM has not tested those products in connection with this publication and cannot confirm the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. IBM does not warrant the quality of any third-party products, or the ability of any such third-party products to interoperate with IBM’s products. IBM expressly disclaims all warranties, expressed or implied, including but not limited to, the implied warranties of merchantability and fitness for a particular purpose.

The provision of the information contained herein is not intended to, and does not, grant any right or license under any IBM patents, copyrights, trademarks or other intellectual property right.

•IBM, the IBM logo, ibm.com, Bluemix, Blueworks Live, CICS, Clearcase, DOORS®, Enterprise Document Management System™, Global Business Services ®, Global Technology Services ®, Information on Demand, ILOG, Maximo®, MQIntegrator®, MQSeries®, Netcool®, OMEGAMON, OpenPower, PureAnalytics™, PureApplication®, pureCluster™, PureCoverage®, PureData®, PureExperience®, PureFlex®, pureQuery®, pureScale®, PureSystems®, QRadar®, Rational®, Rhapsody®, SoDA, SPSS, StoredIQ, Tivoli®, Trusteer®, urban{code}®, Watson, WebSphere®, Worklight®, X-Force® and System z® Z/OS, are trademarks of International Business Machines Corporation, registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the Web at "Copyright and trademark information" at: www.ibm.com/legal/copytrade.shtml.

http://www.ibm.com/legal/copytrade.shtml

Thank YouYour Feedback is

Important!

Access the InterConnect 2015 Conference CONNECT Attendee Portal to complete your session surveys from your smartphone,

laptop or conference kiosk.