IBM Support

Prerequisites and Maintenance for IBM Watson Machine Learning for z/OS, V2.1

Release Notes


Abstract

Hardware and software requirements for installing IBM Watson Machine Learning for z/OS, V2.1 on IBM z and Linux on Z or Linux.

Content

 IBM Watson Machine Learning for z/OS is an end-to-end enterprise machine learning platform that helps you create, train, and deploy machine learning models to extract value from your mission critical data on IBM Z, while keeping the data where it resides.

What's New

v2.1.0.1
  • A more flexible install option. Customers can choose to install WMLz base components    (where model and deployment management services and UI run) on z/OS only. They have the flexibility to connect a WMLz Linux Cluster (where WMLz IDE runs) or disconnect WMLz IDE from WMLz base whenever needed.
  • Support of the generation sample JCL for metadata DDLs for new installation and migration scenarios.
  • Support of skip level migration support from v1.2.0.2 to the current version for both Z components and Linux installer.
  • Support of Spark 2.3.
  • Support of using custom libraries in JVM-based online scoring engine, batch scoring and model evaluation process.
  • Support of anomaly notifications and a new heap map view for KPIs in Db2 anomaly detection solution template.
  • Support of Python 3.7 with scikit-learn 0.19.1.

v2.1.0
  • Ability to migrate from IBM Machine Learning for z/OS 1.2.0.2.
  • New option of installing training and model management services on a single node for both Linux and Linux on Z.
  • New option of using SAF authentication (as an alternative to LDAP authentication) for running WML for z/OS services on Z only.
  • Automated configuration of Spark and Anaconda runtime environments for WML for z/OS.
  • Support of AT-TLS client authentication for Spark job submission in batch scoring.
  • New sample JCL jobs for managing WML for z/OS services as started tasks on z/OS.
  • New sample notebook and API for using the automated feature engineering feature.
  • Integration of IBM SPSS Modeler for developing models on Linux on Z.
  • Enhancements in performance of online scoring of scikit-learn models.
  • Enhancements to Linux installers in support of upgrading WML for z/OS.

Prepare for the installation

For information concerning the preparation of the product installation see the following instructions

Maintenance Level

WMLz V2.1 maintenance level
Installation options APAR for IBM Z Container image installer for x86 Linux or Linux on z IZODA V1.1 maintenance level
V 2.1.0.1 WMLz base
PH13149
PH13150
PH13152
PH13153
PH13154
PH13155
PH13156
N/A
IzODA Spark 2.2.0 or 2.3.0 (FMID HSPK120) with PTFs UI63369 & UI63424 applied

IzODA Anaconda 3.6.0 or 3.7.0 with PTF UI63170 applied (FMID HANA110)

IzODA Mainframe Data Service 1.1 (FMID HMDS120) with UI63532 applied
WMLz IDE - Linux
can only be installed together with WMLz base
N/A 3 nodes container image installer for x86 Linux:
iwmlz_v2.1.0.1_linux_3-x86-64_installer
1 nodes container image installer for x86 Linux:
iwmlz_v2.1.0.1_linux_1-x86-64_installer
WMLz IDE - Linux on Z
can only be installed together with WMLz base
N/A 3 nodes container image installer for ubuntu Linux on z:
iwmlz_v2.1.0.1_linux_ubuntu_3-s390x_installer
1 nodes container image installer for ubuntu Linux on z:
iwmlz_v2.1.0.1_linux_ubuntu_1-s390x_installer
3 nodes container image installer for redhat Linux on z:
iwmlz_v2.1.0.1_linux_redhat_3-s390x_installer
1 nodes container image installer for redhat Linux on z:
iwmlz_v2.1.0.1_linux_redhat_1-s390x_installer
Add-on - ITOA Health Tree Application can only be installed together with WMLz base
PH13913
PH13914
N/A
V 2.1.0 Option 1- Installing and running all machine learning services, including scoring and model management services, on z/OS PH08690
PH08691
PH08692
PH08693
PH08694
PH08695
N/A
IzODA Spark 2.1.1 or 2.2.0 (FMID HSPK120) with PTFs UI60037 & UI60043 applied

IzODA Anaconda 3.6 with PTF UI59973 applied (FMID HANA110)

IzODA Mainframe Data Service 1.1 (FMID HMDS120) with UI58510 applied
Option 2 - Installing and running scoring services on z/OS and model management services on Linux on Z. PH08690
PH08691
PH08692
3 nodes container image installer for ubuntu Linux on z:
iwmlz_v2.1.0_linux_ubuntu_3-s390x_installer
1 nodes container image installer for ubuntu Linux on z:
iwmlz_v2.1.0_linux_ubuntu_1-s390x_installer
3 nodes container image installer for redhat Linux on z:
iwmlz_v2.1.0_linux_redhat_3-s390x_installer
1 nodes container image installer for redhat Linux on z:
iwmlz_v2.1.0_linux_redhat_1-s390x_installer
Option 3 - Installing and running scoring services on z/OS and model management services on Linux. PH08690
PH08691
PH08692
3 nodes container image installer for x86 Linux:
iwmlz_v2.1.0_linux_3-x86-64_installer
1 nodes container image installer for x86 Linux:
iwmlz_v2.1.0_linux_1-x86-64_installer
Add-on - ITOA Health Tree Application can only be installed together with option 1 PH08698
PH08699
N/A






























































V2.1.0.1 Incompatible change

  • The package "com.ibm.analytics.ngp.ingest.Sampling" is deprecated in Watson Machine Learning for z/OS 2.1.0.1. If you referred to the function "Sampling.trainingSplit" in this package, please rewrite it to use spark internal function "randomSplit" to split dataframe.
  • Watson Machine Learning for z/OS 2.1.0.1 upgraded the MLeap framework from version 0.7 to version 0.11. MLeap 0.11 is not compatible with 0.7. If you have Spark/MLeap models saved on the previous releases of Watson Machine Learning for z/OS, they won't be able to be deployed to online scoring server in MLeap format on v2.1.0.1. If you choose to deploy the model to MLeap engine of the scoring server, the model will be loaded in Spark format instead. For models deployed to MLeap engine in CICS scoring server in the previous releases, you need to go through the steps again as instructed in Prepare a model for online scoring with CICS ALNSCORE . If you want to leverage MLeap for high performance online scoring, then you will need to retrain and save the models again on v2.1.0.1.


Related information

IBM Watson Machine Learning for z/OS 2.1.0.1 documentation
Installing and configuring Watson Machine Learning for z/OS

Document information

More support for: IBM Machine Learning for z/OS

Component: --

Software version: 2.1

Operating system(s): z/OS

Reference #: 7049481

Modified date: 17 July 2019