Planning for deployment of Z Anomaly Analytics with Watson

Before you deploy the IBM Z® Anomaly Analytics with Watson™, you must plan for its installation and configuration, and ensure that your environment meets the system requirements for all components. You can use the log-based machine learning system, the metric-based machine learning system, or both machine learning systems in IBM Z Anomaly Analytics with Watson. The installation and configuration steps vary, depending on the types of machine learning that you want to use.

Decisions to make about your installation

Decide whether you want to use log-based machine learning, metric-based machine learning, or both. For an overview of these types of machine learning, see Z Anomaly Analytics with Watson overview.

Some components of IBM Z Anomaly Analytics with Watson must be installed on an IBM® z/OS® UNIX System Services system, and some must be installed on a Linux® system. The common services are the only components that can be installed on either a z/OS UNIX System Services or Linux system. If all of the following statements are true, and are unlikely to change, you can install the common services on a z/OS UNIX System Services system:
  • You do not plan to use the log-based machine learning system.
  • You do not want to view metric-based machine learning scores in the context of a topology.

For all other scenarios, you must install the common services on Linux.

Installation overview

Install the components of IBM Z Anomaly Analytics with Watson as indicated, based on whether the common services will be installed on a z/OS UNIX System Services or Linux system.
If the common services will be installed on a Linux system, install the components in the following order
  1. Install and configure the Problem Insights server.
  2. If you plan to use metric-based machine learning, install the IBM Z Resource Discovery Data Service component, which is provided as a software container that is compliant with the Open Container Initiative (OCI).

    This installation of the component must be separate from any installation of IBM Z Resource Discovery Data Service that is used with IBM Cloud Pak® for Watson AIOps.

  3. Install and configure the OCI-compliant software containers that are required for machine learning. You can select to install software containers for log-based machine learning, metric-based machine learning or both. Each selection will automatically install required common services, including a gateway service and an authentication service.
  4. Configure integration between the Problem Insights server and the authentication service.
  5. If you plan to use metric-based machine learning, install and configure the metric-based machine learning system on z/OS UNIX System Services on the z/OS system where IBM Watson® Machine Learning for z/OS is installed.
  6. Install and configure the Z Common Data Provider on each z/OS system from which you will analyze log data, SMF data, or IMS data.

Also, review the information in Linux: preparation of installation media.

If the common services will be installed on a z/OS UNIX System Services system, install the components in the following order
  1. On the z/OS system where IBM Watson Machine Learning for z/OS is installed, install and configure the following components:
    • Gateway service
    • Authentication service
    • Problem Insights server
    • Metric-based machine learning system on z/OS UNIX System Services
  2. Install and configure the Z Common Data Provider on each z/OS system from which you will analyze SMF or IMS data.

Also, review z/OS: preparation of installation media.

Software version requirements for z/OS systems and subsystems from which you want to gather data

Ensure that your environment meets the software version requirements for the z/OS systems and subsystems from which operational data is gathered.

The following software versions are required:
  • IBM z/OS 2.3, 2.4, or 2.5
  • IBM CICS® Transaction Server for z/OS 5.3, 5.4, 5.5, or 5.6
  • IBM Db2® for z/OS 12.1
  • IBM IMS 15.1 or 15.2
  • IBM MQ for z/OS 9.0, 9.1, or 9.2

System requirements and prerequisite planning tasks

Ensure that your environment meets the system requirements for all components of IBM Z Anomaly Analytics with Watson, and complete any prerequisite planning tasks that are described.