Overview of BMC AMI Ops Insight

BMC AMI Ops Insight represents a progressive solution that uses machine learning to anticipate and preempt potential disruptions before they occur. By ingesting your historical data, it constructs models that represent the expected patterns of operation within your systems.

Leveraging these models alongside real-time data and employing multivariate analysis through cutting-edge machine learning techniques, BMC AMI Ops Insightswiftly identifies deviations from the norm within your system.

Its intrinsic domain and data science expertise ensure meticulous monitoring of the crucial performance metrics you select, providing a comprehensive description of any anomalies. This meticulous approach significantly diminishes the time and effort typically required for configuring a solution through trial and error, and false alarms, and maximizes lead-time available for remedial actions.

Process

The following figure shows the overall process:

BMC domain experts build algorithms to identify KPIs (Key Performance Indicators) and groups of connected KPIs that can indicate problems. This minimizes your cost because only the relevant KPIs are monitored.
Your historical data is used by the product to identify normal levels for the KPIs in your environment and then train the models. This means that the models are not generic models, but are customized to your environment.
The product then uses multivariate analysis to score your real-time data, comparing it with the data in the models to detect exceptions.
When the product detects anomalies, it looks for trends to detect if you are currently experiencing a problem, or that you are about to experience a problem. Reporting trends rather than individual anomalies maximizes accuracy and minimizes false positives.
Important
In some cases, for especially sensitive KPIs, the product calls out individual exceptions as soon as they show an anomaly.

Data source

BMC AMI Ops Insight uses SMF records as source data (currently, SMF 70-79 and SMF 100 records). The SMF 100 records contain hundreds of system-level statistics cut at one minute intervals. The SMF 70-79 records are cut at 15 minute intervals (by default). These are overall system-level indicators based upon the resource usage on the system during the last interval. These represent a useful and meaningful set of data to measure system-level activity based on the resource usage statistics. The product uses a selected subset of these statistics (KPIs), which are grouped based on specific areas of measurement including CPU time, storage, contention, throughput, and others.

Components and flow

BMC AMI Ops Insight includes the following processes:

Training—This involves building a mathematical model based on historical data.
Scoring—This involves compiling data and calculating a score to compare to a model.

The data flow is same for the Training and Scoring processes.

This diagram illustrates how data proceeds through the product's components. To see the detailed architecture of BMC AMI Ops Insight, see Environment-variables.

BMC AMI Ops Insight process flow and architecture

The following table describes the components:

Component	Description
Data Ingest	Processes historical and real-time raw data
Data Prep (Data Preparation Address Space (z/OS)	An interface to fetch data from BMC AMI Ops Monitor products.
BMC AMI Ops Monitor	BMC AMI Ops Insight uses your monitor products for detailed analysis processing.
Historical Data	Historical SMF data from the last 4-6 weeks.
Real Time Data	Current SMF data sent to BMC AMI Ops Insight by BMC AMI Datastream for Ops Insight
Model Generation (Training)	Builds a mathematical model based on historical data.
TOMCAT REST Interface (USS)	A bridge between the Data Ingest or Data Prep component and BMC AMI Manageror Model Generation .
BMC AMI Manager	It process the real-time data and then uses the scores to perform multivariate analysis to create event for a detected anomaly.
Container (z/CX or X86 Linux) – For charts and timeseries DB	(Optional) Container for displaying charts using timeseries database.
Container (z/CX or X86 Linux) – For PostgreSQL DB	(Optional) Container for displaying charts for Workload analysis.
Scoring Engine (Use models to evaluate data)	Uses your models and the real-time data to calculate a score, which is a measurement of how much a KPI deviates from the normal.
HSQL database server (Model Store, History DB)	Stores models and scored data.
SMF Record Handler	An interface between BMC AMI Datastream for Ops Insight and Data Ingest component
BMC AMI Ops UI Discovery	A server that is installed on the mainframe as part of the BMC AMI Ops Infrastructure standard installation.
BMC AMI Ops UI server	A web server that authenticates the user during login. It then connects to the service registry to get a list of registered services, and verifies which services the user can access.
BMC AMI Ops User Interface	BMC AMI Ops user interface

Training

Historical data is ingested into Data Ingest component for processing.
After the data is processed, it is sent to Model Generation.
Model Generation generates a set of models based on the historical data.
The models are stored in the database and are available via the browser-based UI for scoring.

Scoring

Real-time data is collected by BMC AMI Datastream for Ops Insight (separate license is not required).
The data is then sent to the SMF Record Handler. SMF Record Handler passes the data to the Data Ingest for processing.
After the data is processed, it is sent to BMC AMI Manager.
BMC AMI Manager evaluates the data against a current model.
The results are stored in the database and displayed on the browser-based UI.
(Optional) If you have BMC AMI Ops Monitor products, BMC AMI Ops Insightcan connect to the monitors to generate more detailed analysis.