Technology

The Mainframe Challenge

Mainframes are the most stable IT environment in existence today. The problem is that management takes that stability for granted, and now wants it at reduced cost. Add that to dwindling number of mainframe IT skills in the marketplace, and it is clear that there is a need for innovation in mainframe service management technology.

 

ConicIT provides just that. It has developed products based on its patent pending technology that are easy to use and integrate easily with existing monitoring tools.

 

ConicIT provides innovative mainframe service management solutions that enable

to reduce SLA costs and achieve First Fault Problem Resolution.

 

 

ConicIT Technology Highlights

  • Easy to install, no integration needed for the monitored systems (uses data from existing performance monitoring tools)
  • Non-intrusive set-up and installation on a standard Linux server
  • Real-time capture and aggregation of all relevant problem data from existing      performance monitors
  • Self-learning capabilities including automatic adjustment to environment changes
  • Multi-expert prediction algorithms for look-ahead detection of emerging problems
  • Proprietary mathematical and statistical models to track behavioral patterns and to identify normal vs. abnormal behavior
  • Composite problem detection criteria, taking into account the inter-correlation between performance variables of interest
  • State-of-the art data cleansing algorithms that address the non-stationary and non-linearity of the monitored signals


Behavioral Analysis Technology Solution

 

ConicIT mainframe software solution proactively optimizes mainframe costs and keeps Service Level Agreements (SLAs) on target. The software automates the process of system monitoring and control, constantly analyzing performance data using sophisticated Behavioral Analysis Technology to predict and manage mainframe performance in real-time.

Key Features

  • Creates dynamic thresholds using proprietary learning algorithms
  • Compares in real-time between planned and actual data
  • Analyses problems and provides the root cause
  • Provides key diagnostic data and recommendations
  • Detects anomalies and predicts failures using proprietary mathematical models
  • Consolidates and unifies performance data from multiple systems and LPARs
  • Automates data collection
  • Uses expert systems and Artificial Intelligence (AI) to learn the normal behavior of each variable

How does it Work?

Real-time System Monitor and Analysis

ConicIT is a non-intrusive software platform that runs on a dedicated Linux server – either external or zLinux based. Easily implemented into existing infrastructure ConicIT manages other z/OS monitoring products (e.g. OMEGAMON, TMON, Sysview, Mainview) and uses the available information to analyze system performance and predict trouble ahead. Whether in the form of a brewing performance problem, or a pending increase in capacity charges, the system immediately alerts operations and preserves the data necessary for problem analysis and repair.

 

ConicIT Architecture

Using proprietary mathematical models the ConicIT server identifies behavioral patterns of critical computing resources and the relations between them to establish a baseline of normal behavior specific to a certain hour or day. The Artificial Intelligence (AI) engine works to catch any anomaly as it occurs in order to provide alerts on performance deviations, discrepancies, and other indicators of abnormal behavior in real time. Additionally all monitoring statistics are recorded and subjected to data mining and statistical analysis

Detection of the root cause of disorders ensures that IT personnel manage to solve the actual cause of the incident and don’t spend valuable time and resources trying to battle the incident symptoms.

Self Learning Capability

 

The self-learning mechanism is an offline algorithm that uses historical data to periodically check the prediction algorithms in order to select the most appropriate set of active models for the current state of the monitored system. The self learning module applies the historical data to candidate prediction algorithms to create prediction models and a weight for each model.

ConicIT comes out of the box with a large set of prediction algorithms appropriate for most mainframe installations. Achieving optimal performance for a specific installation requires about a week of site specific professional services tailoring and an equivalent automated learning period.

Unified Performance Data

ConicIT Viewer for Multiple LPAR’s

ConicIT resides outside the mainframe (or on a zLinux partition), and uses data from existing monitors that enables it to collect information from any relevant LPAR simultaneously as needed. Existing mainframe monitors only monitor a single LPAR at a time. This enables ConicIT to provide a holistic view of the whole mainframe environment, displaying a dynamic view of each LPAR’s status in relation to the status of the entire machine.