• Control your mainframe z/OS performance

  • Dynamic alerts on z/OS performance based on machine learning

  • Your whole performance data on one screen

ConicIT In a Nutshell  down-arrow

Machine learning based mainframe performance analytics  

ConicIT is studying the normal behavior of your z/OS performance per day and per hour, and is using advanced machine-learning algorithms for recognizing and alerting about mainframe performance anomalies at early stage, much before the performance-problems reach the mainframe capacity

View all mainframe performance data in one screen

Instead of hopping between your different monitors, or having dozen of screens showing different data from different monitors, ConicIT is aggregating all of the z/OS mainframe performance data and presenting it in one friendly web-application with graphs and tables

Connecting to your existing z/OS performance monitors

ConicIT is reading the z/OS mainframe performance data from IBM OmegamonTM, ASG TMONTM, CA SysviewTM, BMC MainviewTM, RMF, TSO, and others

Analyzing thousands of mainframe performance metrics per minute

ConicIT is reading thousands of mainframe performance variables from your z/OS monitors, storing them, and applying calculations for achieving new insights and calculated metrics

ConicIT z/OS Performance supervisor for Mainframe

ConicIT is a software solution for protecting your mainframe from performance problems by automatically recognizing them at early stage. ConicIT doesn't require you to set static thresholds for alerts, but rather implements dynamic thresholds based on statistics and also applies machine learning to give accurate alerts. The alerts are based on automatic recognition of performance anomalies of your z/OS.
ConicIT doesn't replaces your monitors (TMON, Omegamon, Mainview, and others) but rather connects to them and smartly analyzes the information they provide. ConicIT is installed on Linux machine (either standalone or SystemZ-Linux), and connects to your monitors like a regular user. But while human users are limited by their capacity, ConicIT is checking thousands of performance metrics per minutes, creating statistics over them, and recognizing the normal behavior of your mainframe z/OS performance per each hour of each day. Once ConicIT recognizes substantial deviation from normal behavior it will send you alerts with the relevant information. Typically these alerts will arrive at a very early stage, much before your mainframe response times and availability are at risk.
ConicIT also provides web-interface which allows you to see the information, including aggregated information from different monitors and sources, and important calculated information which are very hard for human user to perform. These calculated variables provides unique information such as summary of CPU and transaction-times per CICS-groups or real time CPU of started-tasks in contrary to the CPU-time since IPL which the monitors provide.
ConicIT can and should be used by few groups in the organization: First it should be used by your performance analysts who will be able to analyze the data and the alerts from ConicIT and to gain quick recognition and analysis of z/OS performance problems. In addition it can serve as a mainframe tool for the operators - which will get 24/7 alerts whenever z/OS performance anomaly is recognized. Lastly, it can be used by the system programmers who can use the graphs and information provided by ConicIT in order to better analyze mainframe performance anomalies related to their applications.

Automatic z/OS Performance Analytics

ConicIT Presentation

Early detection and proactive alerts on Mainframe performance problem

ConicIT presents a solution that automatically recognizes performance problems at their very early stages, allowing you to save valuable time, money, and most important helping you to keep performance within the required SLA levels. ConicIT is using state of the art technology in order to automatically recognize performance problems at their early stages. In contrary to traditional alerts system that are using static thresholds, ConicIT is using dynamic thresholds combined with machine learning technologies for recognizing deviations from normal behavior.
ConicIT is studying the typical behavior of your mainframe or open system per day and per hour, and can recognize when a performance problem starts to evolve. With ConicIT, you get the alert when the system still seems to be functioning well, much before the performance problems become acute, and much before it has effect on your business. Typically, when performance problem starts, it gradually affects more and more metrics, making it harder to analyze. The early detection of the problem with ConicIT along with the clear presentation of relevant information helps you to quickly find the root cause of the problem and to resolve it before it escalates.

Clear visualization is power

ConicIT is not only about alerts, but also a about great visualization of all performance-rated data which are processed and presented in the most clear and concise way. No more manual digging into monitor commands, no more manual comparison between dozens of metrics, no more wasting time on writing special applications for processing trivial performance data.
With ConicIT you get a friendly web site which provides you with all the information that you need regarding the performance of your system, and only the information you need. You can click on the type of information you want, and get clear tables summarizing all relevant data. You can get tables that summarize information, or you can dig in for more detailed tables. You can see graphs showing the fluctuations of the metrics through time, or even see graphs that compare between different variables. When watching a graph you’ll also see a green range around the curve, showing the typical range that this variable had in previous weeks in the same hour and same day of the week.
In many instances one table in ConicIT web-application is a combination of many variables from many different screens in standard monitor tools. Without ConicIT, you would have to waste a lot of time to get the full picture, and often it’s not even feasible. For example, if your monitor brings you CPU-time of started-tasks or of transactions, ConicIT can translate it into real-time CPU based on the rate in which the CPU-time changes through time.

ConicIT starts where z/OS performance monitors stop

ConicIT is not coming instead of system Z performance monitors (such as Sysview, TMON, Omegamon, and others). In fact ConicIT usually receives its data from such monitors, by sending them request once a minute, and parsing the required values out of their responses. You can think of ConicIT as your bright personal assistant which never gets tired, and prepares reports for you about the status of your system now and at any moment in the past, up to one year backward. It will process the system Z performance data to make it more meaningful. It will analyze it for you, and will tell you when something gets odd based on its familiarity of your mainframe behavior.

Save money and MIPS

ConicIT helps you to identify z/OS performance problems much sooner than before and helps you to dramatically reduce the time required for analyzing such problems once recognized. It has three implications:

  • Wasting less hours per month in analyzing performance problems
  • Lower MIPS consumption, due to identifying of problems you wouldn’t even notice even in the long run, and also due to solving problems at their early stages
  • Preventing downtime of your mainframe z/OS

In addition, ConicIT can also alert on PSLC terms (e.g. when the MSU of production LPAR is not high enough compared to the total usage).
ConicIT can inspect and guard on any type of variable, including (and not limited to) CPU (of any entities such as LPAR, CICS, PLEX, or started task), transaction-rate, response-time, IO-rate, memory consumption, DB2 variables, resources-wait and locks, length of queues and many more.

ConicIT is designed for you

ConicIT is highly customizable. It’s not only that we configure ConicIT to adapt to the configuration of your Z system, but we can add new abilities and tailor ConicIT to support new types of variables and new types of calculation within few days from request. ConicIT comes with a very powerful configuration tool and plugin-scripts, which allow us to fit our software for your needs. We can sample and inspect any type of data you need, from any type of source, we can add any type of smart calculations on variables and any type of alerts.

Analyzing z/OS mainframe performance without affecting performance

In ConicIT we know that system resources are expensive, and we believe that software that aims to protect you from overload must not add load to your system. ConicIT is agent-less and virtually has no effect on the load of your mainframe. ConicIT can run on a separated Linux machine, and connects to the monitors of your mainframe with TN3270, just like a regular user of TMON, Sysview, Omegamon or others. Thus, the effect on your mainframe is negligible.

Lets talk about Unified Presentation

Aggregation of performance data by ConicIT

Aggregation of performance data by ConicIT

ConicIT aggregates and summarize performance data from different monitors

In this example, ConicIT receives CICS performance from different monitors, summarizes them by groups of CICSes and shows the total CPU and transaction-rates of each group. The detailed information is also available for the user to see.

Mainframe performance graph with green path representing the expected bahavior

Mainframe performance graph with green path representing the expected bahavior

ConicIT Shows mainframe performance graphs along with their expected behavior as green range

ConicIT can show graphs of performance variable up to one year backward. In addition, you can see the green range which represents the normal expected values based on the statistics for the same day or week and hour.

Comparisons graph of LPARs MIPS by ConicIT

Comparisons graph of LPARs MIPS by ConicIT

ConicIT allows you to easily compare graphs of metrics

In this example, ConicIT compares the MIPS consumption of different LPARs. It's possible to compare graphs over long periods of up to one year long.

Find the bottleneck resources

Find the bottleneck resources

Find the resources that causes bottlenecks

ConicIT aggregates the wait-times per resource, and allows you to identify bottlenecks.