Difference between revisions of "PMQ"

Revision as of 19:55, 8 March 2017

HDD Health Analysis

Function

For device with HDD failure prediction intelligent application, users can know current failure degree of HDD that RMM agent deliver. From current failure degree, remind users whether check or maintain releated equipments of device (For instance, system CPU and Memory), then avoid HDD fails in real.

Implementation

Failure degree is based on calculation of HDD failure prediction intelligent application in Device. It can collect current some indicative data related with HDD (Indicative data contain statistic of Abnormal power off count and HDD write/read issue), and put these data in trained regression analysis model that HDD failure prediction intelligent application offers to calculate failure degree of HDD. Once failure degree is over 0.385, this predictive result means current satus of HDD is with failed risk, then users have to proceeded with related maintain proccess.

Trend Forecasting Monitor

Function

• Predict IPC/IOT Data in Future
• Automatically Monitor Alteration of IPC/IOT Data

For sepecific device or sensor data(HDD Health, System CPU Usage and so on), RMM offers prediction function(Trend Forecasting Monitor). It can help users know abnormal data status as soon as possible to prevent some damages.Predictive Type of Trend Forecasting Monitor are hourly averages of data, and they contain averages of current hour and even next 3 hours in furtue.

Otherwise, Trend Forecasting Monitor also offer function of automatical monitoring hourly averages of IPC/IOT data. If it happens situation that hourly average change significantly, RMM will save as events to notify users.  

Predictive Process

The first condiction is predictive object has at least the newest 48 history hourly averages in Database, and these samples are calculated by time-series prediction algorithm. Time-series prediction algorithm RMM adopts is Exponential Smoothing, and is based on APIs in Weka machinne learning java libary. By the above Process, it can produce values of prediction(averages of current hour and next 3 hours), and Trend Forecasting Monitor takes these values and the newest 3 history hourly averages to compare with hourly averages in yesterday. If there are significant differences between values of prediction and hourly averages in yesterday, the status of analytic result is "Potential Risk". If there are significant differences between the newest 3 history hourly averages and hourly averages in yesterday, the status is "Doubtful Risk". The standard of significant differences is based on Tukey's Range Test, and the formula is: [Q₁-3(Q₃-Q₁), Q3+3(Q₃-Q₁)]. Q₁ and Q₃ are the lower and upper quartile of hourly averages in yesterday. If objects are in the above range, there are not significant differences between them, then we can call status of Object is "Normal". 

Predictive Result

Result for each object contains 3 parts: result status, status information and event record list. First, for result status. except "Potential Risk", "Doubtful Risk" and "Normal" in the above statement, there is a "Data Error" situation, and it means disable predicting because of some data problem. For instance, miss some data in the newest 48 hours or query data from database error. Second, according to result status, web UI will show corresponding information, and major statement is explain reason for status and give some suggestions to users.Finally, event record list will show total records for significant change, then it can help users to trace initial time of abnormal data.