IPA Bridging the Gap between Fault and Performance Management - PowerPoint PPT Presentation

IPA Bridging the Gap between Fault and Performance Management www.rizotec.com

What do mobile operators need to do? Monitor their networks carefully Understand growth trends in their network Identify anomalies across the network in real-time Locate problems before they start affecting the network

Current solutions Fault Management Systems Offer comprehensive visibility to the health status of network elements Answers simple binary questions (fault / no fault) Output: Real-time alerts Drawback: Alerts only on fault (no trends/anomalies, no performance)

Current solutions Performance Management Systems Enable managers to prepare the network for the future and determine its efficiency Uses complex KPIs to identify trends and anomalies in the network Based on historical data Output: performance report Drawback: not in real-time, no alerts

The Gap Between Fault and Performance Management The Gap Identify problems in real-time before they occur Alerts based on complex performance indicators Real-time trends and anomalies

The Gap Between Fault and Performance Management PERFORMANCE Trends Reports FAULT Historical Real-time Alerts Fault / No Fault

Bridging the Gap Performance Anomalies IPA PERFORMANCE Trends Reports FAULT Historical Real-time Alerts Fault / No Fault

IPA - Intelligent Performance Analyzer Bridges the gap between fault and performance management Identify performance deterioration Near real-time alerts based on compound performance data analysis Identifying problems before they start affecting the network Reduce and prevent downtime

Technology IPA collects large amounts of data from various sources (OSS, BSS, CRM, DWH…) Real-time analysis Multi-dimensional model > Entity > Time > Configuration

How it Works? Connect to data sources Define KPIs and Rules Tuning Deploy real-time analysis

IPA Studio Loader Builder Defines the connection to the various data sources in order to replicate the data to IPA’s internal database

IPA Studio KPI Builder Defines the KPI’s that will be used in the performance data analysis. > Nominal KPI’s > Historical KPI’s > Time aggregated KPI’s > Configuration aggregated KPI’s

IPA Studio Rules Builder Rules are created by using the KPI’s to define logical conditions for raising alarms. > Static threshold > Dynamic threshold > Configuration threshold An alarm can be a collection of multiple thresholds

IPA Studio Alarms Builder Defines when to run the rules and where to send the alarms. > SNMP > Email > SMS

IPA Studio Simulator > Before deploying a rule it can be tested and fine-tuned on historical data. > Optimization of existing rules. > Try and go, Try and change, Fine- tuning, Optimization

Runtime Data Loader > Connects to the various data sources at predefined intervals > Replicate the data to IPA’s internal database > Independent, non-intrusive

Runtime Real-time Analyzer > Rules Engine – Analyzes in real-time the constant stream of data, based on the predefined rules > Alarms Engine - Alarms are raised based on the rules, while notifying the relevant stakeholders

Runtime Reporter > Generates textual and graphical reports based on the KPIs > Predefined reports > User-defined reports

Example - Abnormal Release Rate The goal locate technical problems in the network by identifying abnormal patterns in the release rate of session. The Rule comparison between abnormal release rate during the last 15 minutes to the average on the same period (e.g. Monday 09:45 AM) during the last 4 weeks. Alarm will be sent after the rule is triggered 3 times in a row.

Example - Abnormal Release Rate

Example - Abnormal Release Rate Results > On average each month between 15 and 20 customers problems were located by IPA > Early identification allowed the MVNO to fix the problems. > $2500 on average saved each month for uncharged traffic.

Example – Traffic Degradation The goal to locate technical problems in the network by identifying abnormal capacity degradation. The Rule identify a decrease in traffic of over 30% by comparing real-time traffic rate and base station’s average for the past 10 weeks. Alarm is raised on a certain threshold.

Example – Traffic Degradation

Example – Traffic Degradation Results > Prevented decrease in consumption > Reduced the length of such problems by 80%

Example – Non-optimal Configuration The goal to identify non-optimal roaming configuration on cellular devices in order to optimize configuration (increase preferred network usage). The Rule identify customers with a total usage more than x MB in a unit time (for example 1 hour) and more than y% of their traffic served by non-preferred operators

Example – Non-optimal Configuration

Example – Non-optimal Configuration Results > Corrected configuration reduces the end-customer costs and at the same time raises the MVNO’s profit. > Approximately 1.5% of roaming usage is corrected > on average $Saved 1 per event

Summary > Connects to DB/Data warehouse, Files, etc. > GUI-based rules builder (no developers needed) > Real-time analysis of compound performance data > Real-time alarms based on complex rules > Real-data simulator > Graphical and textual reports

Thank you.