Causality Event Correlation Using Artificial Intelligence · 2018-10-25 · Causality Event Correlation Using Artificial Intelligence Federator.ai® cross-layer causality event correlation

Causality Event Correlation Using Artificial Intelligence

© 2018 ProphetStor Data Services, Inc. All Rights Reserved

1

Causality Event Correlation Using Artificial Intelligence

Causality Event Correlation Using Artificial Intelligence

Table of Contents

The future of IT Ops .......................................................................................................................... 2

Causality event correlation algorithms ............................................................................................... 3

Learn more ....................................................................................................................................... 3

© 2018 ProphetStor Data Services, Inc. All Rights Reserved

2

Causality Event Correlation Using Artificial Intelligence

The future of IT Ops

Artificial Intelligence for IT Operation (AIOps) is the next generation of technologies in IT management. As

Gartner defined in the Market Guide for AIOps Platform:

AIOps platforms are software systems that combine big data and AI or machine learning functionality to

enhance and partially replace a broad range of IT operations processes and tasks, including availability and

performance monitoring, event correlation and analysis, IT service management, and automation.

AIOps uses AI and machine learning technologies to turn copious amounts of data into meaningful insights

and foresights, allowing systems to function independently, reduce operational costs, and increase efficiency

and productivity. For example, AIOps has machine learning to detect unusual event sequences in the logs,

which are collected from IT devices. This feature can largely shorten the time IT administrators spend on

addressing possible issues on all of the logs.

AI is becoming omnipresent as industry leaders such as Apple, Microsoft, Google, and Amazon use AI for

their products and services. Traditional IT Ops cannot keep up with the dynamic landscape of IT operations

and business models.

Today's AIOps platforms go beyond traditional monitoring solutions like APM (Application Performance

Monitoring and Management). Traditional monitoring solutions may only show the performance metrics

collected, and a more advanced solution is needed to decrease the mean time to detect (MTTD) and the

mean time to resolution (MTTR).

An AIOps platform detects anomaly events, correlates the causality of events

between different layers, and shows the related path routes and the affected

entities and performances by the anomaly events. AIOps provides strategic

insights to resolve a problem and foresights to help execute proactive actions

to prevent repeated incidents. Prophestor's Federator.ai® takes AIOps a step

further

ProphetStor's AIOps software, Federator.ai®, offers advanced solutions that APM cannot provide. Among

other AI features, such as detecting and predicting hardware issues, Federator.ai® correlates causality events

using the different types of AI technologies to quickly track down root causes and effectively and efficiently

fix problems.

Federator.ai® is equipped with an AI engine called Data Correlation and Impact Prediction Engine (DCIE) to

build correlations between different entities of application, virtualization and physical layers. It uses graph

models to process the relationships among objects. The graph-based data structure and search algorithms

help Federator.ai® find the possible causality events of an application anomaly more effectively and

efficiently in a large-scale IT environment.

© 2018 ProphetStor Data Services, Inc. All Rights Reserved

3

Causality Event Correlation Using Artificial Intelligence

Federator.ai® cross-layer causality event correlation process.

Causality event correlation

algorithms

In the figure above, Federator.ai® correlates the

causality events of the underlying layers when an

anomaly is detected in the upper application

layer. In the left diagram (Step 1), a transaction

latency anomaly event is detected in the

application layer, which is raised from a database

system (DB). In this example, users need to find

the possible causes of this event.

Federator.ai® DCIE uses multiple algorithms

based on four types of event correlation analysis

to correlate causality events and diagnose the

root cause of the application anomalies:

Time-related: Events usually occur in the same

time frame or are based on a certain time-order

pattern.

Dependency-related: Events with the related

entities on the different layers usually occur

together.

Content-related: Events with similar contents

usually occur at some specific points of time.

KB-related: KB (Knowledge Base) may include

the indicators of the causality among events

and the root causes of the problem.

In Step 2 of the above figure, a CPU loading

anomaly is also detected in one of the ESXi hosts;

at this point, it is not known whether this ESXi

host is the cause for the transaction latency

anomaly in the DB.

Step 3 shows how DCIE correlates the anomaly

events in CPU loading and network received traffic

in the physical layer to the database event in the

application layer. Federator.ai® DCIE builds this

correlation by using causality event correlation

algorithms to determine if any combination of

events has a higher possibility of correlation

patterns than others from all of the events

detected in different layers.

Traditional IT solutions require the user to find all

anomalies without AIOps, a time-consuming

process. As the figure shows, IT administrators are

no longer left solely to solve anomaly events

detected by APM or reported by users.

Federator.ai® can help them narrow down the

problems to a few suspicious physical or virtual

nodes or devices.

Learn more

To learn more about ProphetStor AIOps solutions,

visit us at http://www.prophetstor.com/federator-

ai/