AIOps：the next five years Memo

Forrester

where we are now
* I&O teams will need to realign much like DevOps
* A single source of monitoring truth
* Next gen systems – Contextual Intelligence

where we will eventually be
* trust for AI to make huge leap forwared
* ai driven technology will go beyond automation
* integrations accross the entire pipeline of IT services
* closer team collaboration: chatops
* leader of business insight data
* deigital transformations will change it

Moogsoft

why AI?
because the traditional programming techniques are pretty rubblish at dealing with highly complex and dynamic environment

Three Dimension of AI
* inference
* pattern discovery
* problem selection/Action

 

Moogsoft + AIops Memo

From Video:《Machine Learning & AIOps: Why IT Operations & Monitoring Teams Should Care》

Address：https://www.youtube.com/watch?v=UYNygjTY4xw&t=1799s

1. High Level Definition

• Originally “Algorithmic IT Operations”
• “Artificial Intelligence for IT Ops”
• Apply ML to IT Ops: monitoring, service desk, automation, other categories

1. Some basic knowledges about some general concepts of AI

2. What should I care: Open box vs Black Box machine learning

3. Open Box machine learning:

• Explainable:
• Editable: the logic from machine learning can be edited, so you incoporate human knowledge & experience
• Preview-able: you can preview the results

Transparency<->Control<->Trust

1. Category

• incident and problem management
• it operations analytics
• infrastructure management
• Capacity Management

1. ML inside AIops

1. IT Role is changing from many aspects:

• IT-outcome-focused, business-outcome-focused
• Cost-Controlling: Revenue-building
• Within IT: every where

1. behind it there is lots of changes

• infrastructure is changing
• software is changing

1. Lots of IT Role and software to run

• lots of noise
• lots of communication

1. Target of IT Ops

• keep up with more dynamic IT landscape
• support innovation and digital transformation
• scale up through automation, not brute-force human effort

1. Why machine learning can help

• learn about the environment
• handle the known and unknown
• cost effective and scale in real-time
• in 24 * 7 * 365
• Open box machine learning, always in controls

1. Architecture of Moogsoft