aiops + it4it - it4it company | it4it training€¦ · product backlog & requirements incident...
TRANSCRIPT
June 2019 1
Rob Akershoek © 2019
Value Outcome
Create a Smart and Automated IT Value Chain
AIOps + IT4IT
Rob Akershoek• IT Management Architect at
Fruition Partners,
• Chair IT4IT Forum (The Open
Group)
Email: [email protected]
Twitter: @RobAkershoek
June 2019 2
Rob Akershoek © 2019
IT Department
Is your IT organization also struggling to get
the right people?
June 2019 3
Rob Akershoek © 2019
June 2019 4
Rob Akershoek © 2019
Machine Learning – Build a Model (using Algorithms)
Is it a Dog or a Muffin? How difficult is this algorithm?
June 2019 5
Rob Akershoek © 2019
Deep learning and neural networks to recognize patterns / pictures
• Is it a dog or a muffin?
June 2019 6
Rob Akershoek © 2019
Storyboard Content of this presentation
1. Setting the Scene
• WHAT ARE KEY CHALLENGES OF THE MODERN IT ORGANIZATION?
• WHY IS IT AUTOMATION BECOMING A CRITICAL SUCCESS FACTOR?
2. Current state of IT Automation
AIOPS INTEGRATIONS
CURRENT LANDSCAPE OF PROCESS, TOOLS, DATA
3. Target IT Operating Model
CREATING A TARGET OPERATING MODEL
FORCES SHAPING THE FUTURE OF IT
6. How to get started on AIOps?
WHAT ARE THE STEPS?
4. AI in the end-to-end IT Tool Chain
AUTOMATION, AI & ML
• HOW WELL DO WE AUTOMATE IT PROCESSES TODAY?
• HOW DOES THE OVERALL IT AUTOMATION TOOL CHAIN LOOKS LIKE?
• WHAT IT AUTOMATON CAPABILITIES DO WE NEED?
• HOW DOES THE AIOPS TOOL CHAIN LOOKS LIKE?
• WHAT IS THE APPROACH TO START WITH AIOPS?
5. AIOpsArchitecture
• HOW DOES AI AND AIOPS FIT INTO THE OVERALL IT MANAGEMENT LANDSCAPE?
June 26, 2019
Setting the scene
May 2019 8
Rob Akershoek © 2019
Smart Car, Smart Home, Smart City, Smart Factory…
Connected Digital Ecosystem
Optimize User experience / Journeys
Real time insight to optimize decision making
The Connected Digital Ecosystem
But how smart is your IT organization?
Seamless integrations
May 2019 9
Rob Akershoek © 2019
Chatbots / Virtual Assistants
Can you assist you?
Yes, it seems my
account or
credentials are not
correct?
I see that you have
issues logging into
the CRM portal….
Sure, lets do it!
I see that your current
laptop is almost end-of-
life. Can I ask you
some questions to
determine what new
model is best for you?
I see that you are a
heavy power user, so
based upon you current
profile you need a more
high end machine..
….Let me check…I
see that your HR role
has changed
yesterday….
Google assistant, Amazon Alexis, IBM Watson, Siri, …
June 2019 10
Rob Akershoek © 2019
How smart is your monitoring and event management?
• Create business rules (scripted knowledge) If this…then that (IFTTT)
June 2019 11
Rob Akershoek © 2019
CynefinModel(“kin-ev-in”)
Ordered
(Cause and effect
relationship)
The same thing will happen in the same way twice.
The difference is in Simple a best practice can be applied whereas in Complicated you have to do analysis (with different specialists) to find it.
Unordered
Chaotic: The priority is containment and damage reduction, and any initial solution is focused primarily on these objectives.
Rule based
Complex systems require experimentation and investigation (exploration)
Solutions are emergent rather
than predetermined
Disorder
May 2019 12
Rob Akershoek © 2019
WHY IS AI THE FUTURE OF
IT OPERATIONS?
May 2019 13
Rob Akershoek © 2019
The cruel reality: This is how we are automating IT today...
June 26, 2019 14Fruition Partners
Data driven IT managementFocus on the data needed to manage IT…
Software
licenses
Discovery data
Configuration data
Application & system Logs
Test logsMonitoring data & events
IT budget &
costs
Availability &
Performance data
Demands
Product Backlog
&
Requirements
Incident &
problem
Change &
Release
Service & Application
portfolioBuild logs
Contracts &
License
Enterprise
Architecture Service
catalogue
Service
RequestsEpics / User Stories
IT risks
API
register
Defects
Source
code
Consumption &
usage data
Vulnerabilities
Security eventsTechnology
life cycle
Test
cases
Knowledge
items
IT Standards
User identities &
Access Rights
Job
Schedules
Project Portfolio &
Portfolio Backlog
Lacking end-to-end traceability and transparency…
Strategy to
Portfolio
Requirement to
Deploy
Request to
Fulfill
Detect to
Correct
Documents
June 26, 2019 15Fruition Partners
Data driven IT management
ONLY 5% OF
AVAILABLEIT DATA IS USED
May 2019 16
Rob Akershoek © 2019
Why is IT automation and AI becoming a critical capability for managing the digital enterprise?
More vendors in the
IT-ecosystem
Multi-vendor SourcingCloud providers
+More changes &
releases
Continuous delivery, patches, …+
Controls & Regulatory
Requirements
Control monitoring,Compliance+
Increased business
dependency on IT More services and
components
Internet of ThingsCloud servicesAPIs+
Exploding
IT management data
Logs,Monitoring data,Config. data…+
More requests e.g.
access…
Service Request e.g. JML
+
Faster, Better, Cheaper with
Lower risk+
Create a Streamlined,
Automated and Smart
IT Value Chain
More security
events & threats
Security Threats, Events
+
Increased consumption
& fluctuation in demand
On/Off,Auto-scaling
+
May 2019 17
Rob Akershoek © 2019
Most IT organization realize they need to change their IT operating model…
June 26, 2019
The IT4IT Standard A reference model for managing and automating the business of IT
June 2019 19
Rob Akershoek © 2019
The IT4IT Standard as a blueprint for automating the IT value chain
URL: http://www.opengroup.org/IT4IT
December 2018 20
Automate the end-to-end IT value chain
December 2018 21
The Service Backbone as the DNA of the IT organization…
IT Service(portfolio)
ServiceRelease
(Blueprint)
ServiceDeployments
& ServiceModel
December 2018 22
The Service Backbone
IT Service
(portfolio
item)
Service
Release
(and
blueprint)
Service
Catalog
Entry
Service
deployment
(and related
resources)
Becomes…
Ordered
Published in… Released/Deployed!
Streamline and automate workflow across the IT value chain (flow / pull / manage work queues)
Transparency, Traceability & Continuous Feedback Loops
June 2019 23
Rob Akershoek © 2019
Service Backbone – How do we manage the end-to-end business chain and related IT services/products?
Technology /
infrastructure
blueprints e.g.
database, load
balancer, storage,
network (design state)
Business models,
business capabilities,
business processes,
Solution Architecture,
Application and
technology landscape
Services /
applications and
product, ownership.
CMDB model,
discovery, license
management.
Deployment /
configuration tools,
Element managers
Service model for
monitoring, event
management and
impact assessment.
Perform root cause
analysis.
Risk and compliance
management tools
Lacking an integrated view on the IT landscape…
June 2019 24
Rob Akershoek © 2019
Simplified view on the end-to-end IT Value Chain…
Requirement to Deploy Detect to CorrectRequest to FulfillStrategy to Portfolio
Business InterfaceIdea / Demand Requirement / Feature Service Request Incident / Question / Complaint
Portfolio Backlog
Enterprise Architecture
Service / ProductPortfolio
Idea / Demand
Service Backbone (service portfolio, service catalogue and CMDB)
Test
Defect
Product / Team Backlog
Code &Configure
CI/CD Pipeline
Event
Problem
Monitoring
Incident
Change Security Operations
Identity &Access
RequestFulfilment
Service catalogue
Deployment & Provisioning
Usage & Charge-
back
Risk & ComplianceIT Financial
ManagementData Analytics &
ReportingVendors & Contracts Knowledge
Management
June 2019 25
Rob Akershoek © 2019
Top IT automation initiatives most organizations are working on…
CI/CD
Pipeline
Test
Automation
Security
Testing
Deployment &
Release
Automation
Cloud Orchestration
(e.g. Infrastructure as
code)
Monitoring &
Log Analytics
Security
Monitoring &
Vulnerability
Management
Event
Management
Self service
portals
Chatbot /
Virtual
Assistant
Request
Fulfillment
Automation
Identity &
Access
Management
Discovery
Control Monitoring
& Compliance
Collaboration &
Communication
June 2019 26
Rob Akershoek © 2019
Artificial Intelligence Use Cases in the IT Value Chain
Monitoring & Event
Management e.g. anomaly
detection, predictive analytics,
root cause analysis, …)
Chatbot / Virtual
Assistant / Conversational AI (
for self help and self service)
Security
Operations e.g.
cyber defense
Incident routing,
diagnosis &
resolution
Capacity
optimization
Test Bots
e.g. code
quality,
security, etc.
Request
Fulfillment
(e.g. JML)
Diagnostics /
Automated
Remediation
Automatically
create new
demands e.g.
technology
refresh
Control & Compliance
Monitoring (e.g. risk
modeling)
Backlog item
prioritization
Change
Approval
(Go / No Go)
CMDB
Desired
State
Access
Control
June 2019 27
Rob Akershoek © 2019
Discovery Tools
Example of an AIOps Platform in the IT Tool Chain
Infrastructure, Platforms & Applications
IT Service Management System Communication & Collaboration
Monitoring & Log AnalyticsDeployment & Element Managers
Change & Request
CI/CD Pipeline
AIOps Platform (Service health, Impact assessment,
Root cause analysis,
event
management)
Monitoring data, logs & metrics
Data Collection / Events / Streaming data
Incident history / data updates
Create / Update Incident
Service Portfolio &
CMDB
Config. Data
Config. Data
Config. Data
Request
Deploy
Backlog Management
SLM
Incident & Problem
Swarming
Change / Releases
Change
Deploy / Configure
Self service portal
Knowledge base
June 2019 28
Rob Akershoek © 2019
Key questions you should ask yourself…
Do you have
insight in how IT
enables your
end-to-end
business chain?
Can you easily
pinpoint the root
cause and
determine
remediation
actions?
Can you
(proactively)
detect issues in
a timely
manner?
Are you able to consolidate relevant IT
monitoring and event data?
Do you find it almost
impossible to keep up with the
changing IT landscape (e.g.
CMDB)?
Can quickly
resolve incidents
in the digital
ecosystem?
June 2019 29
Rob Akershoek © 2019
Typical Objections to get started with AIOps…
My IT processes are
not mature (e.g.
Incident
Management,
Monitoring)
IT processes
My Service
Administration &
CMDB is not up-to-
date…
Service Administration
My monitoring
landscape is not well
defined..
Monitoring Landscape
I have outsourced
my operations
management
activities
Outsourced
June 2019 30
Rob Akershoek © 2019
Select Business ChainIdentify a business / application chain as a pathfinder and setup operations intelligence team.
Data IntegrationIdentity data sources and integrate with the AIOPS platform
Build a Service ModelDevelop, Test and Verify Model and data. Visualizee business chain. Impact assessment, root cause analysis
Continuously improve modelOptimize models and update CMDB. Improve data model.
Onboard new servicesIncrementally onboard additional business chains / services / teams
5. Scale,
Onboard
Other
services
3. Build &
Validate the
Model
2. Integrate
& Aggregate
Data
1. Select
business
chain
Transform IT Operations
Optimal User
Experience
Transparent &
Traceable Reduce costsProductive
DevOps Teams
Prevent issues Automation Smart Data
Bootstrap your AIOps Startup Journey
Demonstrate valueContinuously demonstrate value to the business/
4. Get Insight &
Proactively
Detect Issues
(Demonstrate
Value)
Generate InsightExecute the model; identify incidents, generate value, learning and improving. Automated incident creation.
Setup operations intelligence team
June 2019 31
Rob Akershoek © 2019
June 2019 32
Rob Akershoek © 2019
APPENDIX
June 2019 33
Rob Akershoek © 2019
AI Ops – Operations Intelligence / Smart IT
• AIOps uses Data analytics, and Machine learning in order to analyze big data from
various IT operation tools and devices. As a result, it is able to automatically spot and,
more importantly, react to issues more timely and proactively.
• AIOps requires a move away from siloed IT data. It uses a comprehensive analytics and
machine learning strategy against the combined data to continuously improve the outcome.
June 2019 34
Rob Akershoek © 2019
AIOPS Platform
Service Model
(Context)
IT data lake
(monitoring, streaming
data, events)
Impact Assessment, Root Cause Analysis, Predictive Analytics, …
Visualization & Alerting
Data Collection
(Logs, Data Streams / Metrics, Events)
IT S
erv
ice
Man
ag
em
en
t
Support
Incidents
SLA
Knowledge
Incidents &
Changes
Config.
Data
CMDB
Infrastructure & Applications
DiscoveryDeployment /
Provisioning toolsElement Managers
Monitoring tools e.g. network monitoring, platform,
log analytic tools, APM, etc. )
June 2019 35
Rob Akershoek © 2019
AI and ML Algorithyms