Amidst an industry cloud of confusion about what “AIOps” is and what it can do, these slides--based on the webinar from EMA research--delineates a clear path to victory for business and IT stakeholders seeking to use machine learning to optimize the performance of critical business services.
Use of FIDO in the Payments and Identity Landscape: FIDO Paris Seminar.pptx
AIOps, IT Analytics, and Business Performance: What’s Needed and What Works
1. IT & DATA MANAGEMENT RESEARCH, INDUSTRY
ANALYSIS & CONSULTING
Dennis Drogseth
VP of Research
EMA
AIOps, IT Analytics, and Business
Performance: What’s Needed and What Works
Bernd Harzog
CEO
APM Experts
Marty Pejko
COO
Centerity
2. IT & DATA MANAGEMENT RESEARCH,
INDUSTRY ANALYSIS & CONSULTING
Watch the On-Demand Webinar
Slide 2
• AIOps, IT Analytics, and Business Performance: What’s
Needed and What Works On-Demand webinar is available here:
http://info.enterprisemanagement.com/aiops-itanalytics-
businessperformance-webinar-centerity
• Check out upcoming webinars from EMA here:
http://www.enterprisemanagement.com/freeResearch
3. Featured Speakers
Dennis Nils Drogseth, Vice President, EMA
Dennis joined EMA in 1998 and currently manages the New Hampshire office.
Dennis brings several years of experience in various aspects of marketing and
business planning for service management solutions. He supports EMA through
leadership in IT Service Management (ITSM), CMDB systems, as well as
megatrends like advanced operations analytics, cross-domain automation systems,
IT-to-business alignment, and service-centric financial optimization.
Bernd Harzog, CEO, APM Experts
Bernd assists global enterprises with management and monitoring software purchase
decisions and implementation strategies. Clients have included Credit Suisse,
Deutsche Bank, Chevron, UBS, Nordea Bank, and Allianz Insurance. APM Experts
also provides product strategy and marketing strategy services to monitoring and
performance management vendors.
Marty Pejko, COO, Centerity
Marty has over 25 years of technology experience in enterprise software, security,
networking, and communications. Marty formerly worked as VP of Global Channels
for Guardium (acquired by IBM), as VP of Global Channel Sales for Network
Intelligence (acquired by EMC/RSA), as VP of Business Development for
International Sales for Quantum Bridge Communications (acquired by Motorola) and
as Corporate Counsel for GeoTel Communications (acquired by Cisco).
6. 1980-2005
Bernd’s Background
• Invented Microsoft SNA
Server for Windows NT
• Gartner Research
Director for Intel
System Software
• 2 Patents for
self-learning
performance analytics
at Netuitive
• CEO at RTO Software
(Citrix Performance
Management) – sold
company to Citrix
• CEO of OpsDataStore
– An AIOps Platform
• CEO of APM Experts –
Consultant to the
leaders in the
monitoring industry
PRODUCT SELECTION CONSULTANT
APM EXPERTS (2005 – 2019)
PRODUCT STRATEGY CONSULTANT
7. The Business Imperative - Digitization
Enterprises are digitizing
(implementing in
software) key constituent
facing services
This requires thinking and
operating like a software
vendor, not an enterprise
IT organization
Software vendors treat
revenue generating
“services” as products,
not projects
As a result of digitization,
the demand to implement
business functionality in
software is infinite
8. As a Result We Continue to See More Innovation, Diversity,
Complexity, Change and Dynamic Behaviors at All Layers of the Stack
Accelerate delivery of
code into production
Everything virtualized
Many Clouds
Shifts in application
architecture
Proliferation of
data architectures
In many different
services and
containers
In many different
languages
9. An Unprecedented Situation
• Digitization is an unprecedented business imperative for enterprises to compete
and execute online as software vendors (Time to Market, Agility, Quality of
Service, End User Experience)
• An unprecedented pace of innovation in processes and technology to support the
business imperative of digitization
• The need for continuous availability and performance is driving dynamic behavior
in virtualized and cloud based compute, networking and storage services
• Dynamic behavior in virtualization and cloud platforms, in Kubernetes and
applications and containers complicates understanding the truth
• Time to market pressures are leading to unprecedented levels of diversity in the
software stack with continuous changes on a release by release basis
• Time to market and agility pressures are causing applications to be architected
around microservices and released multiple times a day with CI/CD processes
• Monitoring is now a very hard problem
10. The Resulting Requirements
• The entire stack must now be monitored in real time (1 Min – 1 Sec) to be
able to detect service quality issues in time
• AI (AIOps) must be deployed to cope with the deluge of incoming monitoring
data and automatically understand normal vs. abnormal
• Relationships across the stack must be determined in real time
• What talks to what (traces and flows)
• What runs on what
• What is a member of what
• AIOps and relationships must be leveraged for automated root cause
• The results of monitoring must be made relevant to business constituents
11. Private Cloud (VMware) Relationships Model
Application ContainerTransaction
Virtual
Server
Datastore
Service
Business
View
Micro-
service
Physical
Server
Process
LUN
Raid
Group
Disk
Group
Disk
K8
Node
K8
Pod
Cluster
Datacenter
Virtual
Network
12. Public Cloud (AWS) Relationships Model
Application ContainerTransaction
EC2
Instance
EBS
Volume
Service
Business
View
Micro –
service
Elastic
NIC
Process
Subnet VPC Region
S3
Bucket
S3
K8
Node
K8
Pod
13. • Gartner expects AI to pervade every aspect of managing IT including APM, IT Infrastructure Monitoring, Cloud
Platform Monitoring, Performance Management, Capacity Management, the delivery tool chain (CI/CD), and IT
Service Management
• Gartner also expects the emergence of AIOps platforms which combine data from multiple sources to deliver
enhanced value
AIOps Overview
Gartner's AIOps
Platform
Architecture
AIOps Platform
Data Types
14. Approaches to AIOps
There are many different approaches to AI
and ML
Rule based – works only for constrained
and limited use cases and difficult to
maintain
Neural Net based – requires training –
problematic in dynamic environments
Unsupervised Self Learning – difficult to
focus upon desired KPI’s
Supervised Self Learning – Combines
human expertise with Machine Learning
Automation is only possible when based
upon a deterministic foundation
15. AIOps Use Cases
Automated Baselining, Anomaly Detection and Root Cause
Automated Workload Management (Contention Avoidance)
Eliminate CPU, Memory, Network I/O and Disk I/O contention
Correctly size VM’s and Cloud Images
Place VM’s in the best Hosts and Clusters
Automated Cloud Cost Management
Optimize cost by right-sizing cloud images
Optimize cost by choosing the optimal price plan
Automated Event Management
De-duplicate events
Support a collaborative (DevOps) problem resolution process
AIOPs Platforms
Automated Performance Optimization and Remediation
Automatically learn the performance characteristics of the
application and the entire supporting stack
Automatically optimize for a chosen KPI (performance, efficiency)
16. IT & DATA MANAGEMENT RESEARCH, INDUSTRY
ANALYSIS & CONSULTING
Dennis Drogseth
Vice President
Enterprise Management Associates
AIOps, IT Analytics and Business
Performance: What’s Needed and What Works
22. IT Is More Easily Transformed than the
Business
BASE: (IT Transformation: N=131; Digital Transformation: N=306)
Very or Extremely Successful
79% of IT Transformation
69% of Digital Transformation
1%
5%
15%
45%
34%
0%
4%
26%
54%
15%
0 0.2 0.4 0.6
Largely…
Only…
Successful…
Very…
Extremely…
How successful have you been to date with your
digital or IT transformation initiative(s)?
Digital Transformation IT Transformation
47. All information within this document is confidential and commercially sensitive to
Centerity and must not be copied or disclosed to any third party without the prior
written consent of Centerity.
DYNAMIC SERVICE VIEWS INTO YOUR
CRITICAL BUSINESS SERVICES
Marty Pejko, COO
49. The Problem The Solution
Network
Monitor
Storage
Monitor
Transaction
Monitor
Server
Monitor
Virtualization
Monitor
OS
Monitor
IT
Operations
Log
Monitor
Container
Monitor
Franken-monitors fail
to provide any business
visibility
NetworkComputeStorageOSApplications
Consolidated
Dynamic Service
Views for each
critical Digital
Business Service
50. Data Collection
• Agentless
• Agent-Based
• Any API
• Comprehensive
• Real-Time
Key Metrics
• Availability
• Performance
• Throughput
• Error Rate
• Business State
Real-Time Relationship Engine
• Transaction Flow Mapping
• Infrastructure Dependency Mapping
• Virtualization & Cloud Grouping
• Automatic Discovery
Service Level Engine
• Leverages all metrics, logs & events
• Calculates Business Service Levels
Analytics Engine
• Dynamic Baselines
• Automatic Anomaly Detection
• Dependency Based Root Cause
Analysis
Dynamic Service
Views
• All Services
• Drill Down
• Root Cause
• Cross-Stack
Alerts and
Notifications
• Email
• SMS
• PagerDuty
• ServiceNow
• Slack
Business
Executive
Product
Manager
IT Operations
The Centerity Platform
PLATFORM CAPABILITIES
Real-Time Streaming • Role-Based Access Control • Multi-Tenancy • Scaling • High Availability
Custom
Queries
Events
Metrics
Logs
!
DEPLOYMENT OPTIONS
Bare Metal • Private Cloud • Hybrid Cloud • Public Cloud • Multi-Cloud
Integrations
• APM –
AppDynamics,
Dynatrace,
Riverbed, Nastel
• Virtualization –
VMware
• Cloud – AWS
• Middleware – Java,
.NET, SQL,
NOSQL, Docker,
Kubernetes
• Operating Systems
– Windows, Linux,
Solaris, HPUX, AIX
• Networking – All
TCP/IP, SNMP,
Netflow
• Storage – EMC,
NetApp, HP
Apps &
Business
Services
• SAP
• Medical
• Retail
Store
• Custom
Web
• Custom
Mobile
• IOT
• Digital
• Legacy