Techniques for identifying , analyzing and treating risks to user service quality of software applications executing on cloud computing platforms.

1. © 2017 Nokia1
Risk Thinking for Cloud-
Based Application Services
Public
Eric Bauer

2. © 2017 Nokia2
The Cloud Operator’s Risk Management Challenge
…and those organizations seek to avoid unacceptable downside
consequences, especially
Impaired service reliability
Impaired service latency
Impaired service availability
Risk Management is “coordinated activities to direct and control an
organization with regard to risk” meaning activity to reduce the
uncertainties of achieving desired upside benefits without unacceptable
downside consequences
Organizations invest in cloud in pursuit of two benefits:
1. Deliver new services and value faster
 grow top line
2. Improve operational efficiency
 boost bottom line

3. © 2017 Nokia3
Cloud
Service
Providers
(CSPs)
Primary Cloud Operator Roles
CSP: Management
and Orchestration
as-a-Service
CSP: Functional
Component
offered as-a-Service
Cloud
Service
Customer
(CSC)
CSP:
Infrastructure
as-a-Service
Application
Software
CSP:
Network
Provider
Provider
Role
Customer
Role
Customer
Role
Provider
Role
Customer
Role
Provider
Role
Customer
Role
Provider
Role
Customer
Role
Provider
Role
Customer
Role
Provider
Role
Functional
Component
as-a-Service
Virtual
Compute,
Memory,
Storage &
Networking
Service
Automated
Lifecycle
Management
Cloud
Service
User
Application
Software
Integrator
Role
Transport
Service
Application
Service
Cloud Service Customers (CSCs)
operate application services hosted
on public or private clouds
Cloud Service Providers (CSPs) own and operate the physical
infrastructure equipment, etc, that serve the virtual resources,
automated lifecycle management and functional components to CSCs
Considering risks facing the cloud service customer organization
Roles based on ISO/IEC 17788:2014 Cloud Computing Overview & Vocabulary

4. © 2017 Nokia4
Deliver new services and value faster to grow
the top line, especially via:
1. Enhanced Service Value Chains…rapidly
leverage off-the-shelf service components
2. Agile/DevOps and Accelerated Application
Service Lifecycle
Improve operational efficiency to grow the
bottom line, especially via:
1. Aggressive automation and self service
2. DevOps and application lifecycle changes
3. Perfect capacity management
4. Transparency and continuous improvement
5. Aggressive cost management
Expected Cloud Service Customer Benefits of Cloud Deployment
Service Lifecycle
ServiceTransitionPlan & Code
Service
OperationTest
Build
Development Lifecycle
Cloud
Service
Customer
(CSC’s)
Service
Production
Costs
CSC General & Administrative Overheads
Maintenance & Support Fees
CSC Labor Staffing Loaded Salary
CSC Operations & Business Support Systems
Application &
Software
Components
Usage Usage-based pricing
Infrastructure Usage Usage-based pricing
Functional
Component
as-a-Service
Usage Usage-based pricing
Cloud
Management &
Orchestration
Usage Usage-based pricing
1.Aggressive automation,
including selfservice
5.Aggressive cost
management
3. Perfectcapacity
management
4. Transparencyand
ContinuousImprovement
2. DevOps and application
Lifecycle Changes

5. © 2017 Nokia5
Cloud User Service Quality Risk Fishbone Diagram
Virtual Network Risks
 Packet Loss
 Packet Delay
 Packet Jitter
 Network Delivered
Throughput
 Network Outage
 VN Diversity
Compliance Violation
VNF Product Risks
 Faulty VNF Configuration
Specifications
 Defective VNF Lifecycle
Management Scripts
 Residual Product Defect
Virtualized Application
Latency Risks
 TAIL Application Latency
 TYPICAL Application
Latency
Service Integration Risks
 Wrong Element Used
 Integration Defect
 Element Operational
Conditions Violated
 Faulty Service Delivery
Architecture
 Faulty Service Control
Architecture
 Faulty Service Workflow
Lifecycle Management (Execution)
Risks
 VNF Lifecycle Management
 Network Service Lifecycle Mgmt
 Forwarding Graph Service Lifecycle
Mgmt
 Virtual Link Service Lifecycle MgmtAccountability Risks
 Incomplete Accountability
 Conflicting Accountability
 Ambiguous Demarcation
 Ambiguous Service Level
Objective
 Inadequate/Ambiguous Data
Cloud Service Provider
Catastrophe Risks
 (Semi?) Permanent loss of cloud
service
 (Semi?) Permanent loss of CSC
data
Human and Organizational
Risks
 Organization and Incentive
Design
 (Human) Process Risks
 Human Error
Unknown-Unknown Risks
 Fundamental Disruption
 Emerging Technology
 Technical Debt
 Flawed Standards
 Faulty IaaS CapEx Reduction
 Faulty CSP OpEx Reduction
Virtual Machine
Risks
 VM Dead on Arrival
 VM Failure
 VM Stall
 VM Scheduling
Latency
 VM Clock Error
 VM Placement
Policy Violation
User
Service
Quality
Risk
Virtual Storage
Risks
 Storage access
latency
 Storage access
reliability
 Volume capacity
 Volume outage
 Volume
throughput
Service Policy Risks
 Insufficient Spare Capacity
(Target)
 Faulty Resource Placement
Policy
 Faulty Scaling Decision Criteria
 Inaccurate Demand Forecast
Functional-Component-as-a-
Service Quality Risks
 FCaaS Outage Downtime
 FCaaS Reliability
 FCaaS Latency
Visibility Risks
 Obstructed Vision
 Poor Resolution
 Stale Vision
 Mirage
Graphic from Risk Thinking for Cloud-Based Application Services, by Eric Bauer, CRC Press, April 2017

6. © 2017 Nokia6
Risk Treatment Techniques
Enterprises can treat risks via one or more of the following techniques
 Replace or remove the risk source, e.g., replace risky service component with a higher
quality alternative
 Change the risk likelihood, e.g., more testing to find (and then fix) residual defects
 Change the risk consequences, e.g., redundancy or high availability architectures to
minimize duration of impact from failures
 Share the risk with external parties, e.g., service level agreements with remedies, such
as liquidated damages
 Retain the risk (default option), e.g., business retains the risk that customers will
demand their product or service offering
 Reject accountability, e.g., for misuse or abuse of the product or service
 Avoid the risk, e.g., delaying or declining to bring a product or service to market

7. © 2017 Nokia7
Primary Risk Identification Techniques
Influence Diagrams
Cause and Effect Analysis
Failure Mode Effect Analysis
Structured Interviewing and Brainstorming
Structured what-if technique (SWIFT)
Fault Tree Analysis
Risk Identification and Analysis Techniques
Virtual Network Risks
 Packet Loss
 Packet Delay
 Packet Jitter
 Network Delivered
Throughput
 Network Outage
 VN Diversity
Compliance Violation
VNF Product Risks
 Faulty VNF Configuration
Specifications
 Defective VNF Lifecycle
Management Scripts
 Residual Product Defect
VirtualizedApplication
LatencyRisks
 TAIL Application Latency
 TYPICAL Application Latency
Service IntegrationRisks
 Wrong Element Used
 Integration Defect
 Element Operational
Conditions Violated
 Faulty Service Delivery
Architecture
 Faulty Service Control
Architecture
 Faulty Service Workflow
Lifecycle Management
(Execution) Risks
 VNF Lifecycle Management
 Network Service Lifecycle Mgmt
 Forwarding Graph Service Lifecycle
Mgmt
 Virtual Link Service Lifecycle Mgmt
AccountabilityRisks
 Incomplete Accountability
 Conflicting Accountability
 Ambiguous Demarcation
 Ambiguous Service Level
Objective
 Inadequate/Ambiguous Data
CloudService Provider
Catastrophe Risks
 (Semi?) Permanent loss of cloud
service
 (Semi?) Permanent loss of CSC
data
HumanandOrganizational
Risks
 Organization and Incentive
Design
 (Human) Process Risks
 Human Error
Unknown-Unknown Risks
 Fundamental Disruption
 Emerging Technology
 Technical Debt
 Flawed Standards
 Faulty IaaS CapEx Reduction
 Faulty CSP OpEx Reduction
Virtual Machine
Risks
 VM Dead on Arrival
 VM Failure
 VM Stall
 VM Scheduling
Latency
 VM Clock Error
 VM Placement
Policy Violation
User
Service
Reliability
Risk
Virtual Storage
Risks
 Storage access
latency
 Storage access
reliability
 Volume capacity
 Volume outage
 Volume
throughput
Service Policy Risks
 Insufficient Spare Capacity
(Target)
 Faulty Resource Placement Policy
 Faulty Scaling Decision Criteria
 Inaccurate Demand Forecast
Functional-Component-as-a-
Service QualityRisks
 FCaaS Outage Downtime
 FCaaS Reliability
 FCaaS Latency
VisibilityRisks
 Obstructed Vision
 Poor Resolution
 Stale Vision
 Mirage

8. © 2017 Nokia8
Primary Risk Control Analysis Techniques
Layers of Protection Analysis
Critical Control Point Analysis
Event Tree Analysis
Bow Tie Analysis
Risk Control Analysis Techniques

9. © 2017 Nokia9
0
1
2
3
4
5
Virtual Machine
Risks
Virtual Network
Risks
Virtual Storage Risks
Virtualized
Application Latency
Risks
Functional
Component-as-a-
Service Quality Risks
Lifecycle
Management
(Execution) Risks
Volatile Risk Vector Index
Risk Evaluation Techniques
Primary Risk Evaluation Techniques
Failure Mode Effects and Criticality Analysis
Dose-Response (Toxicity) Assessment
Consequence/Probability matrix
FN Curves
Risk indices
Decision tree
Cost/benefit analysis

10. © 2017 Nokia10
Reconsidering Quality and Risk
Frequency
Magnitude
Opportunity
(very low risk)
Acceptable
Tolerable
Unacceptable
Canonical Risk Map
Opportunity
Acceptable
Unacceptable
Tolerable
Level of Cloud Service Impairment
User Service
Quality
(Defective
Operations
per Million)
Target
Opportunity
to soften cloud
service level
objective (SLO)
10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2
10 DPM
1 DPM
0.1 DPM
0.01 DPM
Unacceptable
cloud service
level objective
Tolerable
Hypothetical Dose-Response Chart
Super acceptable service quality may be an
opportunity reduce resource allocation to
increase operational efficiency and save money

11. © 2017 Nokia11
Quality and Risk-Based Thinking
• ISO 9001:2015 “Quality Management Systems -
Requirements” Clause 0.3.3 – “To conform to the
requirements of this International Standard, an
organization needs to plan and implement actions to
address risks and opportunities. Addressing both risks
and opportunities establishes a basis for increasing the
effectiveness of the quality management system, achieving
improved results and preventing negative effects.”
Available April 2017
Methodically identifying, analyzing,
treating and monitoring uncertainties for
desired benefits and undesirable
consequences is a best practice for
service, risk and quality management