cloud migration with confidence: 7 keys to success
TRANSCRIPT
2
Copyright © 2017. All rights reserved.
Cloud Migration with Confidence: 7 Keys to Success
Explore how to quickly identify elements of your stack that may
require special attention or reconfiguration before, during, and
after migration
Discover how to objectively and accurately monitor your apps and
cloud infrastructure to ensure everything is running reliably
Prepare your applications and cloud infrastructure to run at scale
and learn how microservices and dynamic infrastructure can help
3
Copyright © 2017. All rights reserved.
Speakers
CBS InteractiveEditor
Branded Content
MICHAEL STEINHART DAVID GEWIRTZ ABNER GERMANOW
CBS InteractiveDistinguished Lecturer
and author ofThe Flexible Enterprise
New RelicSenior Director of Evangelism and
Partner Marketing
KALPAN RAVAL
Amazon Web ServicesGlobal Ecosystem
Leader Enterprise Migrations
4
Copyright © 2017. All rights reserved.
Pre-cloud Case Study: ZATZ
Launched in 1997
PHP was still at 1.0
Windows was Windows 95
Mac OS was System 7
PC: 8MB RAM, 500MB HD, $1700
Titanic was the big movie of the year
Blogs were still considered “zines”
Mark Zuckerberg was 13
5
Copyright © 2017. All rights reserved.
Pre-cloud Case Study: ZATZ
Launched in 1997
PHP was still at 1.0
Windows was Windows 95
Mac OS was System 7
PC: 8MB RAM, 500MB HD, $1700
Titanic was the big movie of the year
Blogs were still considered “zines”
Mark Zuckerberg was 13
If we ran out of server capacity, we had to buy a
new box
6
Copyright © 2017. All rights reserved.
Pre-cloud Case Study: ZATZ
Launched in 1997
PHP was still at 1.0
Windows was Windows 95
Mac OS was System 7
PC: 8MB RAM, 500MB HD, $1700
Titanic was the big movie of the year
Blogs were still considered “zines”
Mark Zuckerberg was 13
If we ran out of server capacity, we had to buy a
new box
If we ran out of bandwidth, we had to have an additional T-1
installed
7
Copyright © 2017. All rights reserved.
Pre-cloud Case Study: ZATZ
Launched in 1997
PHP was still at 1.0
Windows was Windows 95
Mac OS was System 7
PC: 8MB RAM, 500MB HD, $1700
Titanic was the big movie of the year
Blogs were still considered “zines”
Mark Zuckerberg was 13
If we ran out of server capacity, we had to buy a
new box
If we ran out of bandwidth, we had to have an additional T-1
installed
If we ran out of rack space, we
had to find a new location to store
our machines
8
Copyright © 2017. All rights reserved.
Pre-cloud Case Study: ZATZ
Launched in 1997
PHP was still at 1.0
Windows was Windows 95
Mac OS was System 7
PC: 8MB RAM, 500MB HD, $1700
Titanic was the big movie of the year
Blogs were still considered “zines”
Mark Zuckerberg was 13
If we ran out of server capacity, we had to buy a
new box
If we ran out of bandwidth, we had to have an additional T-1
installed
If we ran out of rack space, we
had to find a new location to store
our machines
Growth was in expensive chunks
Growth was heavily dependent on CAPEX
9
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
66,014 articles
1.2 million email newsletter subscribers
539 million email newsletters mailed
1.9 billion web pages served
By 2010…The cloud provided
ScalabilityFlexibility
ModularityCost Management
Migrated to cloud hosting services before AWS
Moved to AWS with managed hosting provider
10
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Migrated to cloud hosting services before AWS
Moved to AWS with managed hosting provider
66,014 articles
1.2 million email newsletter subscribers
539 million email newsletters mailed
1.9 billion web pages served
By 2010…
But…
11
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Pre-AWS migration was very challenging
It took three full years
12
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Pre-AWS migration was very challenging
Distributed systems were almost impossible to keep in sync
It took three full years
13
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Pre-AWS migration was very challenging
Distributed systems were almost impossible to keep in sync
Distributed workloads were nearly impossible to monitor and manage
It took three full years
14
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Pre-AWS migration was very challenging
Distributed systems were almost impossible to keep in sync
Distributed workloads were nearly impossible to monitor and manage
When we had some kind of race condition, tracking it down was brutally challenging
It took three full years
15
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Pre-AWS migration was very challenging
Distributed systems were almost impossible to keep in sync
Distributed workloads were nearly impossible to monitor and manage
When we had some kind of race condition, tracking it down was brutally challenging
Debugging and maintaining multi-server code was nearly impossible
It took three full years
16
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Pre-AWS migration was very challenging
Distributed systems were almost impossible to keep in sync
Distributed workloads were nearly impossible to monitor and manage
When we had some kind of race condition, tracking it down was brutally challenging
Debugging and maintaining multi-server code was nearly impossible
It took three full years
Delivering code updates without downtime was absolutely impossible
17
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Pre-AWS migration was very challenging
Distributed systems were almost impossible to keep in sync
Distributed workloads were nearly impossible to monitor and manage
When we had some kind of race condition, tracking it down was brutally challenging
Debugging and maintaining multi-server code was nearly impossible
It took three full years
Delivering code updates without downtime was absolutely impossible
Monitoring each workload process in real-time was impossible
18
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Pre-AWS migration was very challenging
Distributed systems were almost impossible to keep in sync
Distributed workloads were nearly impossible to monitor and manage
When we had some kind of race condition, tracking it down was brutally challenging
Debugging and maintaining multi-server code was nearly impossible
It took three full years
Delivering code updates without downtime was absolutely impossible
Monitoring each workload process in real-time was impossible
As we increased flexibility due to the cloud, complexity increased exponentially
19
Copyright © 2017. All rights reserved.
Case Study: ZATZ Cloud Migration
Pre-AWS migration was very challenging
Distributed systems were almost impossible to keep in sync
Distributed workloads were nearly impossible to monitor and manage
When we had some kind of race condition, tracking it down was brutally challenging
Debugging and maintaining multi-server code was nearly impossible
It took three full years
Delivering code updates without downtime was absolutely impossible
Monitoring each workload process in real-time was impossible
As we increased flexibility due to the cloud, complexity increased exponentially
Diagnosing where problems occurred (infrastructure or code) was very slow
21
Copyright © 2017. All rights reserved.
What We Needed
TOOLS TO ASSURE US MIGRATION WORKED
TOOLS TO CONFIRM WORKLOADS AND APPLICATIONS ARE RUNNING PROPERLY
22
Copyright © 2017. All rights reserved.
What We Needed
TOOLS TO ASSURE US MIGRATION WORKED
TOOLS TO CONFIRM WORKLOADS AND APPLICATIONS ARE RUNNING PROPERLY
TOOLS TO HELP DIAGNOSE ISSUES
23
Copyright © 2017. All rights reserved.
What We Needed
TOOLS TO ASSURE US MIGRATION WORKED
TOOLS TO CONFIRM WORKLOADS AND APPLICATIONS ARE RUNNING PROPERLY
TOOLS TO HELP DIAGNOSE ISSUES
TOOLS TO MONITOR IT ALL
24
Copyright © 2017. All rights reserved.
What We Needed
TOOLS TO ASSURE US MIGRATION WORKED
TOOLS TO CONFIRM WORKLOADS AND APPLICATIONS ARE RUNNING PROPERLY
TOOLS TO HELP DIAGNOSE ISSUES
TOOLS TO MONITOR IT ALL
ACROSS ON-PREM, CLOUD, & CLOUD WORKLOADS
©2008–17 New Relic, Inc. All rights
reserved.
Abner Germanow, New Relic Partner Marketing, November 2017 @abnerg
Get Cloud Ready & Migrate Fast With Confidence
NEWRNYSE – 2014
15k+Customers
13,500+ Disruptors1,500+ Global Enterprises
New Relic is the Catalyst
1.5BScale: Events & Metrics Per Minute
You need to INSTRUMENT EVERYTHING fastFast, easy, complete
29Confidential ©2008–17 New Relic, Inc. All rights reserved.
7 Programming Languages
Cloud & On-Host Integrations
30Confidential ©2008–17 New Relic, Inc. All rights reserved.
We collect at scale and innovate daily..Enterprise class SaaS
+1.5Bmetrics and events/minute
+50 teams deploying every day
+15K customers
..with world-class security and open APIsEnterprise class SaaS
31Confidential ©2008–17 New Relic, Inc. All rights reserved.
✓ SSAE-16/SOC II New Relic Platform✓ SSAE-16/SOC II Data Center✓ CSA-STAR
enterprise security
To make sense of raw data and figure out what matters..Curated experiences powered by Applied Intelligence
32Confidential ©2008–17 New Relic, Inc. All rights reserved.
So the whole team has the Insights they need to succeed. Self Service, Proactive, Predictive
33Confidential ©2008–17 New Relic, Inc. All rights reserved.
34Confidential ©2008–17 New Relic, Inc. All rights reserved.
The New Relic PlatformInstrumentation. Intelligence. Insight
Measuring the migration to the cloud is critical
Create baselines (black-box, frontend and
backend app metrics)
Compare and understand deviations from baselines
each step of migration
Keep your cloud budget in control and monitor the dynamic
cloud(s) post migration
Pre-migration Profiling
& Benchmarking
Performance Post-migration
& During Optimization
©2008–17 New Relic, Inc. All rights
reserved.
Getting Cloud Ready
Dev
Now - DYNAMIC World
Ops
Ops
Previous - STATIC World
©2008–17 New Relic, Inc. All rights reserved.
Private Data Center To Infrastructure AAS
Amazon EC2 Instance
BrowserMobile
Server (Virtual)
Hardware
Server OS
Application &
Application
Microservices
BrowserMobile
Server Hardware
Server OS
Application &
Application
Microservices
Amazon EC2 Instance
BrowserMobile
Server (Virtual)
Hardware
Server OS
Application &
Application
Microservices
AWS Monitoring (Once you get there)
Amazon CloudWatch
AWS CONSOLE
Amazon CloudWatch
Monitors
• EC2 instance
• Virtualization
• Hardware
• [CPU / Disk / Networking]
Doesn’t know about:
• Server OS
• Memory / Filesystem
• Processes
• Configuration
• Application
- Latency
- Error rates
Amazon EC2 Instance
Browser Mobile
Server (Virtual)
Hardware
Server OS
Application &
Application
Microservices
DASHBOARDS
New Relic Monitoring
New RelicApplicationMonitoring
New Relic Infrastructure
Monitoring
Amazon CloudWatch
AWS CONSOLE
Customer Experience:
• Synthetic tests
Application Analytics:
• App health
• App performance
• Microservice Dependencies
Instance Analytics:
• How O.S. is performing
• Configuration Changes
• Files & Packages
• Processes
Doesn’t know
• Virtualization
Amazon EC2 Instance
Browser Mobile
Server (Virtual)
Hardware
Server OS
Application &
Application
Microservices
DASHBOARDS
New Relic + AWS Monitoring
New RelicApplicationMonitoring
New Relic Infrastructure
Monitoring
Amazon CloudWatch
AWS CONSOLE
New Relic
Monitors
CloudWatch
monitors
Integrations
AWS / CloudWatch
Integrations
• Visibility into virtualization
• CPU / Disk / Networking
• Popular AWS Services
New Relic
• CPU / Disk / Networking
• Memory / Filesystem
• Processes
- Infrastructure components
- Configuration inventory
• Application / Microservices:
- Latency
- Error rates
- App insights
1b.Profile Application
Dependencies to prioritize
application moves
Instrumentation enables speed & confidence at each phase. Measurement Depth Will Vary Based On Needs
Plan Migrate Run
Strategy Analysis Design Transition Operations Improvement
1. Establish baselines for
end-user and application
performance and availability.
4. Optimize apps
for the cloud &
business goals.
New
Relic D
ata
3b. Acceptance Test: Prove
end-user, application & Infra
performance and health.©2008–17 New Relic, Inc. All rights
reserved.
3. Compare cloud
environment vs on-
prem environments.
2. Cloud Ready
Infrastructure
Monitoring
Discovery
• How do you identify and prioritize infrastructure, apps, and teams to migrate?
• How can you easily baseline customer experience and availability pre-migration?
• Do you understand infrastructure, app & service, and team dependencies?
Key Discovery & Cloud Readiness QuestionsProve the cloud has improved your application and prepare to show progress
Cloud Readiness
• Are your tools built for and priced for the cloud?
• How can you baseline customer experience and availability pre-migration?
• Can you see where your workloads are in real time?
For time sensitive lift and shift migrations or packaged apps, Synthetic monitoring is fast and easy.
Plan Migrate Run
Strategy Analysis Design Transition Operations Improvement
1. Establish baselines for
end-user and application
performance and availability.
New
Relic D
ata
©2008–17 New Relic, Inc. All rights
reserved.
3. Compare cloud
environment vs on-
prem environments.
Confidential ©2008–17 New Relic, Inc. All rights reserved
Cloud Ready Speed or Packaged App Planning Strategy
5
2
Recommendation:
1. Set up synthetic monitors & alerts for each URL and for significant customer paths
2. Install cloud ready infrastructure monitoring
3. Measure migration progress
Synthetics – Script Example 1 of 2
This scripting action is the basis for nearly all scripted browsers.
©2008–17 New Relic, Inc. All rights
reserved.
Synthetics – General
©2008–17 New Relic, Inc. All rights
reserved.
From
Where?
How Often?
Name &
Apdex
Synthetics – Overview
©2008–17 New Relic, Inc. All rights
reserved.
Load Time
Availability
Slowest
Results
Failures
1b.Profile Application
Dependencies to prioritize
application moves
Instrumentation enables speed & confidence at each phase. Measurement Depth Will Vary Based On Needs
Plan Migrate Run
Strategy Analysis Design Transition Operations Improvement
New
Relic D
ata
©2008–17 New Relic, Inc. All rights
reserved.
Confidential ©2008–17 New Relic, Inc. All rights reserved
APM Service Maps & Transaction Data Easily Surface Dependencies
Re-Architecture Discovery Strategy: Inventory On-Premises & App Dependencies
Recommendation:
1. Inventory on-premises infrastructure
2. Surface app and code dependencies with APM.
3. Map high potential infrastructure to apps to teams and prioritize sequence to optimize learning & success
Monitoring On-prem and cloud hosts with the same tool enables migration tracking and investment protection
Plan Migrate Run
Strategy Analysis Design Transition Operations Improvement
New
Relic D
ata
©2008–17 New Relic, Inc. All rights
reserved.
2. Cloud Ready
Infrastructure
Monitoring
Confidential ©2008–17 New Relic, Inc. All rights reserved
Infrastructure Monitoring
• Choose infrastructure and cloud agnostic monitoring
• Work with a vendor with flexible terms for migrations
• Track your progress at a glance.
Migration
• What types of acceptance criteria should determine if the application is considered stable and performing well in the cloud?
• How is the overall migration initiative tracking across applications, application health, and infrastructure?
Acceptance Testing
• Does the team have confidence in the health and stability of the migrated application?
• Has an application’s performance gotten faster or slower?
• Is an application more or less stable than before?
Key Migration Measurement QuestionsMake sure your migration projects are on time and on budget
When a workload reaches the cloud, prove it’s happy.
Plan Migrate Run
Strategy Analysis Design Transition Operations Improvement
New
Relic D
ata
3b. Acceptance Test: Prove
end-user, application & Infra
performance and health.©2008–17 New Relic, Inc. All rights
reserved.
Confidential ©2008–17 New Relic, Inc. All rights reserved
Acceptance Test Strategy: Understand Key Application Metrics Before And After
6
5
Recommendation:
1. Use APM to understand error rates, response times, and more
1. Use Service Maps to understand system performance in hybrid environments
APM – Acceptance Testing
©2008–17 New Relic, Inc. All rights
reserved.
Web Transaction
Time
Error Rate
Transactions
Apdex
Throughput
Activity
Host(s)
Confidential ©2008–17 New Relic, Inc. All rights reserved
Acceptance Testing Metrics
67
• Backend (APM, Infra, Syn)
Response Time (ms)
Error Rate (%)
Availability (%)
Throughput per host (rpm)
Host CPU (%)
• Front End (Browser, Mobile)
Page Load Time (sec)
JS Error Rate (%)
Crash Rate (%)
HTTP Error Rate (%)
Cloud Ready Infrastructure Monitoring With AWS Integrations
©2008–17 New Relic, Inc. All rights
reserved.
Confidential ©2008–17 New Relic, Inc. All rights reserved 72
● End State Dashboards/Maps
○ Ops Dashboards
■ Full Stack App Health
■ Regional/Local AZ Performance
■ Availability
○ Business Level
■ KPI Tracking
■ Customer Issues
○ Dev Dashboards
■ Errors
■ Version Performance
○ Health Maps
Run & Optimize Software-Driven Business
Unified view of Information across hybrid cloud
Elastic scalability to support
spikes in demand
Baseline metrics support
AWS migration success
“New Relic allows us to build a solid foundation for all
of our next-gen products that helps us deliver great
customer experiences.”
Results
- Stephen WeinbergDirector of Site Reliability Engineering,
Morningstar
New Relic at Morningstar
1b.Profile Application
Dependencies to prioritize
application moves
New Relic enables speed & confidence at each phase. Prove Success For Your Entire Cloud Journey
Plan Migrate Run
Strategy Analysis Design Transition Operations Improvement
1. Establish baselines for
end-user and application
performance and availability.
4. Optimize apps
for the cloud &
business goals.
New
Relic D
ata
3b. Acceptance Test: Prove
end-user, application & Infra
performance and health.©2008–17 New Relic, Inc. All rights
reserved.
3. Compare cloud
environment vs on-
prem environments.
2. Cloud Ready
Infrastructure
Monitoring
Confidential ©2008–17 New Relic, Inc. All rights reserved
Discovery & Cloud
Readiness“Lift and shift”
Cloud migration
Optimize and
improve app +
infrastructure
efficiency
Shift to DevOps,
containerization,
shared services
Agile development
and deployment,
integrated teams
Common Journey 1: AWS To Get To DevOps
Optimize front end
Custom Data & Dashboards For Each Phase
Application Dependences, Health
Cloud / On-Prem Infrastructure & Services
Benchmark Customer Experience & Availability
Move on-prem apps to cloud
prioritizing high infra cost appsReap benefits of cloud and
improve customer experience
Highest biz value, make teams better
at deploying and building
Migrate RunPlan
Confidential ©2008–17 New Relic, Inc. All rights reserved
Common Journey: AWS To Accelerate DevOps
Cloud migration
Optimize and
improve app +
infrastructure
efficiency
Accelerate DevOps
with containerization,
shared services
Accelerate Agile
development and
deployment
Move on-prem apps to reside
with cloud native appsMinor changes to take
advantage of cloud elasticity
Focus change efforts on new high value
apps and architecture of existing apps to
improve customer experience
Adopt DevOps,
but speed is
infrastructure
constrained
Migrate RunPlan
Confidential ©2008–17 New Relic, Inc. All rights reserved
Example Alert Strategy In The Cloud
78
● Base Alert Set (Built Automatically)○ Cloud Instance Health
■ Host Availability Alert
■ Process Availability Alert
■ Perf Metrics (CPU, Disk, Memory)
○ Application Health
■ Response Time
■ Error Rate (Baseline)
■ Throughput (Baseline)
○ Client-Side Health
■ Page Load Time
■ Crash Rate
■ Error Rate
■ Throughput (Baseline)
○ Cloud Products Health (RDS, S3, Cosmos DB, etc)
■ Perf Metric by Type
Disruption demands a response
BCG research indicates that leaders believe the biggest impacts on their industry will come from:
• Speed of adopting new technology
• Big-data analytics
• Technology platforms
• Enhanced operations processes
80
Migration Acceleration Program (MAP)
• Mature organisations are often built upon a complex digital tapestry that embodies years of technical debt
• Addressing this issue allows IT organisations to better serve the needs of the business, and enable digital transformation
81
Source: McKinsey Global Survey
But it is hard to drive real change when so much
budget just supports the status quo
Migration Acceleration Program (MAP)
Responding to disruption requires a new model
Focus IT on what differentiates your business.
Retire technical debt.
Innovate at start-up like speed
Reduce risk
©2016, Amazon Web Services, Inc. or its affiliates. All rights reserved 82
Migration Acceleration Program (MAP)
Migration Acceleration Program (MAP)
This Stages of Adoption path describes a
common journey customers take to drive
innovation and value through the cloud
• Project stage
• Foundation stage
• Migration stage
• Reinvention
The Migration Acceleration Program
helps accelerate your Foundation building
and Application Migrations to more rapidly
build capability and achieve business
results
AWS PartnersAWS and Partner Tools
AWS Investment AWS Training
Migration Methodology
AWS Professional Services
Migration Acceleration Program (MAP)
Customer experiences
Migrated over ~5,500 instances in 9 months, reduced storage costs 50%, compute costs 20%, and sped up provisioning by 10x (4 weeks to 2 days).
Migrated 300+ applications in one year for a savings of $14M.
Migrated data center assets for several divested business units in 6 months.
Consolidating 56 data centers down to 6, moving 75% of infrastructure to cloud in the process, and re-allocating more than $100M toward key business drivers.
Migration Acceleration Program (MAP)
©2016, Amazon Web Services, Inc. or its affiliates. All rights reserved 85
Migrated 8 of 12 data centers to AWS, improving performance, reliability, and cost.
App Migration Patterns
Redesign
Application/
Infrastructure Architecture
App Code
Development
Full ALM /
SDLC
Integration
Purchase COTS/
SaaS & licensing
Manual Install
& Setup
Modify underlying
Infrastructure
Determine
new platform
Retain /
Revisit
Manual Config
Manual Deploy
Manual Install
Retire /
Decommission
Determine
Migration
Path
Use Migration Tools
Transition ProductionValidation
Automate
Discover/Assess/Prioritize
applications
Retain/
Revisit
Retire/
Decommision
Determine
migration path
Migration Acceleration Program (MAP)
Prepared by TechRepublic exclusively for
THANK YOU!
CLOUD MIGRATION WITH CONFIDENCE:7 KEYS TO SUCCESS