2. Your Application Portfolio is a Mix of Things
CONTAINERS
EVENT-DRIVEN
FUNCTIONS
DATA
MICROSERVICES
Batch
MONOLITHIC
APPLICATIONS
ESBJEE
COTS
MAINFRAME
.NET
3. To Start the Journey
Getting apps to run on PCF is easy. Policy, regulation, manual process, organization and politics is hard. Tackling
these things can take time so let’s start right away. Let’s really dig into the path to prod and define the value stream.
Identify the manual process and bottlenecks. Figure out how we measure near-term success. Do it with working code.
identify a
small set
of apps
define a
Funnel to
make
decisions
Baseline
the Path
to prod
Fire a tracer bullet to validate assumptions and push on path to prod constraints
4. Pick Applications That Matter
TECHNICAL BUSINESS Why Move?
● Licensing costs
● Time-to-market factors
● Revenue opportunities
● Business criticality
● Risk tolerance
● Change frequency
What Can we Move?
● Framework / runtime
● Architecture design
● Statefulness
● Use of proprietary tooling
● Dependencies / integration
● Usage / workload
Org Factors (the “When” and “How”)
● Domain expert availability
● Lifecycle stage
● Calendar dependencies
● Etc.
The App Tx Funnel
A framework for disposition
planning, prioritization and
governance
5. Understand the Path to Production
Sample: Banking Customer
65 Steps consisting of 45 days of wait and
180 days of process time for a total cost, per
new application, of $150,000.
● Workshop to define the basic process baseline
● Break down the larger steps into detail
○ Steps
○ Who is involved
○ Time required
○ Dependencies
○ Average cost
● Flag areas of potential automation
● Identify the full duration of time and cost
● Take “1” app and run through the process
○ Validate / invalidate assumptions
○ Try to automate in low-risk way
● Use as roadmap for subsequent steps
6. Application Transformation Strategies
Re-Host
Containerize aka
“lift and shift”
Re-Platform
Minimal efforts to run
natively on PAS
Re-Factor
Convert to Cloud
Native aka
“Modernize”
Re-Build
Decompose and
rewrite from the
ground up
Application Portfolio
Funneling Process
7. Your Code As a
Crime Scene:
App Foundry
Automated suitability assessment
and decisioning framework.
A free tool from Pivotal to help
accelerate good decision making
along the path to PCF.
Upload binary or map
to source control Parsers for most common
languages (Java, .NET,
PHP, Ruby, etc.)
Extensible rules (100s
available currently)
UI dashboard to help
triage decisioning
Understand technical
effort to move your app
8. Build Program Artifacts
We accelerate team velocity using a
“cookbook” filled with recipes written by
developers as new patterns for
transformation and legacy interoperation
are discovered through hands-on work.
Recipes are written in Git
markdown. The Cookbook is a
PCF site, includes search and is
assembled using a CI pipeline
9. How Successful Customers Scale Quickly
Small product team that services
internal customers while
automating everything.
PLATFORM AS PRODUCT
Hands-on training (and train-
the-trainer) for Application
Developers and Architects
CxO dashboard of aggregated
success metrics. OKRs at project
level keeps the work & aligned.
Apps on
PCF
BU or geo based “move hubs”.
Velocity through pairing, rotation
and reuse. Incremental results.
APPLICATION
TRANSFORMATION
+
EDUCATION AND
ENABLEMENT
“5S”
METRICS
Starter
Define a funnel
and path-to-prod
process. Push
initial app(s). Learn
a lot.
Prod Dial Tone
Assemble a
Platform team.
Deploy PCF with a
prod dial tone.
Day 2
Platform team
automates more,
staffs up, settles
into day 2
concerns.
A Few Projects
Targeting
progressively more
complex workload.
Usually within a
single geo and/or BU.
Enablement
Create program;
train people staffed
on App Tx projects
Move Hubs
Transition from
project to program
with multiple “hubs”
on a geo / BU basis.
BIG Wins
Measurement starts on
day 1 but is magnified
as workload increases
(as driven by the App
Transformation
Program!)
10. But Why? Why Do This AppTx Thing?
Note 1: 2017 & 2018 Pivotal customer case studies. n=15. “Process Time” = hands-on work (vs. “total time” being the time to deliver).
Note 2: Averages are exclusive of highest and lowest measures.
37% more time coding
37% relative improvement in time
spent coding (from 56% to 77%)
21% absolute improvement in time
spent coding (from 56% to 77%)
90% faster 61% faster
90% time reduction
87% process time reduction
Total time: from 65 to 6 days
Process time: from 81 to 11 hours
61% time reduction
82% process time reduction
Total time: from 14 to 5 days
Process time: from 115 to 21 hours
69% time reduction
74% process time reduction
Total time: from 79 to 24 hours
Process time: from 34 to 9 hours
69% more efficient
DevelopingProvisioning Operating (Day 2)Releasing
Software Development Lifecycle On-going Ops
82% average reduction in ‘path to production’ time (from 84 days to 15 days)
11. Recent Customer Overview
Planning & Strategy Re-Host & Re-Platform Program DeliveryRe-Factor & Re-Build
Citibank
TD Ameritrade
Aon Hewitt
USAF
Boeing
FedEx
Liberty Mutual
Wells Fargo
Discover Financial
General Motors
Stubhub
Raytheon
CoreLogic
Voya Financial
USAF
Dicks Sporting Goods
Meeting Customers Where They Are Across 100s of Projects