Configuration Management Best Practices

9/9/13

Conﬁguration
Management
Best
Practices

Bob
Aiello,
Principal
Consultant
and
Author
of

Conﬁguration
Management
Best
Practices
:

Practical
Methods
that
Work
in
the
Real
World

http://www.linkedin.com/in/BobAiello

http://cmbestpractices.com

1

CM Best Practices Consulting © 2013

Who am I?
•  CM Lead & Consultant for over 25 years
•  Editor-in-Chief at CM Crossroads
•  Author of CM Best Practices
•  IEEE Management Board
•  Tools and process agnostic
•  The guy called in the middle of the night
when the release doesn’t work!
©
2013

2

April 9, 2013

1

9/9/13

Goals of this Course
•  Implement Effective Source Code
Management practices including variants
•  Automate build, package and deploy
•  Establish effective IT Controls
•  Use industry standards and frameworks
•  Create a CM function that grows &
improves
©
2013

3

April 9, 2013

More Goals of this Course
•  Use CM to support development
•  Understand the classic four CM functions
•  Introduce the core CM framework
•  Examine current and emerging trends
•  Guidance on implementing Agile CM
•  Establish IT governance and compliance
•  Establish your own plan for CM!
©
2013

4

April 9, 2013

2

9/9/13

Goals
of
Code
Management

•  Never lose code
•  Know exactly what is running in Prod
•  Make a two line fix without any
chance of the code regressing (due to
the wrong version)
What exactly is CM?
©
2013

5

April
9,
2013

Conﬁguration
Management

•  Configuration Identification
•  Status Accounting
•  Change Control
•  Configuration Audit
Tracking and Controlling Changes to
Configuration Items
©
2013

6

April
9,
2013

3

9/9/13

Conﬁguration
Identiﬁcation

•  Provides a specific and unique
identity to each configuration item (e.g.
binary, config file, documentation)
•  Selecting the configuration items for a
system and recording their functional
and physical characteristics (Sevocab)
©
2013

7

April
9,
2013

Status
Accounting

•  Tracking the status of a configuration
item throughout its lifecycle.
•  Recording and reporting of
information needed to manage a
configuration effectively (Sevocab)
©
2013

8

April
9,
2013

4

9/9/13

Change
Control

•  Establishing checkpoints including
gatekeeping (e.g. Production, QA, UAT)
and configuration control.
•  Identifying, documenting, approving or
rejecting, and controlling changes to
the project baselines (Sevocab)
©
2013

9

April
9,
2013

Conﬁguration
Audit

•  Inspect and identify the exact version
of any configuration item (physical &
functional)
•  Independent examination of the
configuration status to compare with
the physical configuration (Sevocab)
©
2013

10

April
9,
2013

5

9/9/13

Ensuring
the
Trusted
Base

•  Know what you built
•  Deploy the right code
•  Verify that it got there
•  Detect any unauthorized changes
CM is a full lifecycle endeavor
©
2013

11

April
9,
2013

CM
is
a
full
lifecycle
eﬀort

•  The four functions should be part of a
development lifecycle (e.g. ISO/IEEE
12207, 15288)
•  There needs to be an implicit
requirement for testing CM itself
Leads us to V & V
©
2013

12

April
9,
2013

6

9/9/13

Veriﬁcation
and
Validation

•  Does the CI meet specified
requirements?
•  Have the requirements for a specific
intended use or application been
fulfilled?
©
2013

13

April
9,
2013

Functional
description
of
CM

•  Easier to understand in the context of
a lifecycle
•  Consisting of six core CM functions
•  Closely matches the job descriptions
of the people doing the work
•  Can be tailored to your needs
So what are the six functions?
©
2013

14

April
9,
2013

7

9/9/13

CM
Functions

•  Source Code Management
•  Build Engineering
•  Environment Configuration
•  Release Engineering
•  Deployment
Let's start with a brief overview
©
2013

15

April
9,
2013

Source
Code
Management

•  Control of every configuration item
(e.g. source code, config, binaries,
compile and runtime dependencies).
•  Much more than just checkin and
checkout (version control)
•  Provides sanity to the development
process (reduces cognitive complexity)
©
2013

16

April
9,
2013

8

9/9/13

Terminology

•  Configuration items (CIs) include
binaries, source code, config files and
even documents
•  ISO 1007 notes end user function
•  Bob says, “anything where getting the
wrong version would be bad”
©
2013

17

April
9,
2013

What
is
Control?

•  In CM, control is managing the
evolution of a CI throughout its lifecycle
•  Configuration Control
Is control really the right word?
©
2013

18

April
9,
2013

9

9/9/13

Principles

•  Code is locked down and can never
be lost
•  Code is baselined marking specific
milestones
•  Managing variants using branches
•  Code changed on a branch can be
merged
©
2013

19

April
9,
2013

More
Principles

•  Processes are repeatable Agile and
Lean
•  Traceability and tracking of all
changes
•  Improves productivity and quality

©
2013

20

April
9,
2013

10

9/9/13

Best
Practices

•  How do we establish source code
management that adheres to these
principles?
•  Better question is how does CM add
value and help facilitate the
development effort?

©
2013

21

April
9,
2013

Sandboxes

•  Provide a degree of isolation
•  Support multiple sandboxes
•  Allows the “what-if” scenario
•  Cheap and disposible
•  Make sure that you refresh before
commiting your code

©
2013

22

April
9,
2013

11

9/9/13

Variants
in
the
code

•  Supporting multiple operating systems

©
2013

23

April
9,
2013

CopyBranches

•  Example of a copybranch (versus
delta)

©
2013

24

April
9,
2013

12

9/9/13

Handling
a
bugﬁx

•  We need to change Revision 2, but 3
is already being developed

©
2013

25

April
9,
2013

Inner
Merge

•  You need to merge the change on the
bugfix branch back to the main trunk

©
2013

26

April
9,
2013

13

9/9/13

Outer
merge

•  You also might want some new code
merged from trunk to the bugfix

©
2013

27

April
9,
2013

Software
Patterns

•  Fixing bugs while developing next
version of a product in parallel
•  Support for developers working in
parallel
•  Track component baselines
Software Configuration Management
Patterns By Steve Berczuk
©
2013

28

April
9,
2013

14

9/9/13

Streams

•  Provides a clear usage paradigm
•  Model components and architecture
•  Control flow of changesets
•  Snapshots create baseline of code
•  Ability to load a particular snapshot
•  Strong security authorization and
entitlements
•  Complete history and traceability
©
2013

29

April
9,
2013

Examples

•  Organize code into components
•  Use Streams & branches
•  Make merging viable and traceable
•  Navigate your repository metadata
•  Use Tasks to track your work

©
2013

30

April
9,
2013

15

9/9/13

Defect
&
Task
Tracking

•  Track changesets to workitem
•  Traceability to who made the change
•  Makes release notes a breeze to
create
•  Ties back to requirements and test
cases
•  Allows for ALM and workflow
automation
©
2013

31

April
9,
2013

Globally
Distributed
team

•  Managing work for a globally
distributed team
•  Effective communication
•  Better coordination
•  Traceability
•  Visibility

©
2013

32

April
9,
2013

16

9/9/13

Deﬁning
the
Usage
Model

•  You need to create a clear and
compelling usage model
•  Otherwise everyone will do whatever
worked well on the last project
•  Helps even when you have to live with
an inferior tool

©
2013

33

April
9,
2013

Training

•  Training is the “hill to die on”
•  Best when given by your CM support
team
•  Includes the process you want them
to use
•  Much more than just vendor training
•  Test first and then teach
©
2013

34

April
9,
2013

17

9/9/13

Future

•  More robust Application Lifecycle
Management solutions
•  Integration with the entire ALM
•  Open standards (OSLC)
•  Toolchains for everyone!

©
2013

35

April
9,
2013

Source
Code
Management

•  Makes everything else easier to
manage
•  Helps to juggle multiple code lines
•  Improves productivity & quality
•  Leads us to build engineering!

©
2013

36

April
9,
2013

18

9/9/13

Build
Engineering

•  Reliable and repeatable automated
process to compile, link and package
code components.
•  Must handle complex compile
dependencies
•  Continuous integration (or nightly
build)
•  Visibility into who broke the build
©
2013

37

April
9,
2013

Principles

•  Builds are understood and repeatable
•  Builds are fast and reliable
•  Every configuration item is identifiable
•  Source and compile dependencies
can be easily determined

©
2013

38

April
9,
2013

19

9/9/13

More
Principles

•  Code should be built once, but
deployed anywhere
•  Build anomolies are identified and
managed
•  The cause of broken builds is quickly
and easily identified and fixed!

©
2013

39

April
9,
2013

CI
Identity
Crisis

•  Who am I?
•  What if you cannot reach the version
control system (VCS)?
•  CIs should be identifiable outside of
the VCS
•  Breadcrumbs are not enough
•  Its tagged so I can build it - right?
(not so fast – maybe you can't!)
©
2013

40

April
9,
2013

20

9/9/13

Version
IDs

•  You need to embed an immutable and
unique version ID
•  You must have a procedure to easily
pull out the version ID at runtime
•  Cannot depend upon the version
control system (VCS)
•  Stamp in the tag or label
©
2013

41

April
9,
2013

How
does
this
help?

•  Create a sandbox from the tag or
label which identifies a baseline
•  Create a variant in the code or bugfix
branch
•  Allows you to create a 2 line fix
without any chance of the code
regressing due to a version control
issue
©
2013

42

April
9,
2013

21

9/9/13

Compile
Dependencies

•  Get environment variables understood
•  You must be able to build at the
command line
•  Developers forget what they set in the
IDE
•  What was that classpath?
•  What libraries did I use?
©
2013

43

April
9,
2013

Independent
Build

•  Code must be built using a different
computer account on seperate
computer
•  First time it always fails !
•  You have to find whatever they forgot
to check into version control
•  Verification and validation of the build
•  Satisfy regulatory requirements (audit)
©
2013

44

April
9,
2013

22

9/9/13

Overengineering
the
Build

•  Beware of overly complex builds
•  Don't embed calls to the version
control tool
•  Use components that can be run
separately
•  Automate everything
•  Treat this like any other development
effort
©
2013

45

April
9,
2013

Continuous
Integration

•  Framework for structuring the entire
build, package and deploy
•  Determine build and integration issue
early in the process
•  Should include deployment to a test
environment

©
2013

46

April
9,
2013

23

9/9/13

Code
Analysis

•  Source code repository helps code
analysis in two different ways
•  Static Code Analysis by providing a
repository
•  Instrument the code using variants
•  Security Audits
•  Uncover code defects
•  Code coverage
©
2013

47

April
9,
2013

Build
Frameworks

•  Build agents
•  Preflight builds
•  Allows use of the build farm
•  Moves the build framework upstream
•  Supports rapid iterative development

©
2013

48

April
9,
2013

24

9/9/13

Ergonomics
of
the
Build

•  “Bob-proof” your build
•  Implicit verification and validation
•  Avoid the possibility of mistakes
•  Each step should be easy to
understand
•  One step should not break the stream
•  Use dashboards and reports to
communicate status
©
2013

49

April
9,
2013

Partnering
with
Development

•  Development will always be a step
ahead
•  Set entry criteria and require that you
get advanced notice of when they
change the architecture
•  Development should be able to build
the release on the command line
©
2013

50

April
9,
2013

25

9/9/13

DevOps

•  Set of Principles where development
and operations partners
•  Better communication
•  Knowledge sharing
•  Moving build and deploy upstream

©
2013

51

April
9,
2013

Future

•  Focus on complete deployment
framework
•  Support rapid iterative development
•  Virtualization and cloud computing
leads us to very fast build, package and
deploy
•  Don't forget the automated testing
©
2013

52

April
9,
2013

26

9/9/13

Environment
Conﬁguration

•  Managing the environments including
creation and controlled configuration
•  Procedures to manage compile and
runtime dependencies (e.g. database
access)
•  Monitoring runtime environment for
unauthorized changes (e.g. ports open)
©
2013

53

April
9,
2013

Principles

•  Environment configuration
dependencies are identified and
understood
•  Environments can be interogated for
current status
•  Code is built once and configured
using automated procedures
©
2013

54

April
9,
2013

27

9/9/13

More
Principles

•  Environment configurations should be
changed in a controlled and predictable
way
•  Environment configurations should be
documented and understood by all

©
2013

55

April
9,
2013

Supporting
Code
Promotion

•  Promotion of code throughout the
application lifecycle
•  Environments must be isolated from
each other
•  QA should never “accidently” access
production

©
2013

56

April
9,
2013

28

9/9/13

Using
Tokens

•  Your code should read $DB1
•  Substitute the correct database
•  Centralize environment variable
assignment

©
2013

57

April
9,
2013

Conﬁguration
Control

•  Should identify and control
environment configuration
•  Trust but verify
•  Good example of where a CMDB can
help keep surveillance during runtime
•  Depends upon Change Control

©
2013

58

April
9,
2013

29

9/9/13

Future

•  Cloud computing and virtualization
are having a huge impact
•  Full size environments created on the
fly
•  Sharing a pool of resources
•  Can I rent a super computer for a few
days please?
©
2013

59

April
9,
2013

Change
Control

•  Management of changes including
gatekeeping and configuration control
•  Process related changes managed
through the SEPG
•  Change Advisory Board (CAB) to
evaluate the downstream impact of a
change
•  A priori change control
©
2013

60

April
9,
2013

30

9/9/13

Principles

•  Changes should be planned and not
just last minute events
•  Changes should be understandable,
including their downstream impacts
•  Authority and approvals for changes
should be established and obtained as
appropriate
©
2013

61

April
9,
2013

More
Principles

•  Procedures for emergency changes
should be established to cover
emergency incidents
•  Change control should assess and
confirm that all configuration
management processes are being
followed
©
2013

62

April
9,
2013

31

9/9/13

Types
of
Change
Control

•  A priori
•  Gatekeeping
•  Configuration Control
•  Change Advisory Board
•  Emergency Change Control
•  Process Engineering
•  Senior Management Oversight
©
2013

63

April
9,
2013

A
Priori

•  May I have permission to make that
change?
•  Facilitates an ALM task based
approach
•  Who said you could work on that?
•  Common for defense contractors
•  My friends at the FAA
©
2013

64

April
9,
2013

32

9/9/13

Gatekeeping
&
Conﬁguration

•  Is the release ready to be promoted?
•  What are the downstream impacts?
(Hint – check with the CAB)
•  What happens if we make this
change?

©
2013

65

April
9,
2013

Minding
the
Process

•  Software Engineering Process Group
•  Senior Management
•  What makes something an
emergency?

©
2013

66

April
9,
2013

33

9/9/13

The
29
Minute
Meeting

•  Change control meetings can be
problematic
•  Need to be structured and controlled
•  Entry and exit criteria
•  Control the dynamics!

©
2013

67

April
9,
2013

e-‐Change
Control

•  Routine changes
•  Still requires traceability and
transparency
•  Good way to implement the change
advisory board (CAB)

©
2013

68

April
9,
2013

34

9/9/13

Drive
the
Entire
CM
Process

•  You can drive the entire CM Process
from change control
•  Review the CM Plan
•  Ensure that there are procedures in
place for code promotion
•  Don't forget the fallback plan
•  Identify and manage risk
©
2013

69

April
9,
2013

Retrospective

•  After action review
•  Need open and honest evaluation
•  Opportunity to improve the process
•  Drives the entire release process

©
2013

70

April
9,
2013

35

9/9/13

Release
Management

•  Consists of Release Engineering and
release coordination (PMO)
•  Packaging and identification (e.g.
manifest) of all components built in the
build engineering function
•  Automation to build, package, stage
and deploy releases
•  Don't forget being able to rollback!
©
2013

71

April
9,
2013

Principles

•  Releases should be readily
identifiable with an immutable version
ID
•  Releases should be packaged with all
dependencies identified and controlled
•  Packaging should be automated and
and designed to avoid human error
©
2013

72

April
9,
2013

36

9/9/13

More
Principles

•  Release management should be fast
and reliable to facilitate iterative
development
•  Must be able to audit the release
package
•  Contents of release trackable (audit)
•  Release management source of
information on status of all releases
©
2013

73

April
9,
2013

Manifest

•  Documents contents of release
package
•  Embedded (immutable) version ID
•  Requires procedure to retrieve
version ID
•  Created through automated
procedure
•  Verifiable through configuration audit
©
2013

74

April
9,
2013

37

9/9/13

Release
Maps

•  Complete list of release contents with
MAC-SHA1 or MD5 hash
•  Utility to recreate release map and
compare to version shipped with
release
•  Should be able to verify MAC-SHA1
or MD5 hash
©
2013

75

April
9,
2013

My
Three
Step
Process

•  Common task is to fix the build &
release
•  Observe the first time and take notes
•  Then I drive with my checklist and
developer at my side
•  Third time I have some scripts to
automate parts and my checklist
©
2013

76

April
9,
2013

38

9/9/13

Release
Coordination

•  More of a PMO function that works
closely with Change Control
•  Release Calendar is essential
•  Track requirements completed in
release notes
•  Communicate status of release to all
stakeholders
©
2013

77

April
9,
2013

Staging

•  Essential practice that ensures the
success of the deployment
•  Should be fully automated
•  Must be fully traceable
•  Configuration audit verifies successful
completion

©
2013

78

April
9,
2013

39

9/9/13

Future

•  Complete release and deployment
framework
•  Both open source and commercial
solutions
•  Integration with the ALM
•  Status of my deploy on the dashboard

©
2013

79

April
9,
2013

Deployment

•  Should be the smallest of the
functions
•  Should be engineered to be a push
button lightswitch
•  Requires full traceability
•  Run by Ops
•  Rollback is essential
•  Relies upon release engineering
©
2013

80

April
9,
2013

40

9/9/13

Principles

•  Promoting a release should be as
simple as possible
•  Backing out a release should be as
important as promoting
•  Promoting a release should be fully
traceable with an audit log of all
changes
©
2013

81

April
9,
2013

More
Principles

•  Only Ops should be involved with
deployment
•  Separation of controls essential for
compliance
•  Unauthorized changes should be
detected
•  Configuration audit to verify
•  Retrospective & ongoing improvement
©
2013

82

April
9,
2013

41

9/9/13

I
Make
Mistakes

•  But I will always be able to tell you
what mistakes I made
•  Full traceability
•  Few steps, if any, are done manually
•  Verification that steps were completed
correctly is essential
•  Automate everything!
©
2013

83

April
9,
2013

Communicating
the
Deploy

•  Communication to all stakeholders is
essential
•  Announce outages and completed
deployments
•  Should be automated and part of a
console

©
2013

84

April
9,
2013

42

9/9/13

Smoke
Testing

•  Last step of the deploy is always
testing
•  CM should be part of the QA and
testing function

©
2013

85

April
9,
2013

Current
and
Emerging
Trends

•  Agile principles are impacting CM in
many essential ways
•  We will talk about Agile CM next
•  Agile ALM – related to status accounting
•  Tracking the status of a configuration
item is essentially the lifecycle
•  Agile release planning
•  Focus on process maturity
©
2013

86

April
9,
2013

43

9/9/13

Changing
Landscape

•  Cloud computing and virtualization
•  OSGi – plug and play
•  Application servers to handhelds
•  Can we get that build & deploy done
now?
•  Continuous Integration becomes
Continuous CM
©
2013

87

April
9,
2013

Paradigm
Shift

•  In many organizations deployment means
giving up your weekend
•  We need to shift the way that we look at
deployment
•  Making the deploy a non-event
•  You don't believe so I will quote a few
colleagues
©
2013

88

April
9,
2013

44

9/9/13

Let's
Talk
About
Agile

Agile shifts the focus...
(But you will also see that there will be
much in common)

©
2013

89

April
9,
2013

Goals
of
Agile
CM

•  Rapidly build, package and deploy
•  Reliable and repeatable process
•  Traceability and forensics

©
2013

90

April
9,
2013

45

9/9/13

CM
for
Agile

CM that is adapted to suit the continuous
nature of change that Agile provides
without sacrificing the values of CM.
Adapting Configuration Management for Agile
Teams: Balancing Sustainability and Speed by
Mario Moreira
©
2013

91

April
9,
2013

Agile
on
CM

But we really need to use Agile
principles to implement Agile CM

©
2013

92

April
9,
2013

46

9/9/13

Agile Configuration Management
Individuals and interactions over processes and tools
Working software over comprehensive documentation
Customer collaboration over contract negotiation
Responding to change over following a plan

©
2013

93

April
9,
2013

Characteristics
of
Agile
CM

•  Customer centric (which one?)
•  Rapid iterative development
•  Pragmatic approach to requirements
•  Support for testing
•  Collaborative communication
•  Role in the SCRUM
©
2013

94

April
9,
2013

47

9/9/13

More
to
Agile
CM
than
just
CI

•  Continuous Delivery
•  Lightweight (lean ceremony)
•  Easy to maintain (respond to change)
•  Continuous Integration (of course!)
•  Devops focus on the full ALM
What are the first seven things?
©
2013

95

April
9,
2013

Agile
Focus
On
Seven
Items

1. 
2. 
3. 
4. 
5. 
6. 
7. 

Source Code Management
Build and Release Engineering
Environment Configuration
Continuous Integration
Continuous Deployment
Verification and Validation
Devops

©
2013

96

April
9,
2013

48

9/9/13

Agile Release Train (ART)
Making each product a successful and
routine event – an event that is indeed
planned and eagerly anticipated yet one
one that happens almost on autopilot
Dean Leffingwell’s Agile Software
Requirements, p. 299
©
2013

97

April
9,
2013

Build
in
the
Cloud

•  Virtualization allows you to create a
fully resourced build box
•  Make the build fast with no penalty for
frequent builds
•  Builds should be logged and traceable
•  Make sure you can tag “interesting”
builds and purge unneeded builds
©
2013

98

April
9,
2013

49

9/9/13

Architect
Your
Build
for
CM

•  Architect Your Application to facilitate
CM (e.g. immutable version IDs)
•  CM also helps facilitate an effective
architecture
•  Overly complex builds are a huge
waste
•  Rapidly changing architecture can
outpace the build
©
2013

99

April
9,
2013

Lessons
From
ITIL

•  Configuration Management Database
(CMDB)
•  Federated CMDB
•  Configuration Management System
•  Definitive Media Library (DML)
Devops = Agile + ITIL
©
2013

100

April
9,
2013

50

9/9/13

Puppet/Chef

•  Automate provisioning, patching and
configuration of operating system and
application components
•  Systems integration framework
•  Scalable and extensible
•  Used in other deployment frameworks
www.puppetlabs.com www.opscode.com
©
2013

101

April
9,
2013

•  Rapid iterative deployment
•  Automate everything
•  Keep the deployments small
•  Minimize risk
•  Easier to deal with problems
•  Ability to fall back is important

©
2013

102

April
9,
2013

51

9/9/13

Common Problems
•  Deployments can be risky
•  Missing a single step can result in
problems
•  Too many mistakes
•  Takes too long
•  No way back
•  Assumed defeat
©
2013

103

April
9,
2013

Deployment
Pipeline

A deployment pipeline is … an
automated implementation of your
application’s build, deploy, test and
release process
Jez Humble and David Farley’s
Continuous Delivery, p 3.
©
2013

104

April
9,
2013

52

9/9/13

Aim of the Pipeline
•  Makes building, deploying, testing and
releasing software visible to everyone involved
•  Improves feedback so that problems are
identified, and so resolved, as early in the
process as possible
•  Enables teams to deploy and release any
version of their software to any environment at
will through a fully automated process (p. 4)
©
2013

105

April
9,
2013

Antipatterns

•  Deploying Software Manually
•  Deploying to Production-like
environment only after Development is
complete
•  Manual Configuration of Production
Environments
Continuous Deployment, p. 7 – 10
©
2013

106

April
9,
2013

53

9/9/13

Verification & Validation
•  Deming – build quality in
•  Test your own framework
•  Consider the ergonomics of your
automation

http://cmbestpractices.com © 2013

107

April
9,
2013

Ergonomics
of
Build

&
Release

•  Cockpit of a plane
•  Controls are easy to read
•  Traceability
•  Designed to avoid mistakes

©
2013

108

April
9,
2013

54

9/9/13

Devops
•  Synergy of Agile & ITIL
•  Full lifecycle approach
•  Good communication to all stakeholders
•  Break down barriers
•  Don’t forget separation of roles

http://cmbestpractices.com © 2013

109

April
9,
2013

Dev/QA
Focus

•  Development
•  QA & Testing
•  Operations
•  Self Managing/Organizing Teams

©
2013

110

April
9,
2013

55

9/9/13

What
is
DevOps?

•  New Term for...
•  Portmanteau
•  Agile Systems Administration
•  Agile Operations
•  Group of Principles
Now that we cleared that up!
©
2013

111

April
9,
2013

New
Term

•  Group of concepts
•  Been around for a while
•  Use case is compelling
•  Stimulating discussion
•  Necessary to meet demand

©
2013

112

April
9,
2013

56

9/9/13

Portmanteau

•  Combination of two words
•  Development
•  Operations
Development and Operations have
very different goals
©
2013

113

April
9,
2013

Conﬂict
Between
Dev
&
Ops

•  Development focused on delivering new
functionality
•  Operations is focused on providing
continuous (reliable) services
•  Manage risk!
One time I was asked to break the rules
©
2013

114

April
9,
2013

57

9/9/13

Trying
to
make
the
deadline

•  Trading system was tested and
passed
•  Few bugs discovered
•  I was asked to deliver a different
version than was tested
How does DevOps help balance?
©
2013

115

April
9,
2013

DevOps
is
also

•  Emerging Best Practices
•  Collaboration between Dev & Ops
•  Application and Systems Deployment
•  Software and Systems Development
But is DevOps Agile?
©
2013

116

April
9,
2013

58

9/9/13

What
about
Agile?

•  Agile Systems Administration
•  Agile Operations
•  Waterfall needs DevOps too!
Release Antipatterns...
©
2013

117

April
9,
2013

Release
Antipatterns

Deploying software manually
l  Deploying to a production-like
environment only after development is
complete
l  Manual configuration of production
environment.
l 

So what really is DevOps?
©
2013

118

April
9,
2013

59

9/9/13

DevOps
is
Really...

•  Developer and Operations
collaboration
•  Crossfunctional team
•  Knowledge Management
•  Better communication
Time to get rid of silos
©
2013

119

April
9,
2013

Dev/QA
Focus

•  Development
•  QA & Testing
•  Operations
•  Self Managing/Organizing Teams

©
2013

120

July
15,
2013

60

9/9/13

Agile
Focus
On
Seven
Items

1. 
2. 
3. 
4. 
5. 
6. 
7. 

Source Code Management
Build and Release Engineering
Environment Configuration
Continuous Integration
Verification and Validation
Devops

©
2013

121

April
9,
2013

Skills
for
CM
guru

•  Hands-on technical
•  At least some development skills
•  Strong scripting (e.g. Perl, Python,
Ruby)
•  Knowledge of some frameworks or
standards
•  Process orientation (enjoy improving
the process)
©
2013

122

April
9,
2013

61

9/9/13

CM/Devops

•  Flexible technical background
•  Good knowledge of development
•  Knowledge of QA/Ops
•  Strong automation skills
•  Some systems administration
•  Ability to work across silos

©
2013

123

April
9,
2013

Toolsmith/Devops

•  Strong technical background
•  Strong scripting skills
•  Diving deep into the tools including
troubleshooting
•  Understands toolchains and finds
flexible solutions
•  Process orientation – focus on
traceability
©
2013

124

April
9,
2013

62

9/9/13

IT
Governance
&
Compliance

•  IT Governance needs to be in
alignment with corporate governance
•  Financial reports needs to be
accurate
•  Separation of controls
•  Security measures to prevent
unauthorized access
•  Audit in place for intrusion detection
©
2013

125

April
9,
2013

Sox
Compliance

•  Section 404 of the Sarbanes Oxley
Act of 2002
•  Using ISACA Cobit 4.1
•  34 high level IT controls
•  PCI compliance
•  SAS-70

©
2013

126

April
9,
2013

63

9/9/13

ISO 9001
•  Establishes the quality management
system
•  ISO 90003 is the software standard in
the 9000 family of standards
•  Uses ISO 12207 (or 15288) to specify
lifecycle processes
•  ISO 10007 for CM
•  IEEE 828, EIA 649-A, Mil Std coming!
©
2013

127

April
9,
2013

Which
Standards?

•  IEEE 828 – CM Planning
•  EIA 649-B – Non compliance
•  ISO 90003 to support QMS
•  Full lifecycle ISO 12207
Tailor !

©
2013

128

April
9,
2013

64

9/9/13

Moving
Upstream

•  Dev to CM to QA to Ops
•  Cross functional focus
•  Speed up development
•  Build a great deployment architecture
•  Give it to Devs as a service!

©
2013

129

April
9,
2013

Frameworks

•  ITIL v3 including CMDBs, federated
CMDBs, CMS, DML…
•  Cobit for SOX
•  CMMI ->>>> Agile

©
2013

130

April
9,
2013

65

9/9/13

The
CM
Process

•  Should be Lean
•  Processes need to be reviewed
•  Tailor down or tailor up
•  More collaboration and consensus
building
•  Use standards and frameworks

©
2013

131

April
9,
2013

Assessment

•  First step is to assess current
practices - “As-Is”
•  Compare to industry standards and
frameworks
•  Determine “To-Be”
•  Create a plan for improving your CM
processes
©
2013

132

April
9,
2013

66

9/9/13

Plan
for
Improvement

•  Improve training and use case for
source code management
•  Improvement build automation
•  Setup or improve continuous
integration
•  Automate package and deployment
•  Create procedures for configuration
audit
©
2013

133

April
9,
2013

Conﬁguration
Management

•  Configuration Identification
•  Status Accounting
Tracking and Controlling Changes to
Configuration Items
©
2013

134

April
9,
2013

67

9/9/13

Goals of this Course
•  Implement Effective Source Code
Management practices including variants
•  Automate build, package and deploy
•  Establish effective IT Controls
•  Use industry standards and frameworks
•  Create a CM function that grows &
improves
©
2013

135

April 9, 2013

More Goals of this Course
•  Use CM to support development
•  Understand the classic four CM functions
•  Introduce the core CM framework
•  Examine Current and emerging trends
•  Implement Agile CM – the first 7 things
•  Establish IT governance and compliance
•  Establish your own plan for CM!
©
2013

136

April
9,
2013

68

9/9/13

Conﬁguration
Management

Best
Practices

Bob
Aiello,
Principal
Consultant
and
Author
of

Conﬁguration
Management
Best
Practices
:

Practical
Methods
that
Work
in
the
Real
World

http://www.linkedin.com/in/BobAiello


137

CM Best Practices Consulting © 2013

69

Configuration Management Best Practices

Recommandé

Recommandé

Contenu connexe

Tendances

Tendances (20)

En vedette

En vedette (15)

Similaire à Configuration Management Best Practices

Similaire à Configuration Management Best Practices (20)

Plus de TechWell

Plus de TechWell (20)

Dernier

Dernier (20)

Configuration Management Best Practices