WP How APIs Enhance A Devops Model
WP How APIs Enhance A Devops Model
Executive summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Introduction .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
About MuleSoft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Executive summary
4
Introduction
5
wider company. Not only are these artifacts replicated in
business silos, killing business productivity, but they are locked
away from other parts of the business, who might be able to
make use or and extract further value out of them.
Successful DevOps adoption has to include both an accelerat-
ed and efficient production model and an efficient, balanced
consumption model. Companies that successfully implement
DevOps should be stewards of a new operating model with
a virtuous cycle in which teams produce reusable assets in a
manner visible to central IT, which then enables consumption
of those assets by making them discoverable and available for
self-service, all the while monitoring feedback and usage.
In this model, those developing applications and other assets
must change their mindset to think about producing assets
that will be consumed by others in the business. And, in
turn, central IT needs to change its mindset to an enabler of
reusable, self-service consumption. Any assets built need to
be discoverable and developers need to be enabled to self-
serve them in projects. The virtuous part of this production-
consumption cycle is to get active feedback from the
consumption model along with usage metrics to inform the
production model.
6
DevOps: A profound cultural
and technological shift
8
The benefits of adopting DevOps
PROFITBRICKS PRESENTS:
LANDSCAPE VIEW OF INFRASTRUCTURE AUTOMATION COMPANIES LATE 2015 (V 1.0)
NETWORKING
SECURITY
TESTING
PROVISIONING
Docker
Swarm
Created by ProfitBricks, Inc. 2015 ProfitBricks logo are trademarks or registered trademarks of ProfitBricks, Inc. All other registered trademarks or trademarks are property of their respective owners.
9
buckets, from development and test, to deployment and
monitoring, down to security and collaboration. Operations
solutions like ServiceNow or Remedy for IT Service
Management and JIRA for issue tracking are also common.
Many of these technologies are open-source, or at least have
open-source versions, so it’s very easy for our customers to get
started relatively quickly in building out the foundations of their
DevOps practice.
10
How to use Anypoint Platform to set
up DevOps in your organization
11
common Java testing framework, are used quite often within
our customer base to deliver test automation successfully.
Version control is something that most developers today are
very comfortable with. Developers know how to commit code
across a disparate set of teams so they can drive parallel
development and get instant visibility into changes that are not
compatible with what others have been doing.
MuleSoft enables version control best practice via extensibility
to GitHub, SVN, Team Foundation, and many others version
control systems commonly used within enterprises.
Then, finally, there is deployment automation. As with
creating a recipe for how an application is built, a recipe
for how applications are deployed should also be created.
Which environments should they be deployed to? How do
you promote applications across environments: dev, QA,
performance, security, stage, and production? How do the
configurations change by environment? How do you create
rules to block software from being deployed if it fails critical
tests? Jenkins is a common solution that our customers use to
deliver on this aspect of the pipeline.
Great DevOps practices focus on one additional element:
reuse. As an asset is moving through the pipeline, how can
you ensure others are aware of it, can access it securely, and
know how to use it effectively? The API-led connectivity model
is very informative here, as most developers are comfortable
using Developer Portals and requesting API client keys and
secrets. Industry trends suggest organizations are building
an automation pipeline across the entire lifecycle of their
APIs; there are a number of technologies across that lifecycle
that have emerged as best practice, many of which MuleSoft
supports and our customers are already leveraging today.
12
Lifecycle stage Stage activities Anypoint Platform
components
13
DevOps and API lifecycle capabilities
on Anypoint Platform
+
Engage Mocking service
te
API notebook
Fe
Engage
API spec
Engage Studio
Si m
la
te ign
u
D es Engage MUnit
14
When the code is ready to be shipped, MuleSoft customers
often leverage Jenkins, Puppet, Chef, or HP ALM for both
deployment automation and the orchestration of release
processes across a number of different environments: dev,
test, stage, and production. Puppet and Chef can also be
used to automate the installation and configuration of Mule
runtimes. Docker is rapidly gaining popularity to reduce
the complexities of continuous deployment, and the Mule
runtime can be shipped in a Docker container. And through
the management agent, Anypoint Platform for APIs and
Anypoint Platform CLI, users can further automate processes
using configuration automation and management tools such
as Puppet, Chef, Ansible, and Salt, instead of the Anypoint
Platform user interface.
Deployment Operations
Prod
Anypoint Platform
Exchange
Studio
Test
Commit
Dev
GitHub
kubernetes
JUnit
Mgmt agent
Jenkins
Ship
Image
15
Finally, for monitoring and analytics, developers and Platform
operators have the option to either use the out of the box
capabilities provided by Anypoint Platform or plug into any
existing third-party tools: monitoring and analytics tools such
as Splunk, Nagios, AppDynamics, New Relic, or the open-
source Elastic Stack: Elasticsearch, Logstash, and Kibana. All
this is made possible through the extensibility of the Mule
runtime engine.
16
Shifting from production to
production + consumption
The more discoverable you make the assets that you create
through your DevOps pipeline, the faster you will become
at making them ready for reuse, enabling line of business IT,
partner developers, and internal developers to securely self-
serve on top of these assets. This will make your organization
more innovative in creating new products and services because
you’re exposing capabilities to audiences that may or may
not have otherwise known about them, or have had access
to these pieces of data or connectivity artifacts. Reusability
enables speed, innovation, and agility.
Consumption
LoB IT, developers
Discoverable Reusable
assets
Production
Central IT/LoB IT
17
exposing the assets they create for maximum business value, is
an approach to integration called “API-led connectivity.” In our
experience, this approach provides the answer to producing
reusable assets and enabling their consumption to drive useful
speed and innovation.
API-led connectivity provides an approach for connecting and
exposing assets. With API-led connectivity, every asset becomes
a managed API, a modern API, which makes it discoverable
through self-service without losing control.
App Dev
Experience APIs
Innovation and digital products
Knowledge/asset exchange
Developer communities
LoB
Decentralization
Dev/IT
Process APIs
Agility and new value creation
Accessibility
and
ownership
19
The benefits of the MuleSoft
approach to DevOps
21
C4E asset base API consumption–Internal API consumption: External
Number of assets Q1 ‘16 Q2 ‘16 Target Number of Q1 ‘16 Q2 ‘16 Target Number of Q1 ‘16 Q2 ‘16 Target
Asset and API
11 11 5
in the catalogue) 8 30 0 10 consumers 4 10
(+3) (+1) (+2)
APIs
Q1 ‘16 Q2 ‘16 Target Time to MVP (i.e. Q1 ‘16 Q2 ‘16 Target Time to POC Q1 ‘16 Q2 ‘16 Target
enablement
*
A subjective measure of knowledge/capacity
22
Conclusion
23
About MuleSoft
24