A Contributed Article by Wesley Pullen, General
Manager & Vice President, Deployment Solutions at Electric Cloud
What once was considered a grassroots movement in the
software world, Release Automation/Deployment Automation is now at the heart
and soul of the largest and most innovative companies in the world. The
transformative benefits of this practice are now critical for the organizations
to facilitate in order to scale and automate the lifeblood of the digital age -
better software faster with greater quality and innovation.
A critical piece of the software delivery process within large and small IT organizations
focuses on something called Application Release Automation (ARA), which helps
enable the consistent, repeatable and auditable process of packaging and
deploying an application(s) or update of an application(s) from development
across various environments, and ultimately into production. ARA is
increasingly important for many reasons, as organizations struggle with long
release cycles, configuration-related errors and not being able to track
processes for audit and compliance purposes.
Beyond the key benefits of implementing ARA, here are five tips/trends
that we see at Electric Cloud that demonstrate the reasons why enterprises (from
financial services and retail to telecom and healthcare) are adopting ARA tools
and strategies:
1)
Root
Cause Analysis: As development and operations teams have begun working
together (via DevOps), in 2015 these teams still have an opportunity to try new
methods to accelerate and automate processes. For example, they can be doing
things like comparing servers, snapshotting the servers, etc. In the event of
finding a difference between those servers, these teams have an opportunity to provide
a delta package which is deployable to the server of concern. Instead of needing
to deploy dozens of pieces of an application(s), if there's only two pieces that
are different (out of sync...), you just deploy those two pieces. This will reduce
the time it takes to deploy by only needing to pull out the logs that are needed
to see what those differences are.
2)
Need for
Insight & Discovery into App Development: Oftentimes when companies
leverage ARA or DevOps solutions, getting them stood up in middleware server
environments with tons of applications can be a daunting task. But, what if you
are able to point to the server and have it tell you what applications already
exist so you can drag and drop those applications into a model? And then go from
there to do your deployments? In 2015, there will be moves towards these types
of discovery processes to make deployment faster and easier to gain greater
insights in a much quicker pace. When it comes time to train the trainer or
train other resources on newer technology, it can now happen in minutes versus hours
or days.
3)
Smarter
Deployments: In the past, a lot of companies did "spin up" and "spin down"
for testing resources, almost providing QA and development a self-service
portal. It's important to have on-demand resources because when someone needs a
testing instance, they'll need to layer down the firewall rules, layer down the
application, the database, the middleware server and all these pieces, and then
run the tests, parse the results, and then spin the resource down so that they
are no longer charged. In the future, organizations will start leveraging a
smart deploy methodology. Under this methodology, these "spin up" and "spin
down" processes are called constructs where you can do cloud scenarios and the
delta package deployments to minimize time as well.
4)
Containerization:
Going forward, we will see companies start to migrate to containerization for
the purposes of having ready-made containers to provide a bill of materials
inside the container and manage that container through deployment automation as
well. Docker and its fast popularity is
just the tip of the iceberg and many enterprises are now starting to employ
Docker containers within their environments.
The reduced overhead in system resources and not having to manage
hypervisors is very appealing.
Deployment Automation and working and integrating with Docker is a must
for any ARA vendor in 2015.
5)
Big Data
Trend Analysis: Now that software development teams have done hundreds, if
not thousands of deployments to the lower environments, they can start to analyze
and study those deployment trends and provide windows for smarter server
allocation and reservation in the future. So if someone knows that they're doing
a deployment and someone else has done this deployment 100 times in QA and Dev,
by the time they get to my staging and production environment there is already some
knowledge about this deployment. For example, they will know how long it will
take, the kind of the resources needed, etc. So, when it's time to schedule a
maintenance window, people can leverage actual, trended data versus guesswork.
These types of evaluations will continue to emerge in 2015 and beyond.
##
About the Author
Wesley
Pullen is the General Manager and Vice President of Deployment
Solutions at Electric Cloud.
He brings more than 18 years of experience
in software development methodologies and design standards, applying
this experience to both commercial and private sectors. Prior
to joining Electric Cloud, he was the vice president of the Emerging
Solutions Group and Global ARA Solutions Group within Automic Software.
Pullen earned his bachelor's degree in Electrical Engineering from the University of Pittsburgh in Pennsylvania.