The World in this century is moving very fast and especially
software Development. The advent of Social Media has further made this complicated,
as the Customer Expectation as well as Interactions is Intense. An increased
focus of User Experience and Expectation for Quicker Change implementation is
evident. The changes here include both the changes proposed by the Customers as
well as the changes, which the process requires such as the process itself being
flexible enough to be changed quickly. The traditional development practices
are unable to cater to such a massive change in the Mind-set and Expectations.
Another important aspect to deal with is the performance of the Applications
should be competitive so that it leads to an actual adoption of the software
product.
With all these challenges in
place, any change in the process becomes next to impossible, as there are
barriers to adoption.
This need of rapid delivery applies to both type of projects
which are in Development or which are in Operations. In projects which are in
Development, the changes are frequent as well as the deadlines are
aggressive. Any changes in the later
phases can pose serious threats to successful delivery. In Operations, there is
a strong concern of introducing minimum changes. Although the changes are less,
however they need to checked rigorously that they should not hamper the
existing product being used by Customer on a daily basis. Further, the change
should be quick and prompt as per customer expectation. These demand a complete
revamp in the process as well as the mind-set to own such a process, which did
not exist before.
Some of the Critical Questions in Industry regarding to this
are:
- How can we Release on time and in less time?
- Is there a way in which the Operations are happy to have frequent changes?
- Is it Possible to Deploy Simultaneously and frequently on multiple Servers?
- Can we Develop and Test in Parallel in as Good as Production Environment?
- Is there a way to avoid last minute surprises aka Pandora’s Box?
- How can we perpetually reduce IT Costs and control any effort or resource wastage?
- I have adopted agile practices. How can I Extend them further?
And, the List goes on. The Solution to the above problems is
by implementing DevOps.
DevOps is not about technology, but a cultural phenomenon.
Hence, it is important to communicate effectively with the IT and delivery
teams. Building Such DevOps Culture would require a top down approach in
organization where leaders would lead by example so that the flow is imbibed
all across. This can be more effective with enablers such as Continuous
Delivery and Continuous Testing.
Continuous Delivery employs Continuous
Integration. It also demands the use of agile and lean principles across the
Software Delivery chain. This means that the speed of Delivery with an equal
focus on quality should be there from Inception to production release and
Customer Satisfaction.
It is the need of this hour where commitments to ship
reliable deliveries is even stronger. DevOps can be implemented using Team
Foundation Server effectively. TFS allows us to construct a release pipeline to
implement DevOps. Since Visual Studio 2012, there has an increased focus on
Combined Engineering and DevOps. The following features employ quicker
implementation of DevOps using TFS:
1.
Creating Product Backlog – DevOps needs
interaction of all the stakeholders like Business Analysts, Testers and
Developers on a frequent basis so that we can implement changes. To start with
this, we need a product backlog ready and accessible to all. This enables us to
know what all is pending and to be dealt with. TFS in the enhancement after VS
2012 introduced the enhanced templates for Scrum as well as Agile
processes. These are specifically
designed to optimally create product backlog.
Additionally, the Kanban board is of great importance to visualize the Progress in a centralized manner across all the Stakeholders. Board in TFS allows us to implement this across phases and stages. The communication is loud and clear so that there is no gap in understanding of teams.
Customization of stages supports the
lean thinking and changing the process easily and efficiently leads to
Continuous Improvement.
An important aspect of this is to avoid any kind of Over Development or Under Development. Hence, anomalies are critically analysed in the Initial Phase by multiple Stakeholders. This supports the cultural movement as required in DevOps as there is complete collaboration as well as Visibility to evaluate the adoption amongst teams.
2.
Building a Release workflow: Continuous
Integration, Continuous Testing and Continuous Delivery are pillars required
for DevOps implementation. Furthermore,
whatever can be and thereby should be automated. This would allow minimum
hindrance to the Process Changes and would rather facilitate it.
“Continuous Integration is a software development practice
where members of a team integrate their work frequently; usually each person integrates
at least daily - leading to multiple integrations per day. An automated build
(including test) to detect integration errors as quickly as possible verifies
each integration. Many teams find that this approach leads to significantly
reduced integration problems and allows a team to develop cohesive software
more rapidly.” - Martin Fowler
TFS has the provision for implementing Continuous
Integration by customizing the Build Definitions. It heavily relies on Build only Once
strategy. This in simple words means that once a Developer checks in code that
the Code is built simultaneously to check that the code submitted by the
Developer is not interfering with the existing code. In case there are errors,
the Code check-in is unsuccessful. This
further is an automated process.
The Next Step to this is to have Simultaneous deployments
across multiple environments. This again
is automated depending on the Success of Previous Step. Along with this data
for monitoring this is gathered to closely watch the success per stage. DevOps Workbench express allows us to create
Orchestrations. This has support for
multiple Microsoft Technologies including SharePoint, MSSQL, Azure, .NET
Applications, ASP.Net Applications. We can create Deployment Orchestrations for
performing different Steps in order to enable Continuous Integration and
Delivery.
Creating this kind of a Build workflow is a complex
activity. To reduce this complexity, DevOps Workbench Express is a rescue. Such
Customized orchestrations aka Build Definitions can be readily used to create
the Build Pipeline using this.
Until now, we have an orchestration to handle Building the
Source Code and deploying it to target environments simultaneously. The Next
Big Step is to adhere to Continuous Testing. Unit Test is effective to do this.
Employing Automated BVT in the form of Unit Tests can sanitize a Build. Any
failure of the tests would lead to subsequent rollback of the build and thereby
ascertain fool proof Regression Testing. This necessarily is automated and
automatically triggered with a build. Using Visual Studio 2013, we create unit
tests and add them to the project so that they are automatically invoked with
the build using some basic Settings in the build definition.
Unit Test Generator is as Add on for VS 2015 and provides
facility to quickly add unit tests. These Unit tests are tightly couple to the
projects and are an efficient mechanism to maintain regression testing.
Continuous testing can be achieved by this.
4.
Continuous Monitoring: We all want progress, but
if we are on the wrong road, progress means doing an about-turn and walking
back to the right road; in that case, the man who turns back soonest is the
most progressive. Continuous monitoring thereby becomes extremely important to
identify issues and act upon them at the earliest.
TFS provides reports at various
levels to cater these needs. Some of these reports for constant monitoring are:
a.
Project Management Dashboard
b.
Progress Dashboard
c.
Quality Dashboard
d.
Test Dashboard
e.
Bug Dashboard
f.
Build Quality Indicator Reports
g.
Bug Backlog Reports
h.
Burn down and Burn Rate Reports
Using these reports provides a
holistic approach with complete end to end visibility for all the stakeholders.
Having said all of this, it
should be clearly understood that enablers to change could only help if the
DevOps culture is in place. Leads of the organization must stress on the
implementation of such culture and should be open to feedback for further
process improvement. It Might take time to adopt this but the results are very
satisfactory.
To Sum it up, the following are
the vital benefits of DevOps:
Minimized Conflicts between Development and
Operations
Greater Scale ability and Flexibility in Process
Implementation
Collaborative and Integrated Development among st
Teams
Frequent and Faster time to Delivery
More Stable and reliable Releases
Continuous Monitoring for Complete Visibility
across Stakeholders
Reduced Outages and minimal errors to Operations
No comments:
Post a Comment