This article will provide an example delivery pipeline used through continuous integration, as well as the DevOps tools involved in the automation of building, testing and deploying code through your software development life cycle (SDLC).
Background
These days, it’s almost impossible to go anywhere without hearing about DevOps. In short, DevOps is the effort to improve communication and synergies between development organizations and operations organizations—hence the term, “DevOps.” The main focus is to improve the efficiency, quality and speed to market within the software development world.
Today, you also may hear and see a lot about continuous integration and continuous delivery/deployment. There are differences between the two for sure, but for simplicity’s sake, this blog focuses on breaking down the different stages of the process. This way you can get the basics and hit the ground running with a delivery pipeline in your organization. The delivery pipeline can be broken down into a few major buckets of work, or stages, as mentioned below.
In my opinion—and, again, trying to keep things simple for now—the stages can be broken down as follows:
- Source Code Control (Management)
- Build Automation
- Unit Test Automation (could also include Integration Testing here as well)
- Deployment Automation
- Monitoring – not included in this discussion, and can be added at any time
I have outlined the progression of these stages in Figure 1 below. Let’s review each of the stages in a little more detail.
Figure 1 – Continuous Integration Delivery Pipeline Sample
Source Code Control (Management)
Background
Source code management, or source code control, certainly is not a new topic. This has been around for decades and has evolved over time. In a nutshell, an organization stores its code in a source code control system or repository so it can be tracked, maintained, versioned and audited. Developers should not store the code on their laptops or virtual machines and trust that will suffice for managing the code.
Possible Tools
- Git is probably the most widely used SCM system out there. It is an open-source system.
- Subversion (SVN) has been around for quite some time. It is also an open-source system and is heavily used in many organizations, but there has been more of a push toward Git.
Git and Subversion comparison can be found here.
Build Automation
Background
Once a source code management system is in place and actively being used by a development team, the team will need to be able to compile and build their code. This is probably the first step in the whole chain of continuous integration events—it’s what gets the ball rolling. The code needs to build cleanly before an organization can even think about deploying out to its environments for testing and production.
Possible Tools
- Gradle is an open-source build automation system used by top companies including Netflix, Google and LinkedIn.
- Maven is another open-source build automation system.
Gradle and Maven comparison can be found here and here.
Unit Test Automation
Background
Developers unit test their code to ensure the functionality they are building works as expected. In an ideal world, the development team should be saving these unit tests so they can be reused and put into a regression test bed.
Possible Tools
There are multiple tools out there for helping developers unit test their code. Many of these tools are open-source and can be used freely.
- JUnit is an open-source unit test framework pretty widely used.
- CA DevTest allows for the automation of unit testing, as well as a few other bells and whistles, such as service virtualization.
Deployment Automation
Background
For the last stage in the process, delivery teams need to deploy their code/applications out to various test environments and, of course, production. To reduce errors and overhead in the deployment process, while increasing speed to market, this step can be automated through a variety of tools and methods.
Possible Tools
- IBM Urbancode uDeploy allows users to model a process and orchestrate their deployment. This process then can be repeated across all an organization’s environments, and tweaked for each environment as needed.
- Ansible is an open-source IT automation tool. It can be used for everything from configuration management to product installation to application deployments.This tool is rapidly gaining acceptance and momentum in the DevOps community.
Conclusion
This was simply a quick overview for companies to get started down the path of continuous integration and the DevOps world. Figure 2 is an example of what the whole flow looks like from committing the code to repo to deploying the code to an environment, to help put things into perspective. It’s easy to see how the tools interact with each other on the journey to production.
Figure 2 – Continuous Integration Process Flow
About the Author / Seth Gagnon
Seth Gagnon is a Middleware / Infrastructure / DevOps architect with more than 12 years’ experience in the field. Please visit sethgagnon.com for additional related articles. Connect with him on LinkedIn and Twitter.