background

What is DevOps?

27 Jul 2022 in

Introduction

DevOps is part of the Project to Product revolution. Project to Product is a fast-emerging software development approach that helps organisations stay nimble and flexible by moving away from projects and moving to product delivery. This big picture approach helps development teams move away from a siloed, project-centric approach, to a higher-level, holistic-focused, business product approach. DevOps is a critical component of this fast-emerging shift designed to increase business value.

DevOps unites development and operations to provide better quality software and a better experience for the user. Cross skilled teams are centred around a product and have all the skills required to build and run software. Development and operations activities are all treated like engineering problems to be solved with everything as code.

DevOps practices provide a faster way to deliver high-quality applications than traditional software development. As a result, businesses better compete by reducing inefficiencies and cutting costs.

To understand DevOps, you need to learn about automated software development lifecycle (SDLC) and the Continuous Integration/Continuous Delivery/Continuous Deployment (CI/CD) pipeline. These concepts each play a part in streamlining workflows and speeding up innovation.

Contents

DevOps Culture

DevOps culture is not only tools and practices, but also a high performance mindset. DevOps culture focuses on automating and integrating processes between development and operations teams. This philosophy stresses empowerment, communication and collaboration, and automated technologies.

The main cultural belief at the centre of the DevOps culture is “we are better together than alone.” This idea is supported by specific practices: working in the open, not blaming, learning, and failing fast.

DevOps Principles

Although DevOps principles can vary from organisation to organisation, there are a few consistent concepts. Among them are collaboration, communication, automation, and continuous improvement.

Collaboration is an essential premise of the DevOps approach. Teams communicate, offer feedback, and work together through the orchestration of the development and deployment cycle. The result is better full-stack development as teams share responsibilities from start to finish.

A primary goal in DevOps is to cut as many manual processes as possible from the software development lifecycle. Automation frees up dev teams to write more code and develop new features. This crucial part of the CI/CD pipeline helps reduce errors and boost productivity.

An excellent solution to help developers automate code more quickly is Test Driven Development (TDD). TDD is an agile development framework that focuses on coding, testing and refactoring. The process starts with writing the test – first. Then, write just enough code to pass the test, and use the test to implement the code more efficiently. This is a highly efficient process where coding, testing and design are tightly coordinated.

DevOps for Product Development

Under the umbrella of the Project to Product philosophy, the DevOps methodology is the next evolution in software development. This system elevates the development process by inviting dev teams to take a more holistic view of the business. Instead of thinking about software development as a defined project with a beginning and end date, thinking about DevOps for product development shifts the focus to Continuous Improvement.

The Continuous Improvement delivery process is a crucial part of the iterative product development and management approach. Teams focus on experimentation, eliminating waste, and optimising speed, cost, and delivering a great customer experience. Continuous Improvement and Continuous Delivery enable new releases through the pipeline that improve efficiency while also increasing business value.

DevOps for Product Management

DevOps practices are proven to be valuable tools when working with commercial off-the-shelf software (COTS), new product development, and any and all software engineering. Beyond that, DevOps principles are also quite valuable for ongoing product management. Like development, product management benefits from a holistic approach that looks at the product and how it serves the customer.

DevOps philosophies help aid product managers in ensuring that at every stage — from software development, through testing and even pricing — that the product is absolutely focused on the customer and serves their needs. This systematised and strategic approach promises an efficient way to go to market, and to continue to optimise and iterate.

Customer-Centric Action

DevOps gathers feedback from users to create products based on their needs. Rapid deployment and real-time live monitoring provide instant feedback on product usage. Teams use these insights to make improvements sooner rather than later.

Begin With the End in Mind

Understanding a product’s purpose is to solve a user’s problem is a significant part of DevOps. With this mentality, teams don’t work in silos or create a product based on how they think the customer will use it. Instead, DevOps teams share a holistic view of the product from development to implementation.

DevOps Practices

DevOps aims to improve development and boost efficiency by encouraging teamwork. Best practices include agile project management, shifting left with CI/CD, cloud, infrastructure as code (IaC), toolchains, automation, monitoring, observability, security, and continuous feedback.

Agile Project Management

In agile project management, teams constantly check requirements, plans, and results. Then, they can use feedback to make adjustments.

Agile workflow phases include:

  • To do: Work in queue
  • In progress: Work teams are inspecting
  • Review: Finished work, but teams are reviewing
  • Done: Finished, deployed to live, and customers can use it

In agile software development, teams divide large projects into minor tasks, breaking down work into small stories that can be delivered within a sprint. Core project management frameworks like scrum and kanban enable teams to plan, track, and measure work increments. Utilising these proven frameworks also help with DevOps adoption.

Shifting Left With CI/CD

“Shift left” means DevOps engineers start testing early in their code development processes. Test Driven Development fills this function nicely. Teams start with a test, and they test early and often, in order to improve code and fix bugs while working in the relevant codebase section. If teams don’t test until the code is ready for production, it’s harder to backtrack and correct issues.

TDD is a highly functional approach that helps DevOps engineers shift left, but unfortunately simply testing sections of the code is not always enough. That’s where test coverage comes in. Test coverage tools help teams understand which lines of code have been checked by the test, revealing the percentage of the code that has been validated.

When testing, avoiding mistakes made with manual processes is crucial. For that reason, test automation is a must. It’s especially useful for finding bugs and examining use cases. Test automation is not only faster, it ensures quality by uncovering problems early, so they can be fixed earlier. This efficient process helps to get products to market faster.

DevOps Tool Management: Toolchains

DevOps is about integrating development and operations. To facilitate integration, teams work with a set of tools, called a toolchain, to ensure they use the right DevOps tools in each lifecycle phase.

Teams can select an all-in-one toolchain, which is great for getting a quick start, but it may not integrate with all of the open source or proprietary third-party tools. Or, teams can build a customizable toolchain to bring their existing tools into the fold. But what if everyone’s tools don’t integrate? This solution may be time-consuming and make sharing information difficult.

Automation

CI/CD allows teams to automate the merging of code into the primary repository. Automated testing is crucial to DevOps practices, including end-to-end, integration, unit, and performance tests. Embracing automation moves teams into the feedback loops, which is a continuous feedback and iteration cycle.

Monitoring

Monitoring the DevOps pipeline prevents a failed test or broken build from causing delays. Automation may speed up the development process, but it’s counterproductive if a failure goes undetected. Teams also watch applications to identify performance deficiencies. This way, they don't hear about the problem from the customer.

Observability

Moving from on-premise servers to cloud-based applications makes monitoring more complicated. So DevOps focuses on the three pillars of observability: logs, traces, and metrics. This information helps teams better understand internal functions by observing external behaviours.

Continuous Feedback

Gathering continuous feedback ensures DevOps teams have the information they need to succeed. A steady flow of input alerts teams about code test results and pipeline failures as soon as possible. Regular responses notify operations about performance deficiencies, production failures, or reported bugs. Continuous feedback optimises both speed and quality.

Benefits of DevOps

There are many technical, cultural, and business benefits when teams execute DevOps the right way.

DevOps teams release deliverables more often with higher quality code than siloed teams. Also, automation and collaboration reduce complexity and mistakes. This benefit improves the Mean Time to Recovery (MTTR). Finally, provisioning through DevOps can automatically provide additional IT infrastructure capacity exactly when it’s needed.

Transparency and communication lead to less downtime and faster problem resolution. Teams enjoy increases in productivity, efficiency, and customer satisfaction. A mix of greater collaboration and trust leads to better working environments and more engaged employees.

Automated SDLC

Software Development Life Cycle (SDLC) automation removes many manual processes from the software delivery lifecycle. The goal is to produce high-quality software within budget and time constraints.

Processes

While teams follow the same SDLC process, their approach to each phase can vary. SDLC systems include Waterfall, Agile, Lean, Site Reliability Engineering (SRE), and DevOps.

Traditional SDLC

Traditional SDLC is also known as the waterfall model. Most teams regard the Waterfall model as the oldest SDLC approach. Teams finish one phase, move on to the next, and don’t look back. Waterfall is a simple model, but it doesn’t bode well for flexibility or ongoing projects. Teams can’t fix issues until the maintenance stage.

The phases of traditional SDLC are:

  • Planning
  • Analysis
  • Design
  • Development
  • Testing
  • Implementation
  • Maintenance

Agile SDLC

The Agile process differs from the traditional model, in that it is focused on collaboration and iteration. Agile project management can be used in many contexts, including software development.

Agile processes involve a continuous release cycle, where teams make and test gradual changes after each release. This approach allows agile teams to find and fix minor problems before they become significant issues. Also, Agile development encourages stakeholders to offer feedback throughout development.

The phases of Agile SDLC are:

  • Backlog refinement
  • Continuous prioritisation
  • Sprint planning
  • Sprint (including design, development, testing, and deployment) - usually over 2 weeks
  • End of Sprint show and tell
  • End of Sprint retrospective
  • Customer feedback
  • Repeat

Lean

Agile and Lean principles are both focused on increasing speed and efficiency. However, while Agile focuses on collaboration and sprints, Lean focuses on efficient work processes.

The Lean method includes seven principles:

  • Waste elimination: Getting rid of unnecessary tasks
  • Amplified learning: Trying ideas and getting feedback
  • Late decisions: Gathering facts before making choices
  • Fast software delivery: Producing quickly to receive input and improve
  • Team empowerment: Trusting workers to get the job done
  • Built-in integrity: Sharing information between end-users and teams
  • Big picture view: Optimising the whole by catching defects along the way

Teams following the Lean approach drop any activity that doesn’t increase customer satisfaction.

Site Reliability Engineering (SRE)

SRE is complementary to DevOps in that both practices are focused on collaboration, continuous change, and improvement. The main goal of SRE is to improve the user experience by achieving service-level objectives (SLOs) and service-level indicators (SLIs).

SLOs are measures of success based on how often the system is available and able to perform its function, and correspondingly, SLIs measure success by indicating whether the system is behaving as expected. Together, SLOs and SLIs tell you whether the system is available and doing what it is supposed to do. Accordingly, SRE developers are focused on operations challenges, whereas DevOps has operational engineers focused on development challenges.

DevOps

The DevOps model is the most recent SDLC methodology. This approach combines the Agile and Lean principles and includes small but frequent updates. Teams focus on automating manual processes, gathering feedback, and improving the entire process.

What are the Benefits of SDLC Automation?

There are many benefits to SDLC automation. It reduces human errors, increases productivity, and creates transparency. In addition, it improves communication gaps, eliminates bottlenecks, and introduces agile components through standardised processes. And importantly, SDLC automation ensures consistency, scalability, speed and flexibility.

Consistency

SDLC automation helps identify behavioural issues and errors in application development. The result is predictable and consistent because less human interaction equals fewer mistakes.

Scalability

It’s easier to scale automated processes by creating more processes to achieve the heightened requirements. The only constraint is software and hardware availability, which is not a problem when dealing with the cloud.

Speed

An automated process takes place whether there’s time available or team members are ready to execute the task. So the ability to speed through application lifecycle stages enhances project deliverability.

Flexibility

SDLC automation offers flexibility in scope and functionality. Usually, the only constraint is configuration management, which teams can adjust to suit the requirements. It’s easier to reconfigure an automated process than train a team member.

CI/CD Pipeline

The CI/CD pipeline uses automation to build, test, and deploy code. Practices include Continuous Integration, Continuous Delivery, and Continuous Deployment.

Teams follow a consistent set of practices to build, package, and test software with Continuous Integration. Then, Continuous Delivery directs validated code changes teams made in Continuous Integration to code repositories or select environments. Finally, Continuous Deployment sends code changes to end-users after several tests.

Automating certain phases of this pipeline leads to higher-quality software that’s faster and more reliable.

CI/CD Phases

Although the CI/CD steps can vary between organisations, the main phases of the CI/CD pipeline are building, testing, and delivering or deploying.

Build

Building is the Continuous Integration phase where teams create and compile code. They work together to build on source code, integrate new code, and pinpoint issues.

Test

Testing happens during Continuous Delivery and Continuous Deployment. Teams use automated tests such as performance testing, unit tests, integration tests, and functional tests.

Deliver or Deploy

Then, the software is ready to be delivered. Delivery involves sending an approved codebase to a production environment. Ideally, teams will automate this phase in Continuous Deployment and (after approval) Continuous Delivery.

Deployment occurs in Continuous Deployment. This automated phase puts the final product into production. During Continuous Delivery, teams send products or code to repositories then move them to production or deployment following approval.

CI/CD Principles

To ensure a safe, quality product that meets and exceeds customer expectations, teams must adhere to a set of shared CI/CD principles:

  • Create reliable and repeatable processes
  • Automate wherever possible
  • Create version control with a shared codebase
  • Embed testing and quality control in every step
  • Focus first on the most challenging parts and avoid procrastination
  • Every team member accepts responsibility and accountability
  • Define “done” as released to the customer
  • Go beyond CI/CD to achieve Continuous Improvement

Create Reliable and Repeatable Processes

Turning manual playbooks into automated software scripts ensures tasks are repeatable. This procedure allows other team members to execute a software script faster than going through the manual process again.

Automate Wherever Possible

Automating manual processes guarantees that they are reliable and repeatable. So teams can spend more time on creativity by committing processes to code with the ability to execute them on demand.

Create Version Control With a Shared Codebase

Teams working on a shared codebase can revert to previous release candidates with version control. They can also track edits on scripts, configuration, databases, and documentation with a version control system such as Git. Version control helps integrate changes made on independent branches of the code.

Embed Testing and Quality Control in Every Step

Teams use the central feedback loop to re-evaluate quality for the end-user in every phase of the CI/CD pipeline. They deliver new features through automated testing to verify new code meets expectations and is bug-free.

Focus First on the Most Challenging Parts, Experiment, and Avoid Procrastination

Teams address arduous tasks as they come up to avoid compounding issues‌. Avoiding procrastination helps find areas of improvement early and prevents unnecessary struggle later. Experimenting during the most challenging parts is a necessity to keep the project from getting stuck.

Team Members Accept Responsibility and Accountability

All team members focus on delivering the highest quality product possible to the end-user. The organisation motivates product managers, security, quality control, and developers to meet this goal.

Define “Done” as Released to the Customer

The customer must receive viable software. Work is not complete until the product satisfies the end-user. Even if individual team members finish their contributions, the project isn’t complete until the customer can use the product to fulfil a transaction.

DevOps Metrics that Go Beyond CI/CD to Achieve Continuous Improvement

Once deployment occurs, automated tools track and measure effectiveness for the end-user. Teams analyse DevOps metrics—such as business revenue, conversion rates, and engagement time—to reveal the end-user experience.

Benefits of CI/CD

Automating the software release process is the most critical benefit of CI/CD. But development teams receive even more benefits from this process.

Quicker Deployment

Teams use automation in Continuous Delivery and Continuous Development to speed up the development lifecycle.

Reduced Costs

Thanks to automation, teams spend less time on development, testing, and production, which results in less cost.

Continuous Feedback

The continuous build, test, and deploy cycle allows teams to take immediate action on feedback and improve code.

Errors Addressed Earlier

Automated testing for each code version makes it easier to discover and fix issues earlier in the pipeline.

Improved Collaboration

All team members can change code, address feedback, and handle issues as they happen.

Next Steps: DevOps Assessment

To get DevOps rolling, start small. First, identify an app or service a small group of people can experiment with. Then you can consider incorporating the whole mindset shift in your entire organisation.

A DevOps assessment can help you figure out your state of DevOps, where to go next, and how to get there.

Catapult CX works with DevOps Institute to conduct an Assessment of DevOps Capabilities (ADOC) for our clients. Book a meeting today to get started.

If you would like to learn more about how we can improve your digital service delivery, visit this page

Or, if you want to check out how we can upgrade your legacy system, visit this page