Drive a race car on a reliable and resilient track
November 17, 2021
Every company is looking to deliver the best value for their customers. But in finding their unique competitive edge, they may be overlooking an important part of the equation. Scaling DevOps & Site Reliability Engineering (SRE) is the way to take Agile to the next level, and with it, the opportunity to accelerate business change.
What is DevOps and SRE? At Accenture, we talk about how application delivery has shifted from large-scale, project-based implementations to a continuous evolution. DevOps and SRE make this possible by bringing business, development, and operations teams together and applying automated processes to streamline IT. The ability to roll out improved business capabilities has become critical in today’s digital world.
DevOps has been around for over a decade now. But in our fast-paced, rapidly changing world, we need to extend its reach beyond the development teams to the enterprise—in other words, we need to Enterprise DevOps.
Let me share with you the four key phases for maturing your Enterprise DevOps and what is learned along the way. After all, we all live by the Agile adage: fail fast and fail often, because that is the true-and-tried path to continuous learning. The speed of Agile, with its constant iterations in IT development, requires ever higher levels of automation—and all the organizational adjustments that come along with that.
<<< Start >>>
<<< End >>>
Why do we need Enterprise DevOps?
Technology is making Agile—and Enterprise DevOps—transformation possible. But it is just one component. In my experience, DevOps teams are being held back from delivering their full value because they do not have the power of the enterprise behind them. It is not possible for teams to do everything bottom up. Company-wide commitment to scaling DevOps makes all the difference, by driving change from the top down.
Three main drivers are pushing companies to recognize they need to scale DevOps. The first is the extensive adoption of Agile practices. The second is that the business case for automation is now positive, with the more frequent execution deployment of test cycles. And the third is the focus on the cloud, with Enterprise DevOps as a key practice to realize the full value of cloud migration.
In short, what we need is not the old siloed DevOps. It has to be crucial to the mindset of everyone in the organization. It cannot be an add-on, but rather the focus for the entire company. When we talk about Enterprise DevOps and SRE, it is no longer about the tools, which already exist on the team level. Now, the stakes are higher—it is about people and processes and cultural change. It is about the organization and the structure of the teams, and how we not only adopt the Agile approach but also the mindset.
One more key point: Enterprise DevOps works best when leveraging Site Reliability Engineering, the practice of taking software engineering concepts and applying them to IT operations, which originated within Google.
I describe it this way to my clients. With Agile, you are building a fast car. With DevOps, you are building the highway (the pipeline). And SRE are the guardrails. Each part is working in sync, so that you can drive your car really fast down that highway, but safely, within the protective framework of the guardrails that allows for resilience and reliability.
<<< Start >>>
DevOps teams are being held back from delivering their full value because they don’t have the power of the enterprise behind them. Scaling DevOps makes all the difference, by driving the changes from the top down.
<<< End >>>
The four phases of Enterprise DevOps and SRE
There are four distinct phases of maturity to Enterprise DevOps:
Phase 1: The focus is only on the development teams. Continuous integration is part of building the minimum workable pipeline. The practitioners learn version control and to write unit tests, prepared to take over the responsibility for quality in the next step towards maturity. To be independent from other teams, these development teams need to align to the products. This business value alignment is a central aspect of Agile transformations.
Phase 2: Continuous delivery breaks down the wall between development and quality teams. To make it successful and balance the workload, some team members moved from the former quality team into the development team. Testing is automated, tests run in the pipeline with each deployment. Now the teams are responsible for writing new automated test cases and maintaining the existing ones.
Phase 3: Combine development and operations teams: You build it and you run it. For most organizations, this is a huge step, but our numbers prove that it pays off. Why? Reduced communication effort for knowledge transfer, deployment coordination and incident detection. Fewer incidents after changes because of the focus on quality and shared responsibility. And reduced time to restore a service. All of this is enabled by continuous deployment.
Phase 4: Continuous monitoring, with a focus on Site Reliability Engineering (SRE) practices. SRE provides practical techniques, including the service level indicator (SLI) and the service level objective (SLO). SLI/SLO makes the customer experience transparent for example by showing latency time and error rates. The user experience is captured in a real-time dashboard.
<<< Start >>>
<<< End >>>
Start small, but aim high
The best way to have a successful Enterprise DevOps and SRE transformation is to start with a pilot team to showcase positive DevOps and SRE behavior at a smaller scale as a proof of concept. During the pilot phase, we write a playbook that the entire organization can use to adopt the practice. This is complemented by a centralized enablement team which actively helps teams critical to the business. At the same time, the enablement teams centralizes and publishes all information to help other teams building up the capability without hand-on help. This central self-services reduces bottlenecks from the centralized enablement team and all teams can adopt the capabilities independent.
In changing from a siloed organization to empowered teams, leadership must agree on changing the power structure in the organization. Team leads may feel that they have lost control over team members. But this is not the case. It simply refocuses everyone on their main priority—the success of their product.
By scaling DevOps and SRE, we can free up the speed of our technology teams, enabling companies to build an empire. Leading companies have shown that scaling DevOps and SRE on the enterprise level increases the efficiency and quality of their products. Is it your turn next?
