In every industry, business cycles, driven by digital disruption, are accelerating. Now more than ever, IT service delivery cannot become a bottleneck. IT processes need to become faster and more agile along with the rest of the business cycle. More and more, standard tasks and processes are executed by increasingly autonomic systems that are data-driven, augmented by artificial intelligence and automated. This enables human work to be reorganized so that it becomes more agile and efficient. As the boundaries between infrastructure and application services are beginning to blur, I believe we should see the rapid emergence of DevOps practices within service delivery processes.
DevOps was first practiced in "born-on-the-cloud" organizations, leading to the integration and even merging of development and operations teams. In recent years, it has been adopted successfully in ‘traditional’ enterprises, IT service delivery organizations and hybrid cloud environments.
In 2014, IBM defined DevOps with a business outcome perspective as an "essential enterprise capability for continuous delivery of software-driven innovation that enables organizations to seize market opportunities and reduce time-to-customer feedback."
DevOps practices are designed to dramatically improve both the efficiency and effectiveness of the delivery process to increase speed and lower cost and risk. Companies like Nationwide (a former customer of mine) have seen a 50% improvement in code quality and a 70% decrease in end-user downtime, with 58% of their teams in the top productivity quartile of the industry.
DevOps can bridge the cultural divide between application development teams and IT operations teams by fusing the lean approaches adopted by operations with the agile practices favored by development teams. Then, these processes can be extended across the entire software and service delivery process end-to-end, from client requirements to development and operations. From there, a continuous delivery feedback loop between the IT and development teams can be created that gives both teams a shared view of outcomes.
To ensure that a solution delights the target audience, DevOps is increasingly blended with Design Thinking, a disciplined iterative design approach that focuses on establishing a small number of design goals with verification through sponsor users, which can help you stay in touch with what end users will want out of the product.
Continuous Delivery Pipelines
For efficient delivery, DevOps draws heavily on the lean concept of a continuous delivery pipeline. This requires removal of waste and bottlenecks at every step of the development process to achieve faster throughput. Process improvement methods like Value Stream Mapping are an integral part of the DevOps toolkit, as are the elimination of hand-offs, the automation of tasks and the reduction of work in progress.
Feedback Loops
For effective steering of the delivery process, DevOps draws on the Design Thinking and the agile concept of continuous feedback loops, with honest measurements and a focus on quality and operational stability. The principle of "working software" at every iteration is extended to "working software in a production (or production-like) environment." Design Thinking also adds structured end-user feedback.
Continuous Integration And Continuous Testing
Moving tests, especially integration tests, to the front of the development process will help reduce risk and lead to the concept of continuous integration and continuous testing. A high degree of test automation is essential for successful DevOps implementations, in particular for maintaining quality while significantly reducing delivery cycle times.
Measurements
Measurements are key to DevOps for ensuring continuous improvement. Applying analytics and artificial intelligence allows for fine-tuning and better insights into success patterns, especially when you combine them with extensive use of application performance monitoring (APM) and sophisticated event management.
What DevOps Can Do
DevOps could have implications for a business's organization, culture, process, architecture and tools. A study by Forrester based on a survey of 600 IT professionals identified the most important factors for satisfying business stakeholders. The study found that keeping teams focused on one project at a time, allowing small batches to flow through the software delivery cycle and creating an application architecture based on loose coupling and application programming interfaces (APIs) were the most important critical steps in driving rapid delivery cycles.
All these elements need to come together for a successful transformation of IT service delivery processes and organizations. Eliminating routine tasks through process streamlining and automation allows practitioners to focus on higher value tasks like resolving complex incidents, requests or exceptions, or developing automation capabilities. For these tasks, teams need to be structured differently so they can collaborate quickly and comprehensively. The agile squads/tribes/guilds model can be successfully adapted to delivery centers, together with the principles of small batches and reduction of "work in progress" that are essential to optimizing throughput in a lean approach.
DevOps is complemented by an "IT as a Service" approach, which allows rapid deployment and flexible consumption. Only an "as a Service" model with APIs as the defining interface of a service allows frequent updates that become instantly visible to consumers without any disruption. This is especially important for services provided by partners within a services supply chain. As the portfolio of standard services offerings and solution building blocks becomes richer, composition and orchestration of existing services will become the norm, as opposed to building services from scratch.
Blending DevOps with Design Thinking enables deep and systematic client engagements, leading to solutions that match clients’ requirements. Drawing on curated knowledge and composable building blocks enables differentiated solutions that are tailored with a common architecture from standard components, with faster time to value and more intensive co-creation of unique elements.
The principles and practices of DevOps are essential for service delivery transformation, both for enterprises and IT service providers, providing a perfect complement to automation and "as-a-service" delivery models.
