DevOps according to Atos
- new business models
- disruptive new parties such as Lean start-ups who are able to rapidly deliver new products to market
- a shift to different and / or new channels
- customers (consumers & citizens) demand for online services
- myriad of channels with greater need for differentiation
- increased pace of changing customer requirements
- the increasing importance of brand value & loyalty
- organizations are adapting to the need to innovate in order to survive
IT departments must change:
- in order to tackle the often rigid legacy of existing IT landscapes
- to support the business and respond to changing needs
- evolving from being a supplier into an innovative partner
- to facilitate more intense and new ways of cooperation between organizations and customers.
The challenge facing an IT organisation ‘of the past’ is to find a way to respond to the above changes ‘for the future’. The answer has quite recently been sought in Agile / SCRUM methodologies that enable Continuous Integration with the promise of faster projects, more value for money and involvement of the business through a Product Owner. It has proved a step in the right direction; but not quite enough. You could say that DevOps finishes what Agile has started. Now ‘DevOps’ is the powerful answer that takes Agile a step further by enabling Continuous Delivery with faster Time-to-Market, lower Total-Cost-of-Ownership, Automated Provisioning, Automated Testing and Continuous Deployment that significantly reduces feedback loops and mitigates risks.
DevOps breaks down the traditional silos with multidisciplinary teams providing a continuous flow of new innovative features to the customer; eliminating waste through maximum automation (e.g. reducing manual hand-overs); and dramatically decreasing lead time from “Go” to “Show” by using unified ‘product teams’ instead of separate ‘project & maintenance teams’. The extension of Continuous Integration is seen clearly in a far higher level of automation in respect of both QA and the deployment, whereby every commitment flows into a production release.
Testing forms an integral piece in an organisations’ ability to deliver the full potential of DevOps. Automated Deployment requires a different release strategy with Smoke tests and automated rollbacks; automated data migration testing and database upgrade testing become repetitive; rigor in the build (unit) testing is essential, while an industrialised approach to requirement testing, automated functional testing and robust non-functional testing (notably performance, security, localisation, accessibility, production readiness and user-experience) is witnessed. In fact, it doesn’t stop there: deployment is automated, with the requisite automated testing and checks, but also once in production, performance and security monitoring is also required.
The above addresses two of the three principal drivers out of ‘People, Process and Technology’. Process is important as ever, with Agile / SCRUM / Kanban / Lean; and Technology is even more important than ever before with tools and solutions that enable the higher level of automation (Testing, Provisioning, Deployment and Monitoring); but it’s the aspect of people that really makes the difference in DevOps. Organisations need to accept a culture change, or at the very least allow an ‘incubator’ pilot a DevOps approach with a full mandate to build a bridge / break down walls between the Development and Operations organisations. It takes different skills to excel in a product team as opposed to a project team or maintenance team and the people in these product teams need to be given the freedom to deliver excellence. An investment well worth making for all concerned.