DevOps legitimized the friction that existed between Development and Operations and gained popularity for the right reasons. Improvements were already underway in pockets and the DevOps movement gave fuel to the larger Sysadmin transformation.
DevOps started with optimization of software development and release management area and has since then led to a gamut of innovations in other ‘Ops’ heavy areas such as Infrastructure and Service operations.
DevOps is about breaking siloed behaviours and simplifying delivery process for improved speed and efficiency. Teams that adopt the DevOps discipline have become high-performing, building better products faster and for greater customer satisfaction. DevOps has become synonym to a workStyle/culture change for teams and organisations. DevOps has many interpretations as there is no defined playbook or tool.
DevOps-ify in its true form is driving towards the following key goals — Collaboration. Agility. Consistency. Transparency.
Consultants and businesses have converted DevOps into a buzzword which has led to a proliferation of X-Ops. Popular implementations of DevOps are explained below —
- DevOps/ DevSecOps — Release engineering morphed into DevOps and is often referred to as CI/CD or Continuous delivery. It refers to a set of best practices that combines planning with software development, testing and release lifecycle to achieve continuous delivery of high quality software. With Security becoming a shared responsibilty and shifted to the source, DevOps became increasingly known as DevSecOps
- AppOps/ ProdOps / CloudOps — The application developer is also responsible for the application in production or relies on a separate Production Operations role to comply with SOD. This space includes operational side of application management, including release automation, remediation, monitoring and maintenance processes such as patching, access control, inventory mgmt, provisioning, backup. DevOps transformation involves stitching together activities with change management and incident management tools for continuous operations and visibility. With the advent of Cloud, responsibilities extend to multi-cloud operations. While cloud provides these capabilities, operating requires knowledge of cloud native toolset.
- SRE — Site Reliability Engineering was coined by Google in 2003 and is focused on the stability and reliability of Production Infrastructure. While DevOps focused on the what, SRE provided a prescriptive guidance on the how. SRE deals with ProdOps more objectively with SLIs and SLOs and error threshold with suitable consequences. SRE provides teams with a better framework for observability, negotiation and decision making. SRE shows great promise for organizations with dedicated ProdOps roles. SRE brings about the right mix of real user monitoring and metrics and combined with support closes the continuous feedback loop.
- InfraOps / GitOps — A relatively new mutation of DevOps that started in 2017. Previous approaches did not solve for the disconnect with live system. ‘GitOps’ is an operating model for the delivery and management of cloud native applications on Kubernetes. Based on Git repository and declarative syntax to define the desired state of production and an automated process to match production with the current state in the repository to continuously manage environments. GitOps operationalizes the Infrastructure as code and Configuration as code.
- ITOps / AIOps — Refers to EnterpriseIT modernization. The heart of the DevOps and ITSM culture clash comes down to ITSM being process-intensive while DevOps preaches collaboration, breaking silos and agility. A mixed approach brings in the strengths of both values. As it is increasingly important to be agile, ITOps is the lever to drive process and culture changes in the organisation. The Pheonix Project book by Gene Kim is a great reference novel that highlights the ITOps transformation story. AIOps is not ‘DevOps’ in the true sense as it refers to AI driven IT operations. AIOps replaces traditional IT Operations tools approach with rich data platforms that harvest events, tickets and other operational data sources to provide prediction, anomaly detection and auto-recovery capabilities.
To conclude; like Agile, DevOps is iterative. It does not stop here, the same principles are being applied to other areas to coin terms such as SvcOps, noOps, ChatOps, DataOps, MLOps, FinOps …
Note: DevOps is a broad term and doesn’t differentiate between different sysadmin sub-disciplines — “Ops” is a blanket term for system administrators, operations staff, release engineers, DBAs, network engineers, security professionals etc. “Dev” is used to refer to developers i.e all people involved in developing and designing the product” which can include Product, QA, designers etc.