Bridging the Great Dev/Ops Divide
In today's hyper-competitive landscape, software consumers require their development requests to be responded to quickly, while providers want to demonstrate momentum and serve their markets with frequent software releases. These pressures demand software providers constantly offer new features and functionality. Older approaches to software development are being replaced. Enter DevOps.
Jun 30, 2011 5:00 AM PT
Under the legacy approach to software development, developers write code, which is then frozen, tested by another group, released and ultimately supported by yet another team again. Under this highly structured approach, large enterprise software applications are typically updated every 6 to 12 months, and up to several years can pass between major operating system releases. In enterprises, risk appetite and the desire for stability typically drive these intervals.
This "silo" approach to software development practices (that has been used by development teams since the 1970s) is clearly out of synch with today's real-time-focused business world. In today's hyper-competitive landscape, internal and external software consumers and customers require their development requests to be responded to quickly, while providers want to demonstrate momentum and serve their markets with frequent software releases. These pressures demand software providers constantly offer new features and functionality, and as a result, older approaches to software development are being replaced with a development cycle based on continuous improvement and deployment.
While the advent of agile development has helped to bring software development to the market faster, it does not address the impact of faster cycles beyond the development organization; engineering, testing, security, operations and support teams are also deeply affected by a torrent of software releases. Many technology organizations are still plagued by developmental and operational inefficiencies, poor coordination, and a lack of clear accountability for delivery and management of products and services. How can software development, engineering and operational functions stay in synch?
DevOps aims to bridge this gap between development and operations and encourage better cooperation and inevitably provide agility in both environments.
Think of DevOps as a framework of ideas and principals to create more cooperation and coordination between development, testing, security and other operational groups. In a DevOps environment, developers and operational staff build relationships, processes and tools that allow them to better interact with each other, representing a new way of thinking between the people who make the software and the people tasked with ensuring its smooth operation.
DevOps breaks down the traditional silos between development and operations. This can include extending core tenets of Agile Development from application development into the realm of testing and production. In turn it's also important for operational teams to embed their production orientation into the application development lifecycle. They need to ensure their operational requirements are communicated clearly, widely and most importantly early to development.
Critical server functions -- including automation, monitoring, capacity planning and performance, backup and recovery, storage, security, networking and provisioning -- can all benefit from the DevOps model to enhance the nature and quality of interactions between development and operations teams.
Through the collective DevOps community, best practices have started to emerge and are shared, fostering large-scale improvement. It provides operational and development teams an opportunity to:
- Discuss which best practices are most effective
- Share the tools they tend to converge upon
- Apply these principles (such as version control, bug tracking, staging and production testing) across the entire software lifecycle.
To achieve DevOps fluidity between software development, testing, quality assurance and the operational teams, the environments used for each of these functions should be automated via the same means, and subject to the same processes. This allows software developers to understand the process of how an application is to be deployed, and the IT infrastructure that supports it. System administrators supporting operational functions also learn about the role of server administration in the development process. Open source tools like Puppet and Chef help bridge this divide, giving system administrators across the various functions a common platform for sharing knowledge.
Businesses expect technology organizations to deliver products and services in an efficient, cost-effective, reliable and timely manner. Failure to do so results in slow speed to market, lack of responsiveness to market requirements and reduced customer satisfaction. In many technology organizations, these failures are the result of process inefficiencies, poor coordination, and a lack of clear accountability across development, engineering and operations organizations.
The DevOps movement has created a rapidly emerging framework of ideas and principles designed to foster cooperation, learning and coordination between development and operational groups. In today's DevOps environment, developers and operations staff build relationships, processes and tools that allow them to better interact with each other, and ultimately better serve the customer.