The set of concepts we have been describing is collectively termed DevOps and is a part of a larger process called continuous delivery. DevOps is a shortened form of development operations and describes a close working relationship between the development of software and the deployment and operation use of that software. Continuous delivery is a set of practices that enables software to be developed and continuously deployed to production systems on a frequent basis, usually in an automatic fashion that happens multiple times a week or day.

Each year, a company called Puppet Labs surveys over 4,000 IT operations professionals and developers about their operations procedures. Of those surveyed, companies that have implemented DevOps practices have reported improved software deployment quality and more frequent software releases. Their report states that these companies shipped code 30 times faster and completed those deployments 8,000 times faster than their peers. They had 50% fewer failures and restored service 12 times faster than their peers.

Results such as the ones shown in the Puppet Labs survey show that organizations adopting DevOps are up to five times more likely to be high performing than those who have not. It's a cumulative effect; the longer you practice, the greater the results from adoption and the easier it is to continue doing that. How DevOps enables high performance is something that centers around deployment frequency.

To define and explain the entirety of DevOps and, since continuous delivery is out of the scope of this book, for the purposes here the goals can be summarized as follows: to improve the deployment frequency, lower the failure rate of new releases, and shorten the recovery time if a new release is faulty. Even though the term implies strict developer and operations roles as the only ones involved, the concept really applies to any person or department involved in the developing, deployment, and maintenance of the product and the servers it runs on.

These goals work toward one end: minimizing the risk of software deployment by making changes safe through automation. The root cause of poor quality is variation, whether that is in the system, software settings, or in the processes performing actions on the system or software. The solution to variation is repeatability. By figuring out how to perform an action repeatedly, you have removed the variation from the process and can continually make small changes to the process without causing unforeseen problems.