SDLC Series: NoOps Primer
If you haven't already heard the term "NoOps" as it pertains to enterprise software development and delivery you probably will soon. NoOps is an emerging movement that seeks to relieve a bottleneck created by traditional IT operations and on-premise application hosting by utilizing solutions rooted in automation and cloud-based infrastructure. At Polyrific, we have developed an outstanding NoOps solution called Catapult and we offer this article in hopes that it helps you better understand why Catapult is such a big deal.
From DevOps to NoOps
Perhaps the best way to begin understanding the NoOps movement is to first understand the DevOps movement. The term "DevOps" is an amalgamation of "Development" and "Operations" and refers to the interplay between software developers and IT operations during the process of deploying applications to the world. In every enterprise, it is necessary for these two departments to stay close to one another in order to best serve the needs of the business.
At most enterprises, responsibilities for developers generally include the following:
- Work with stakeholders to understand the needs of the business
- Distill those needs into requirements and specifications
- Develop applications that fulfill said requirements
By contrast, IT operations are generally responsible for interfacing with network hardware:
- Allocation & management of server resources
- Fault planning & monitoring
- Security & compliance
- Device management
Obviously, applications that are developed to suit the needs of the business have to be deployed somewhere so that they can be consumed, and this is where the interplay between the developers and IT operations managers comes in: they must work together to take the developer's work and deploy it to the world on their enterprises resources. This makes perfect sense if the picture were so simple but, as we will see in the next section, reality is a bit more complicated.
Agile & Continuous Deployment
In the early days of enterprise software solutions, very few enterprises created custom software solutions or applications of their own. However, as workplace environments have become more dynamic and reliant on smart hardware and software solutions, the demand for rapid release of custom software applications has grown dramatically. The Agile movement was largely a response to this exponential growth in application demand and it is founded on principles inspired by the Silicon Valley "fail fast & fail early" philosophy. Gone are the days of months of planning, tedious software architecture design, and software release schedules following a waterfall schedule into a deployment phase that is given equal weight by the IT operations team. Today's software development teams are expected to respond immediately to a seemingly never-ending stream of features and demands requested by the business.
Often, projects are started as bare-bones applications that are immediately thrust into production environments where they will be constantly updated and expanded upon as the business requirements evolve. This sounds great, but it presents a few challenges to software development and IT operations teams, especially with regards to quality of the end-user experience and application uptime. To counter this, the development and ops teams employ a set of automation tools and checkpoints, collectively referred to as "Continuous Integration" or "Continuous Deployment" that smooth out the problems caused by rapid iterations in the software development life cycle. For example, when properly configured, and CI pipeline can trigger a series of automated tests whenever a developer checks in new code to ensure that the new code does not break anything or cause "regression" bugs.
The (Traditional) IT Bottleneck
It operations experts are fantastic but, in our view, their role is best executed when the evidence of their work is everywhere, but their presence is not so apparent. A good server at a restaurant will keep your glass full and your food coming without you noticing them much at all and it should be the same with IT operations managers but sometimes--often through no fault of their own--this is not the case. Without considerable depth of automation in your software development life cycle (SDLC), it becomes necessary for the development team to spend significantly more time with the IT operations team in order to coordinate downtime, deployments, rollbacks, and so forth. This is especially true in the case of on-premise deployments. This close coupling between IT ops folk and the developers is bad for at least three reasons:
- It takes the developer's focus away from understanding the needs of the business stakeholders
- It cuts into development time
- It can influence the engineering and delivery schedule of the application
Given the above, you can probably start to see where this is headed: interaction between development and IT operations should be automated so that the software engineers can remain focused on what they do best: delivering application-based solutions that serve the immediate needs of the business.
NoOps Produces Better Outcomes
So in order to respond to ever-changing demands of the business, development teams must be capable of quickly organizing the stakeholder's needs into business requirements and then parlaying those requirements into working code that is tested, quality assured, accepted by the end user and deployed into production environment on a frequent and recurring basis without being slowed down or distracted by hardware and deployment challenges on the IT ops side of things. Does this mean that IT operations professionals must be removed from the SDLC? Of course not. What it does mean is that IT operations personnel should join forces with the developers to implement game-changing solutions that help to automate the business of getting the developer's changes into production with very little interfacing required between development and operations.
In a NoOps world, developers don't check with IT operations before deploying code or to schedule downtime. In fact, they don't deploy code at all--they simply check their changes into source control and the rest happens automatically, behind-the-scenes, just like the server who always keeps your drink full without your noticing they were there at all. Similarly, developers do not need to request allocation of new resources from the IT department. They can, in theory, "spin up" a new ecosystem of server and database environments for a special purpose app while they sit with the stakeholder during a requirements gathering session.
The Catapult Digital Developer & NoOps Solution
As previously mentioned, we have developed a software solution called Catapult that takes automation of enterprise software delivery to the extreme. Using Catapult, even non-technical stakeholders can create new application projects on a meta-level that immediately spin up server resources using popular cloud platforms such as Azure and AWS. Catapult then allows entry of high-level data models in order to populate databases (or it can connect to existing ones) and generate, and deploy comprehensive codebases all without the user knowing how to write the simplest of SQL queries.
Like the restaurant server that deftly keeps your needs satisfied without making his or her presence known, Catapult allocates hardware resources, creates code bases, sets up source control repositories, allows stakeholders to manage content and seed test data, manages branching strategies, communicates with engineering team members to let them know of code changes, and pretty much anything else a competent developer and IT operations professional on your team would do. That is why we refer to Catapult as the "enterprise digital developer".
If you'd like to learn more about Catapult or any of our other software development solutions, please contact us or call us at 833-POLYRIFIC.