There’s an old saying in DevOps: if your DevOps capability toolset doesn’t include quantifiable tasks metrics and call stack analysis technology, then it’s probably spin, puff and fluff. Actually, they don’t say that at all, but they should.
DevOps is one of the ‘new’ darlings of the IT world, except that it isn’t. A coming together portmanteau of Dev for developer and Ops for operations, it is meant to express a new notion of orchestrated workflow unification between these two sometimes seemingly incongruent technology functions. Like cloud computing, it is of course not that new at all… and major vendors has been offering tools in this space for way more than a decade. It’s just fashionable to give it a new name and talk about it nowadays.
The ‘problem’ with DevOps
The issue here, if there is one, is that orchestration and automation technologies are very much needed in the world of big data analysis, cloud and Agile rapid application development. Because of this, DevOps fits really nicely and we see every vendor from the cloud providers to the security specialists suddenly claiming to sport DevOps capabilities. Except they don’t, not in the thoroughbred sense anyway.
At the recent
TechTarget defines DevOps as follows: “DevOps is a philosophy or cultural approach that promotes better communication between the two teams as more elements of operations become programmable. In its most narrow interpretation, DevOp is a job description for an employee who possesses the skills to work as a both a developer and a systems engineer."
What’s inside ‘real’ DevOps, really?
Ed Blankenship is Microsoft product manager for Visual Studio Online, ALM and DevOps. He used his appearance at Build to explain that DevOps is a specialized discipline and practice -- such that there is now the option to create Visual Studio ‘extensions’ so that users can bake in their own specific DevOps-related functionality and abilities into the Microsoft software they use.
Once again, ‘real’ DevOps should incorporate granular testing tools such as new IntelliTest function in Visual Studio 2015 that works to test individual code methods and help discover the minimal amount of tests needed for any piece of code. Real DevOps has automated testing, it has peer review tracking and it has multiple code-commit merge conflict management.
Don’t say software release, say a ‘cloud-based sprint cadence’
Jumping forward from last month’s Build event, this month sees Microsoft introduce several DevOps resources to help development teams identify opportunities and avoid pitfalls. Microsoft’s developer division explains DevOps in the context of its own experiences moving from a boxed release cycle of every couple years (for Team Foundation Server) to today what is known as a ‘cloud-based sprint cadence’ of every three weeks (for Visual Studio Online).
So, according to a Microsoft spokesperson, inside ‘real’ real DevOps we find elements such as -- performance monitoring, usage analysis, a feedback loop from production all the way through to the developers, release management and continuous deployment capabilities, agile planning and workflow support, cross-team collaboration (including test), even end-user input directly integrated with the toolchain.
The company has now just launched a new DevOps e-book by Sam Guckenheimer, "Our Journey to Cloud Cadence, Lessons Learned at Microsoft Developer Division," which is now available for free download (PDF here). In this book, Microsoft discusses lessons learned as the Developer Division transitioned from a 'box' software team, delivering on-premises software releases on a multi-year cadence, to becoming a SaaS provider continuously delivering to the public cloud. It covers the DevOps engineering practices, the tools and culture that the organization needed to evolve as it transformed to the second decade of Agile.
Microsoft has also just released aDevOps online self-assessment tool to determine where teams need to focus their DevOps attention and next logical steps to take in their journey to adopting DevOps practices. The self-assessment tool is designed to assist teams to gauge their proficiency in seven practice areas that the company has identified as critical to DevOps excellence.
There’s also Application Insights - an application performance monitoring (“APM”) tool that claims to make it easy for iOS, Android, Windows and server-side app developers to capture and visualize crash and deployment data, usage analytics and other telemetry via a robust, cloud-backed app health dashboard. Announced at Build and currently available in preview, Microsoft Application Insights provides development teams 'on-ramp' to DevOps, as they say.
DevOps done right can (potentially) lead to an infinitely virtuous loop where teams plan, code, build, test, release, deploy, operate and then monitor – and then plan again, and so on.
Wary of DevOps-washing?
This story is meant to do one thing i.e. highlight the real guts and mechanics inside DevOps. During Microsoft Build I asked a very senior developer division manager if he got angry about the rest of the industry trying to ply its wares by DevOps-washing without real DevOps tools? He was too polite to say so and that confidence probably comes out of the fact that the firm’s tools mostly really do what it says on the tin – well, mostly, anyway.
