By Bradley Clerkin
CTO, BreakFree Solutions
Everyday someone develops a new flavor or tries to coin a new phrase when it comes to DevOps. MLOps, SRE, AIOps, NetOps, CloudOps etc., are all essentially DevOps. DevOps is the reimagining and optimization of IT management processes and supporting technologies for use in digital transformation and cloud. We optimize these processes even further for each cloud and digital use-case, but each flavor has more in-common then not.
In general, we see the following commonalities (this isn’t an exhaustive list, just the highlights):
Use automation and DevOps software to minimize any value delivery slowdowns or bottlenecks caused by processes or technology.
Build custom software that continually increases the effectiveness and simplicity of technology use.
Select best-in-class technology solutions that are optimized for the specific use-case.
Maximize the use of cloud capabilities and design paradigms.
Ensure developers and engineers build the features and capabilities as close to production as possible.
When I think about this, I picture Fred Armisen’s Venezuelan Parks Director character from the TV show Parks and Rec listing off how everything leads “straight to jail.” In the world of cloud and digital, everything is indeed “straight to DevOps.” MLOps is DevOps, NetOps is DevOps, SRE is, believe it or not, DevOps. This should bring more clarity to IT leaders that are confused about all the different types of DevOps terminology and marketing.
However, I am always highly skeptical when I hear about organizations automating IT shared services through software like ServiceNow and claiming to be doing DevOps. In reality, components of DevOps need to be applied to IT shared service delivery. In the IT shared services use-case, new processes and technologies pioneered in DevOps are pragmatically applied to existing processes where there is a clear opportunity to deliver shared services more efficiently. That being said, let’s be clear: if you’re automating VMs, that doesn’t mean you are doing DevOps. Instead, you’re using a technique DevOps teams pioneered. Our customers tend to have the correct mindset, which means putting the majority of their DevOps efforts and funds toward cloud and digital.
Circling back to the many flavors of DevOps, here is a high-level breakdown of the DevOps categories we hear about most often:
DevOps for Development (DevOps): Automated continuous deployment of applications on cloud-native platforms.
DevOps for Databases (DataOps): Automated configuration and changes to stateful database technologies.
DevOps for Machine Learning & Data Science (MLOps): Optimized development systems and processes for the unique challenges of machine learning modeling and data science.
DevOps for Platform: Automated management and deployment of software that enables cloud-native development at scale on cloud platforms.
DevOps for Cloud (CloudOps): Automated operations of shared elements of cloud environments via automated guardrails and APIs.
DevSecOps: Restructuring security processes and systems to allow for effective, high-velocity cloud-native development activities.
Site Reliability Engineering: Cloud-native operational best practices established by Google for operating cloud-native applications at scale.
AIOps: This category currently and primarily focuses on bringing machine learning and automated incident response and handling to cloud-native operations. In the long term, processes may also be designed to support AI engineering.
Operationalizing DevOps capabilities is hard, but the terminology surrounding it doesn’t have to be. Yes, where we should be applying DevOps in cloud and digital seem to be infinitely expanding, but don’t worry. It’s just slightly different flavors of the same foundational approach as long as it’s being applied to support digital transformation and cloud efforts. We use this as a litmus test for our clients, and we hope our clients use this as a litmus test for their partners. You don’t need a bunch of partners for each type of DevOps. You only need one really good one that understands it’s all DevOps. That’s where BreakFree comes in, and why we are the leading DevOps professional services team in the Midwest. We know and help with all the flavors of DevOps.