X-OPS a CTO perspective:
DevOps; DevSecOps; GitOps; MLOps; DataOps; AIOps; DevSecComplianceOps;… FinOps… Ops in IT has never been so popular. Ops was always a villain for users, developers, security, finance and anyone associated with enterprise IT. But now we can’t get enough of it. There are 3 key takeaways here.
1. Rapid incremental and stepwise, reversible change supports stability, agility and innovation.
2. Feedback from Production Users continuously re prioritizes the backlog in a more meaningful way.
3. Business agility and clear Return on Investment ( ROI) result from all engineering functions embracing speed.
It does not matter whether you are an application development engineer, data engineer, data steward, data scientist, site reliability engineer, cybersecurity engineer, infrastructure engineer or operations engineer; there is a discipline and toolset for you in today’s world.
Agile Methods, Cloud and DevOps are correlated in their adoption maturity and I would never want to make them synonymous. However, for the purposes of this article, I make the assumption that the correlated developments are one smooth adoption curve.
Simultaneously, the terms DataOps in the area of data preparation and cleansing for analytics and MLOps in the area of machine learning and AI development have taken hold with their respective development constituents. These are not simply DevOps for the respective domains but actually have meaningful differences, toolsets and practices added because of their domain. AIOps now signifies the use of AI in IT operations. DevSecOps stands for the shift left of security occurring earlier in the DevOps cycle.
DevSecComplianceOps stands for adding compliance evidence gathering being integrated into the DevOps cycle. Last but not least, I have seen references to FinOps… the business of integrating budget management more tightly into the development process for a cloud consumptive model to be well managed. I have consciously avoided other X-Ops candidates in order to reduce reader fatigue. The Table below gives you a set of definitions as a reference. It also shows you the roles in the IT organization that bring together the practices in each domain.
In my discussions with CTOs, I have found it useful to focus on the outcomes these X-Ops domains bring to the business. The focus on the outcomes rather than the technologies provides a simple way to measure the IT organization’s progress on the Journey to Cloud and Journey to AI.
When you define a cloud journey, various practices in the X-ops need to be adopted by the corresponding domain so that the full benefit of the transformation is realized by the customer. For example, without adopting MLOps practices while AI models are built and deployed, models degrade. Model performance doesn’t just depend on model construction, but also on the data it relies on, algorithms chosen, tuning, regular updates and retraining.
In the next article, I will compare the recommended metrics for maturity in DevOps , DataOps and MLOps.
