Accelerating Cloud-Enabled Transformations With Data-Driven DevXOps
Industry experts now support the expanded role of DevOps to impact business operations through a DevXOps frame. At the same time, as enterprises scale out DevXOps organization-wide, it is difficult for them to derive actionable metrics. Organizations need on-demand metrics to derive actionable insights. Here, Sriram Krishnamachari, VP, cloud portfolio solutions, Mphasis, discusses how companies can accelerate cloud-enabled transformations with data-driven DevXOps.
Enterprises are rapidly transitioning and maturing to cloud-native business operations that ensure high-speed service deliveries and highly resilient systems and operations. To achieve these transformative outcomes, it is indeed critical that they evolve their core business operations on industrial cloud foundations built around deep automation factories, pipelines, assembly lines, service catalogs etc.
Several industry experts now support the expanded role of DevOps to impact business operations through a DevXOps frame by applying the successful principles of deep automation and DevOps to now include wider enterprise stakeholders from sales, marketing, finance, and even ecosystem partners.
As I see enterprises scale-out DevXOps organization-wide and move to the next level of maturity, it is also becoming increasingly difficult for them to derive actionable metrics that matter, such as feature velocity, risks in the project, quality of delivery, failure rates etc. at a systemic level. They quickly realize the need for having a data-driven DevXOps approach, enabled by transparent real-time visual dashboards, as being even more critical to scale operations seamlessly.
Several of our clients we consult with have expressed the lack of on-demand metrics, making it difficult for them to derive actionable insights to plan their next best actions and prioritize resources based on business value.
Before we look into the details of how to address the specific evolving challenges, I would like to touch upon a little bit on the foundational capability pivots that I have seen that accelerate your cloud-enabled transformation imperatives.
6 Key Foundational Capability Pivots Powering Cloud Adoption
In pursuit of cloud-enabled transformation, as enterprises scale their internal systems and operations from DevOps to DevXOps, there are some important transitions and choices they will need to make to ensure they are scaling meaningfully around solid foundations. I will attempt to describe a few critical ones and share why they are critical and the expected impact they make.
1. Embrace zero-touch ops
The role of DevOps assembly lines is emerging as the core capability to automate enterprise processes, especially those that involve collaboration on the tasks between multiple teams. Assembly lines implement a pipeline for each activity instead of ad hoc scripting each time.
Assembly lines as a capability are focused on automating and connecting activities performed by several teams, such as CI for developers, automated infrastructure provisioning and configuration management for operations, test automation for test-driven development teams, security patching for security operations, semantic versioning and approval gates for release managers, deployments for multiple environments. Pipelines automate this entire process. Further, enterprises are providing self-service capabilities for the development team to instantiate a pipeline with configuration in an automated way. They help DevOps teams better connect, ensure seamless continuous delivery and move towards a zero-touch ops model.
2. Adopt on-demand continuous delivery with serverless computing
DevOps teams, especially those already operating on the cloud, are typically already working with modular components programmed via pipelines. Specifically, moving these modular services to serverless can enable teams to eliminate some of the tasks related to pipeline management and have them focus on development and deployment.
This capability is especially valuable for teams that are just getting started or are looking to switch to DevOps practices enterprise-wide. As productivity increases, the serverless on-demand infra scales seamlessly with minimal to no effort on the part of the ops team and enables application development without the cost of physical hardware. The pay-as-you-go pricing makes this model cost-effective for many types of applications.
This model may not necessarily suit all types of application workloads but is an important capability foundation that you could leverage and design your operations around.
See More: Five Major Cloud Security Challenges Businesses Should Prepare for in 2022
3. Pivot on machine learning and AI to solve complex problems
Deep automation leveraging machine learning is fully automating the back-office systems with minimal human intervention. AI/ML, on the other hand, is solving deeper and more complex problems that humans cannot solve efficiently. With connected systems from edge to cloud, there is going to be a massive data deluge and instantiating a solid Data Ops/ML Ops foundation is going to play a key role. This capability by way of outputs should answer a critical question on “ What should be our next best app that we should plan and release” to the product owners and developers, mining insights from the data deluge.
For example, a large pharmaceutical system is figuring out a demand forecasting and supply chain model optimization pipeline for 25,000 units of SKUs. They are developing various ML models that crunch massive amounts of data. Once these models are trained, they help with process re-engineering and automation of tasks that could be prohibitively time-consuming for humans.
Git-repos have largely taken over application development source code management and are also taking over infra as-code projects. Users want AIOps integrated as well within the CI/CD pipelines and with the control tower observability because their critical feedback impacts the next iteration of the software being developed. Some vendors, such as GitLab, are automating the feedback from AIOps into the issue tracking system to improve the overall value chain effectiveness.
4. Shift left security with DevSecOps
Security is a very serious concern, and DevOps teams are increasingly shifting security to the left to mitigate potential security risks early in the development life cycle. A DevSecOps model applies security practices throughout the lifecycle.
Adopting DevSecOps brings in an integrated approach to reduce the overall risk by bringing in key capabilities such as data protection, continuous compliance factories, static and dynamic code analysis, threat modeling, vulnerability detection, etc. If we break this capability down further, there are a few key elements that they need to adhere to — integrate security to build and dev testing, integrate security testing into workflows to capture vulnerabilities etc., automate deployments, deploy infra as a code, remediate vulnerabilities quickly with virtual patching or even run-time technologies such as RASP (run-time application self-protection).
5. Lean on open-source leveraged distributions for hybrid multi-cloud platform strategy
Developers are increasingly moving to lean in on open source platforms and customizing them to meet enterprise-scale requirements and find it the best way to keep pace with the rapidly evolving technology landscape — one that avoids vendor lock-in. There is also a lot of support and sponsorships for open-source projects from tech giants Google, Netflix, Facebook, IBM, and VMware through the Cloud Foundry and Cloud Native Computing Foundations, which have now become de-facto standards for shaping up the underlying platform substrate for cloud-enabled transformations.
There is a huge ecosystem out there supporting the development of the hybrid multi-cloud management operations and platform development centered around K8s and the toolsets required for observability, cluster scaling, service mesh, the day 0, day 1, day 2 concerns, etc. that are getting addressed through open source development. However, particularly the transition from where you are to the target state needs a lot of expert interventions from strategic partners.
6. Tie in FINOPS from the get-go
As developers get empowered to design, deliver software and drive enterprise agility, the development and support models are changing such that they are equipped with the inputs to make faster decisions for rollouts or rollbacks while taking on the accountability for the total cost of running and supporting the app. This is a huge change from what it was pre-cloud, where it followed a pre-designed workflow with several stakeholders in the approval chain.
FinOps as a culture and practice is fast evolving around open standards and practices. I see it as a “codified way to decentralize decisions”. It is a way for teams to self-manage their opportunities targeting revenues and optimize costs with choices they are enabled with by the cloud service providers.
In the next generation of evolved FinOps, everyone takes ownership of their cloud usage supported by a central best-practices group, with cross-functional teams in IT, Finance, Product, etc., organizing to enable faster product delivery. At the same time, they gain more financial control and predictability for their service deliveries.
See More: Why Telecoms and Insurers Need To Launch Personal Cloud Solution?
The Gap and the Need for Data-Driven DevXOps
Given the continuously evolving, leaner but an order of magnitude more complex decentralized operating model and the expanded role of the DevXOps model to seamlessly smoothen the operations, I see a very clear need for our clients to become a lot more data-driven and to drive operations through on-demand metrics. There is a clear lack of advanced analytics for what they want to achieve and when they need to be able to then marshall the teams transparently and cohesively and manage decentralized and transformed development through common dashboards.
The value stream mapping that is often treated as a one-time activity for modernization can really be done on a real-time basis through continuous feedback loops. Becoming data-driven further supports continuous value streaming and offers the best path to scale your DevXOps, bringing stakeholder transparency. Tech-driven companies, as well as the open community, now recognize that data-driven insights power velocity, relevancy, and overall quality delivered, and the feedback loops are built right in. Having visibility into every aspect of the DevXOps lifecycle will make value stream management easier and more effective.
To be continued in part-2…
In part-2 of my blog, I will touch upon some solutions and approaches that deliver the desired outcomes that enterprises can think of to accelerate cloud adoption.
Did you find this article helpful? Tell us what you think on LinkedIn, Twitter, or Facebook. We’d be thrilled to hear from you.
