Evolve your operating model to succeed at digital transformation
While organizations embrace a cloud strategy for many reasons, one that stands out is a desire to get IT operations working more efficiently. Companies want to streamline value chains. They want to do more work and do it faster, with less friction.
Cloud-based operating models, done the right way, enable just that. In fact, for many organizations we work with, it is the adoption of the new ways of working that public cloud demands that creates the greatest benefits to business agility, and not merely the technology platform itself.
With this in mind, one of the top initiatives our customers prioritize is the development of cloud operating practices across their IT organizations and their entire IT ecosystem. Ironically, however, the operations domain is also one of the areas in which organizations struggle the most to generate momentum.
Breaking with the past
The biggest challenge organizations face is the need to shed a troublesome piece of baggage: their legacy operating models. Too many companies try to adopt cloud platforms without changing the way they work. They assume they can follow the same procedures they always have and make simple tweaks.
What they end up with is a collection of bad habits, tribal knowledge, and sets of processes, procedures, and tools that don't respond well to the demands of a modern ecosystem. And, in turn, they fail to leverage operational learnings across their operations, missing opportunities to deliver agility improvements across the entire organization.
Organizations that are born in the cloud sometimes avoid developing these bad habits. They tailor their operating models to the way business is done in the cloud. They set up their operations to other organizational domains, incorporating DevOps, enabling innovation, sustaining applications, and driving strategy and governance.
But companies that have lengthy histories and may not have the luxury of spinning up new models from scratch have work to do to get their operations functioning at such a level. They face steep learning curves, including getting comfortable with new tooling, shared-responsibility models and practices, and job functions.
They have to learn how complex cloud transformations are so they can equip their teams with the resources needed to carry out the necessary work. Rather than relegate the work of getting to a cloud operating model to a side-of-desk task, they need to dedicate a group to adopting the new cloud-everywhere operating model.
Where things stand
Based on our engagements with customers, we've evaluated enterprise progress in the eight domains making up the HPE Edge to Cloud Adoption Framework .
Figure 1: The Operations domain in the HPE Edge to Cloud Adoption Framework
Operations is a domain that most customers prioritize. It's a tangible and significant part of most IT organizations, and one where progress is being made overall. Operations is one of the domains where we see the greatest overall progress in capability across our engagements, with an average maturity of 2.1 on a scale of 1 to 5, where a score of 3 indicates a cloud-ready organization.
However, there are some subdomains within operations where organizations are experiencing greater difficulty in achieving traction (see Figure 2):
- Service operations
- Platform operations
- Pipeline operations
Figure 2: Organizational maturity in the Operations domain
Let's take a look at what's behind these challenges and what can be done to overcome them.
Service operations: Slow, ineffective incident response
Many of the organizations we engage with have struggled to progress incident response capability to the point where it can work at the velocity of cloud. Though incident response is considered a security function, the operational component is an important measure of a company's overall success in the cloud.
Companies with immature service operations often fall short in their ability to be proactive and prevent events from happening in the first place. And if a security event takes place, they will often exhibit slow, ineffective incident responses. The time it takes to detect, address, and resolve an issue often places them outside of their service-level agreement (SLA) targets due to a lack of automation and orchestration.
To be proactive, companies need to set up systems to leverage metadata and characteristics from historical events. They start to learn from those events and build out proactive incident responses. Using techniques like comprehensive logging, event curation and correlation, and forensics can determine root causes and prevent issues.
The flip side is when an event has taken place, organizations take a long time to respond. Without automation and orchestration, you can't scale the volume of elements that need to be remediated during an incident. If you've automated their response, it gives you a relatively short window for remediation.
Platforms operations: Lack of clearly defined infrastructure patterns to deliver consistent services
Organizations with a low level of cloud operational maturity lack a set of finite and clearly defined standards for developing applications and corresponding infrastructure. Without these patterns, they struggle to master the art of automating the provisioning of infrastructure and supporting the corresponding services that make applications tick.
This ties back to the legacy hangover issue. Organizations that are used to creating applications outside of cloud-native environments by default model their application architectures to support customized apps that may or may not generate high degrees of value. These architectures can't scale to satisfy the emerging demands of the business.
As a result, operators find themselves having to provide white-glove service and custom work on an ongoing basis for customized applications. If they started from scratch, they'd be ready to scale. They could focus on those application archetypes that have a certain amount of critical mass. What starts to happen is organizations learn how to establish enterprise standards and policy and leverage cloud-native tools and that newfound operating model. It's easier to bring others on board because they have that consistency and they generate results.
The new operating model creates its own gravity and starts to attract the attention of application and business owners. They can provide a valuable service as long as organizations play by a set of standards and rules to support scale. Otherwise, organizations are translating bad habits into the new cloud model.
Pipeline operations: Inconsistent management of container Images
Containers have changed the way organizations develop and deploy applications. Their lightweight structures and portable natures make them ideal packaging tools for organizations looking to add new features quickly and cost effectively. Still, a container operation is only as sound as the images that make up the containers themselves.
Organizations that are early in their cloud transformations tend to struggle with image integrity. They haven't set up systems for hardening container images in a timely and repeatable fashion. They haven't set up an automated testing and compliance certification process. They haven't created secure container registries that identify images for use in the continuous integration and continuous delivery (CI/CD) pipeline, retire out-of-state images, and manage artifacts in various states of transit.
Pockets of skill
Some organizations see their cloud operations mature in a scattershot manner. In other words, it's common to see cloud operations showing up as pockets of individual skills concentrated among certain people and certain groups. This can be helpful. Companies that promote a learning agenda can create first movers that act as a force multiplier. They can become change agents, with an eye toward centralization and supporting federation when the enterprise is ready.
Please read: How to achieve extreme quality with DevOps
But it isn't always a good thing. If IT operations don't prioritize the optimization of a comprehensive cloud operating model, some application owners will head to the cloud by themselves. The way these "experts" support their own applications won't support all of the other application archetypes and the larger ecosystem out there. Costs spike, and the C-suite gets frustrated with the lack of progress. This is a big reason why cloud initiatives fail.
Where leaders succeed
A deliberate, well-planned approach toward operational maturity is the most effective way to succeed in the cloud. Across Hewlett Packard Enterprise's base of cloud transformation engagements, leaders display this kind of discipline in these four areas:
- Service operations: The good news is that organizations have grown more sophisticated in their ability to log a wide variety of activities. The more advanced organizations are doing a better job curating and identifying the right pieces of metadata to turn into insights. They have meta-models in place, but they need to put intelligence around them to put the data to work for them.
- Architecture and governance: Formal governing bodies have been established, but many organizations still are working to streamline decision-making processes so they can reduce lead time.
- Availability management: Leaders are developing their ability to define business availability requirements in SLAs. They have well-thought-out plans to improve metrics such as RTO (recovery time objective) and RPO (recovery point objective), especially for tier-one and tier-two applications. Where organizations still need help is in meeting those requirements. Many establish partnerships with vendors that can take on management functions to help achieve these objectives.
- Platform operations: Leaders have created a mechanism for meta-data-based visibility across their ecosystems. This is important when they integrate with CI/CD pipelines; they know which components to pull from their repositories and where to put them. Infrastructure patterns and standards are important here too. A major bank we worked with was looking to increase adoption of its own internal private cloud. By using well-defined infrastructure patterns and corresponding storage standards, it was able to reduce infrastructure provisioning time, which in turn improved the reputation and buy-in for its initiative.
Getting your house in order
Moving to the cloud presents significant opportunities for companies to transform their operations—to make them more efficient and more focused on delivering business value. But to mature to the point where they're accomplishing meaningful transformation, organizations need to get their own houses in order.
They need to commit to a new operating model, assess their strengths and weaknesses, and forge a plan to set their operations up for success. This operating model will yield benefits not only for the portion of operations that move to public cloud but across the entire enterprise, edge to cloud.
If IT operations don’t prioritize the optimization of a comprehensive cloud operating model, some application owners will head to the cloud by themselves.
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.