There is still an enormous amount of repeatable work done by people in many established organizations. Think of administration, monitoring, reviews, and responding to tickets among other tasks. Over the last decade, cloud vendors have demonstrated how automating processes that remove repetitive work can help increase overall efficiency. Automated processes are consistent and auditable, which can help reduce errors and improve quality. It can also free skilled tech talent to focus on higher value-added tasks.
IT leaders, for various reasons, have been slow to pursue these opportunities. This, however, is beginning to change. In what we recognize as an emerging trend, some CIOs are disrupting their organizations and the army of technologists that currently execute many manual tasks and handoffs across systems, architecture, development, and deployment.
Beyond leveraging investments made by cloud providers to accelerate their journeys, CIOs are following the cloud providers’ playbook to identify and standardize processes. They are attacking opportunities in infrastructure, software components, security, and applications. Once their enhancements mature, CIOs and their teams optimize the new service delivery and automation using advanced techniques such as AI and ML.
Early participants in this trend have already seen gains in efficiency and lower labor costs. In a recent survey of IT and engineering leaders, 74% of respondents said that automation has helped their workforce work more efficiently. Fifty-nine percent reported cost reductions of up to 30% on teams that have embraced process automation.1 Add to this noticeable increases in quality and security, and it becomes clear why 95% of respondents are prioritizing process automation, with 21% saying it’s a high priority.2
The pace of change only continues to increase. The business is asking for more, and they want it more quickly than ever before. The talent market is white hot, with growing demand for advanced skill sets (that are in perpetually short supply.) Everyone is trying to do more with less.
The time to (finally) disrupt IT is now.
The journey from manual to automated activities isn’t new. Indeed, in previous Tech Trends reports, we have examined this transition in areas such as cybersecurity, advanced networking, and the dynamic provisioning of hardware and software. So, what is different this year? Simply put, competition. The pandemic is upending the labor market. Perhaps more importantly, it’s in digital natives’ DNA to push automation to its limits. Hence, startups can achieve greater scalability, reliability, resilience, and efficiency at lower costs than their established counterparts. They hold an additional advantage in that they aren’t held back by technical debt or organizational compromises that require handoffs and manual interventions. For digital natives, such old-school actions become a last resort rather than the norm. This approach is fundamentally different from those often taken by established organizations. Today’s competitive marketplace requires a more robust IT posture, which can translate into a competitive advantage.
Organizations looking for opportunities to disrupt their status quo can focus their efforts in three areas:
Standardize and automate on-premises infrastructure
The first leg of an automation journey involves enabling all infrastructure and management functions to be controlled by code. Programmatic control of resources makes it possible to apply policies consistently and to store previously manual configurations in automated code and configuration files. These solutions require deploying some mix of compute (containers, virtualized servers, and functions), networking (software-defined), and storage.
For automation to scale, processes must be executed consistently across the enterprise. However, if you look at the operational landscapes of many organizations today, you will find a mixed bag of processes, applications, and workarounds. When processes work in one fashion on server A and another on server B; when environments do not have parity; or when networks behave differently, then operations become more costly and inefficient.
If this sounds familiar, consider creating a standard, common approach to developing, deploying, and maintaining your solutions and components. Cloud vendors realized early on that the more you can programmatically control resources, the easier it becomes to treat environments as a program to be managed. Many of today’s infrastructure-as-code platforms trace their roots to early cloud-based automation initiatives.
As organizations explored infrastructure-as-code, they recognized they could also deploy security-as-code or operations-as-code, controlling them all with configuration or code files. The goal of ”as-code” is to push toward an environment in which everything—even bespoke systems—aligns on a set of optimized rules. With rules in place, a single engineer can control a large pool of resources that would have taken several administrators to manage. This frees infrastructure teams to work like the cloud providers: automating, taking advantage of opportunities for self-service, and getting out of the way.3
As organizations streamline operations and management with automation, they should also revisit their initiation processes. Historically, creating new infrastructure involved elaborate procurement exercises with escalating approval levels. In today’s world, adding another virtual instance may not warrant any level of prior approvals. Identifying and automating (or eliminating) similar handoffs and approvals that made sense in a legacy environment can help streamline operations, augment developer productivity, and provide much sought-after organizational agility.
When approached methodically and strategically, automation can deliver significant economies of scale. It offers other benefits as well:
- Greater accuracy. Individuals will no longer be subjectively interpreting documents, queries, and forms.
- Increased security and resilience. Rules will be applied more consistently. It’s worth noting that a nascent “security-as-code” trend is gaining momentum.
- Improved reliability. Problems fixed in the code typically won’t reoccur.
We offer a word of warning for those using the providers’ “as code” services. Make sure you have organized your processes and operations to get the most out of those capabilities. If you don’t, you may reproduce existing limitations in a modern environment.
Standardize and automate software, management tools, and applications
Leading-edge IT organizations no longer manage infrastructure; they now develop code that manages infrastructure, an approach that can boost scalability, efficiency, and consistency. This same approach can apply to software components, management tools, and a variety of applications. Modern IT organizations manage software code that, in turn, manages aspects of development, maintenance, operations, and security. Ultimately, it is easier to manage a single piece of code than an array of manually configured solutions. For example, infrastructure-as-code enables us to bring the agility of software development to infrastructure management. From a deployment perspective, it is possible to manage full-stack solutions rather than separate components that several teams must coordinate.
Like infrastructure, a few on-premises software components are prime candidates for automating. For example, database management, integration tools, security, systems management, and O/S patching can be easily virtualized and abstracted.
For organizations using cloud infrastructure, vendors offer an expanding menu of platform-as-a-service (PaaS) options that feature enhanced automation, programming interfaces, integrated middleware, and management capabilities. Maturing PaaS offerings may also provide enhanced developer self-service, programming interfaces, and more tightly integrated middleware and management capabilities.
How can you decide where to start? First, identify the “user journeys” of those trying to deliver functionality to end users, and the points of friction these users encounter. Ruthlessly prune unnecessary approvals and handoffs and then automate or create self-service options for the steps that stand between developed code and production deployments. Finally, once your automation journey is underway, old performance metrics may no longer apply. It’s important at this point to incentivize an “automation culture” by defining metrics for the organization you want to become.
Optimize automation with ML and rules
Typically, a first pass at automation is rules-based. For example, “if process x doesn’t respond, restart the process.” Over time, IT staff members can identify issues that cause outages and malfunctions and optimize automation tools to address them, just as cloud providers demonstrated a decade ago. Eventually, you can move beyond rules-based to machine learning–based automation. An automation journey that begins immaturely can subsequently grow in terms of sophistication.
Many types of ML—predictive, capacity modeling, action response, outage recovery, among others—support different IT activities. Yet for most organizations, identifying outages early and harnessing predictive modeling to prevent future outages is a top ML priority. By focusing on these areas, ML-enabled teams can measurably improve uptime and decrease outage severity. Moreover, a growing number of PaaS offerings feature embedded ML capabilities. For example, PaaS offerings often use ML to maintain and optimize routine operations that were previously managed manually by developers, administrators, and engineers. The net effect is that development and operations can run in higher gears.
Another optimization technique involves applying rules consistently. Consider this: Enterprise architecture is a set of decisions around what you can use and how. The resulting rules represent an optimum approach to architecture design and function. As part of your automation journey, consider prioritizing consistency. Do this by methodically embedding these rules into systems and processes across the enterprise. Consistency delivers optimum performance.
The way forward
For CIOs and other leaders who are exploring automation opportunities, time is of the essence. In today’s rapid-fire innovation climate, there is not a lot of business value to be found in paying humans to maintain servers and data centers. As CIOs disrupt their IT organizations with automation, there will be ripe opportunities to shift employees’ focus from patching, monitoring, and measuring to higher-value engineering activities. More broadly, automation’s possibilities extend to areas such as development, deployment, maintenance, and security, thus making it possible to gain efficiency and consistency across more of IT’s operations.
The journey from managing things to managing code that manages things won’t happen overnight. For example, there may be some cultural resistance from tech workers and the C-suite, or legacy systems may have manually configured components that make automation difficult. Finally, change is hard, even for the nimblest IT teams. People accustomed to handoffs and human-to-human interaction may adapt slowly to self-service and automated provisioning. For organizations just getting started, it may be helpful to create a dedicated team that develops and deploys automation and self-service to standard processes. This team can methodically broaden its approach, transforming more processes over time as they climb the stack.
Luckily, some needed automation is readily available in the form of cloud-based solutions. The rest is achievable through engineering and a deliberate, consistent focus on building an automated future.