Your company’s early DevOps projects went off without a hitch, resulting in automated deployment, continuous integration/continuous delivery, and other promised benefits. So why did your DevOps plans go off the rails when you tried to implement them throughout the organization?
You may have succeeded in changing the processes employed in your pilot projects, but you failed to adapt your IT department’s culture to the DevOps way of doing things. In an October 17, 2017, article on DataQuest, Siba Prasad Pulugurty cites a Gartner report forecasting that by 2018, 90 percent of I&O organizations attempting to implement DevOps without altering their culture will fail.
Here is Pulugurty’s seven-step plan for establishing a DevOps culture:
- Break down silos by ensuring that all teams collaborate seamlessly with one another — they all must function as “one team.”
- Early in the process, make sure all stakeholders understand and agree on their shared objective.
- Train from the bottom up (line workers first) to prepare everyone for the change in work processes.
- Welcome failures as a sign of progress, but only if they are detected early and corrected quickly.
- Implement DevOps bit-by-bit, choosing key projects, and continue to run existing systems in parallel with new ones (bimodal IT).
- Reinforce the “one team” concept by emphasizing the importance of continual communications and transparency among all parties for all projects.
- Remember that your project’s success will rely not on the automation of systems, but rather on the collaboration of humans.
When expanding the use of DevOps in an organization, information issues and technology issues are dwarfed by process issues and people issues. Source: Gartner
I&O configuration/change management proliferates
The emphasis on cultural change for successful implementation of I&O automation is echoed by Forrester principal analyst Robert Stroud in an October 31, 2017, article on Information Management. Stroud identifies two important benefits of continuous delivery and release automation: the ability to deploy automatically in multiple formats, and to restore the environment instantly should a particular deployment fail.
The onus is on I&O professionals to shorten their existing change-management practices by automating manual procedures. As Stroud puts it, I&O must “speed up deployment cadence.” The trick is doing so while also validating the quality of the deployed applications just as quickly. This requires the choice of the most effective metrics for the particular system, ensuring that the metrics “align to velocity throughput and success.” No single DevOps tool is capable of collecting, analyzing, and reporting crucial performance metrics on hybrid cloud systems. According to Gartner, organizations will mix and match I&O automation technologies from 20 or more services for many years to come.
Research firm Gartner expects the number of separate I&O automation technologies to proliferate, requiring a unified method for planning, implementing, and managing automation components. Source: Gartner, via Advanced Systems Concepts
Before selecting their custom automation toolset, companies need a big-picture view that identifies the needs of all areas. As Jim Manias writes in a September 20, 2017, article on ITProPortal, the goal of automation is to streamline workflows and processes. These includeworkload automation, IT process automation (ITPA), and application release automation (ARA).
A clear-cut plan keeps you ahead of the I&O automation jigsaw puzzle
Most companies find themselves with a patchwork of automation tools in place. It’s natural that staff picked up a tool on this DevOps project and another tool on that project. When you transition to an enterprise-wide approach to automation, you have to account for the many disparate tools already in use.
The resulting mish-mash lacks any central control. Even worse, you’ve set off on a journey without knowing your destination. As Manias points out, this is a recipe for wasted resources and underutilized IT skills. The solution is to hit the pause button and devise a comprehensive automation infrastructure, even if this means pulling back from some existing projects.
The benefits of a unified I&O automation infrastructure are the provision of a single point of control, the ability to incorporate new technologies seamlessly, and “strategic resource utilization.” In addition, the improved transparency of a unified automation platform simplifies governance. However, enhanced manageability is only one of several factors that need to be addressed. The long-term success of your I&O automation strategy hinges on the three IT standbys: Security, efficiency, and usability.
Make the break from ‘deterministic’ automation
The pattern for most organizations is to implement automation one manual process at a time. While the piecemeal approach leads to lower costs and improved quality of service in discrete areas, it doesn’t scale, so the benefits can’t be extended throughout the enterprise. Vijaya Shanker writes in an April 27, 2017, article on CXOtoday that the current “deterministic” approach to I&O automation must give way to the creation of a comprehensive service automation management strategy.
Shanker highlights three keys to service automation:
- Reducing the need for human intervention eliminates the number-one source of system errors.
- Auto-remediation, automatic resource fulfillment, and auto-provisioning deliver process efficiencies that translate directly into improved productivity.
- All the efficiencies in the world are worthless if they don’t make life easier for your customers in business units, so keep their needs in mind first, last, and always.
Key technologies enabling I&O automation are infrastructure as code (IaC), which adds a software layer atop the physical infrastructure; and containers, which make it possible to define and manage complete application stacks as standardized text files and images. Source: Gartner
Now that hybrid clouds are the mainstay of IT operations, your workloads must migrate seamlessly between the public cloud and local hosts. Creating the unified platform hybrid clouds require begins and ends with the physical infrastructure. Gartner’s 2018 Planning Guide for Infrastructure and Operations (Gartner account required) emphasizes the critical role automation plays in providing the scale and agility today’s cloud operations depend on.
Gartner executive Tony Iams writes in an October 12, 2017, post that automation frees I&O staff to work on activities that provide “real business value” rather than on “low-impact management tasks.” In addition, you’re better able to change your infrastructure quickly to keep pace with “evolving business requirements.” Another benefit of infrastructure as code (IaC) is immutability, which ensures that the individual components comprising your hybrid network are always in a known-good-configuration state.
The elimination of siloed “snowflake” applications and subsystems means that I&O staff must broaden their view of business processes. The new collaborative environment means any new architecture component can impact all existing and future systems. No longer can a single team or team member claim total control over any slice of the operating environment. Once again, a culture of sharing, openness, and transparency is paramount as a prerequisite for modern I&O management.
Hybrid cloud automation relies on unified policy enforcement
I&O leaders have been deterred from jumping on the hybrid cloud bandwagon by the inability to automate enforcement of a single set of policies uniformly across the public and private components of their hybrid setup. As Rhett Dillingham explains in a June 1, 2017, article on Forbes, without automated policy enforcement, application owners are free to pick and choose which cloud services meet their specific needs.
The result is a disjointed collection of “bespoke application implementations” that vary in their approach to security, capacity management, and tool selection. Despite the recent flood of cloud management services from top vendors, there is still no comprehensive hybrid cloud solution that enterprises can implement right out of the box.
The need for seamless, transparent service automation for hybrid clouds is heightened for I&O leaders because CEOs and CFOs increasingly look to cloud solutions as a way to improve productivity and reduce expenses. To address these expectations, I&O managers will consolidate to a handful of cloud vendors to “maximize discounts and reduce training and expense costs,” according to Dillingham.
Now that enterprises have acknowledged the central role of hybrid clouds in their business strategies, it falls to I&O leaders in the organization to deliver the consistency, reliability, and security their customers have come to expect from their experiences with on-premises systems. Your plan for automating service management for hybrid clouds starts with a single control point from which a uniform set of policies can be enforced on all the various public and private components.