Cloud computing has become established as a major enabler for pursuing new business opportunities, and many organizations are adopting cloud-first strategies for deploying new workloads. While cloud services are clearly becoming more central to IT, private infrastructure remains highly relevant, even as its operation takes on increasingly cloud-like attributes. Infrastructure plays a critical role in integrating existing on-premise resources with public cloud services, requiring infrastructure managers to embrace the skill sets, practices and procedures needed to accommodate cloud-based operations. Further, when workloads migrate from the public cloud back to localized hosting (either for reasons of cost efficiency, or in support of IoT initiatives such as edge computing), organizations expect to apply the same operating style with their private infrastructure as in the public cloud. The operation of private infrastructure therefore has to evolve to take on the same model as public cloud services, i.e. optimized for agile delivery of shared services, leveraging DevOps management tools and processes, and designed with failure in mind.
Gartner recently published its 2018 Planning Guide for Infrastructure and Operations (I&O) [Gartner clients can access the document here], which provides guidelines for technical professionals to expand the skill sets, practices and procedures needed to accommodate cloud-based operations. The report highlights the growing role that automation plays in delivering the scale and agility necessary for cloud operations. Staffing growth cannot possibly keep up with the increased demands that the cloud imposes on infrastructure. Automation gives administrators the necessary leverage to efficiently manage the resources that are available, freeing I&O personnel from having to spend time on low-impact management tasks so that they can focus more on activities that provide real business value. The elimination of manual configuration tasks also enables infrastructure to change more quickly in response to evolving business requirements.
Several key architectural patterns help to enable automation, including infrastructure as code (IaC), which may be based on either scripts for driving legacy infrastructure or on new API-driven infrastructure, i.e. a layer of software that sits on top of physical infrastructure and enables automated control. Examples include Software Defined Data Centers (SDDC) and Hyperconverged Integrated Systems (HCIS). Containers are another important enabler of automation, in that they allow complete application stacks to be defined and managed as standardized text files and images. These methods all enable infrastructure to be configured and maintained using artifacts that are managed the same way as application software.
When these artifacts are coupled with revision control tools and frameworks for automated configuration such as Bosh/Kubo, Packer/Terraform, and Ansible/Puppet, it becomes possible to apply proven lean and agile principles to infrastructure management, using the same toolchains and iterative continuous integration/continuous delivery (CI/CD) pipelines that are used for applications. This representation of infrastructure state also enables immutability, an architectural pattern that ensures systems consistently remain in a known-good-configuration state in order to eliminate “snowflakes”, i.e. uniquely configured systems that are difficult or impossible to replicate if they have to be rebuilt.
This kind of operating model reduces much of the day-to-day administrative burden on I&O teams, but that does not mean that these teams will become obsolete. However, it forces them to confront the limits of their control, making them prioritize what aspects of operation they turn over either to automation or external cloud services. As part of this evolution, I&O teams will have to move up to the stack in order to stay relevant. Indeed, consistent toolchains and workflows for managing infrastructure help to transform siloed, isolated I&O teams into much more collaborative environments. In these environments, I&O must consider the impact of new architectures, such as those implemented by cloud, containers and microservices on infrastructure, while sharing the responsibility for and control of the operating environment with other teams, like application developers. I&O’s role will increasingly become one of a broker, whose main responsibility is to identify and provision the correct infrastructure service for a use case, in the right location and at an appropriate level of reliability and performance.