Delivering DevOps through Platform Engineering

In software engineering, the Golden Path aims to provide the smoothest way forward via a self-service template for common tasks. It is enabled by platform engineers – who provide developers with the simplest possible internal developer platform and the tools they need to deliver innovation.

Here we look at the emerging discipline of platform engineering, and the benefits it brings to application development via easier and faster access to services and resources, in particular using modern data management platforms built on Kubernetes containerised environments.

When DevOps emerged in the late 2000s, it brought with it key principles of shared ownership, rapid feedback and workflow automation to help deliver on the vision of agile software development. It requires a high degree of autonomy for the developers and in exchange empowers them with the tools they need to be efficient.

Automation is one of the key principles of DevOps since the quick pace of changes it drives is incompatible with “human in the loop” workflows.

The mode of operations preferred by developers (and many technical specialist roles such as data scientists, AI researchers etc.) can often be boiled down to 3 main asks:

Using the above as the “north star” when building services geared towards technical profiles is a great way to enable innovation and ensure fast adoption. While providing instant resources and results might not be always possible, getting as close as possible to instant will drive greater satisfaction.

Today, we see the coming of age of DevOps through the rise of platform engineering, a new function for a more mature era in application development, that provides a suite of self-service tools to empower developers. Platform engineering operates behind the veil to provide an easy-to-use, self-service catalogue of services and infrastructure components to support the day-to-day development experience.

Best practice platform engineering aims to help application developers get on board and start building faster by providing everything that they need to experiment, develop, test and deploy. The platform made available to these developers often takes inspiration from the services popularised by the public cloud and its mode of operation. It is designed to provide instant access to not just the latest and greatest tools and software that underpin innovation, but also provide easy access to the data itself, protected by pre-determined guardrails and security protocols.

The ideal developer-focused platform also includes data management. It may build on top of Kubernetes as the means to orchestrate, deploy, run and scale cloud-native applications as well as to manage the data services required for those applications. Data management capabilities are key to platform engineering because they enable exploration and testing in realistic conditions, for example using an instant copy of production data instead of a somewhat unrealistic synthetic data set.

Ideally, the data management capabilities will also be designed with self-service in mind, and deliver access to data in a highly available, reliable, elastic, multi-tenant and secure manner. Portworx from Pure Storage is an example of such a modern data platform. Fully integrated with Kubernetes, it allows the developer to easily get access to persistent data options (including data protection capabilities such as data replication, backup and archiving) but also to data sets themselves through instant data cloning, even enabling the use of self-service instant snapshot creation and restore so that developers may experiment with changes and roll back to previous states quickly and easily.

Additionally, Portworx Data Services provides a catalogue of curated data services, including MongoDB, Elasticsearch, Cassandra, Kafka and PostgreSQL, simplifying deployment into just a few clicks or a single API call, so that developers can deploy or scale these data services easily with the optimal data storage configuration and protection. This foundation brings these easy-to-specify toolchains and data services to the developers so that they can easily use them as building blocks, even if they don’t have extensive knowledge of Kubernetes or how to deploy a given database engine in a secure and scalable manner.

Platform engineering teams are busy working unseen in the background to bring the self-service Golden Path to application development. With Kubernetes as the orchestration framework, and containers and data services as key resources, the platform engineers can finally deliver fully on the vision of increased agility and greater productivity of DevOps.