Here's how we solve the most common cloud hosting (AWS) problems

As Harborn, we are operationally responsible for hundreds of production environments. Environments that vary widely in size, security requirements and complexity. One thing they all have in common, however, is the platform in the public cloud on which all these environments run. This platform was developed by Harborn and forms the basis of our cloud deployments. The platform provides a standardised interpretation of the various environments, based on industry-wide best practices. In this series of articles, we are going to give a sneak peek into how this platform came into being, and what problems it solves for the customer, and therefore for us. We'll take you through why we believe such a platform is essential in the face of ever-increasing complexity in operational IT.

Harborn has been using the cloud for more than 15 years. During that time, mainly through trial and error, we learned very clearly what the do's and don'ts are when using the cloud. That includes how to properly set up a FinOps (cloud financial management) process, but also how to monitor consistency in one (or more) environment(s), while losing as little as possible of the flexibility that makes the cloud so powerful.

Before starting a project or a product, we like to think carefully about what problem such a project or product actually solves. To understand properly first before starting with the solution. We use tools like Focus and Design Sprint, to answer for ourselves the important questions like Who, What, Where, When, but mainly the Why.

In the case of our cloud platform solution, our Focus Sprint revealed quite quickly that, in this case, it’s very important to have a clear vision and mission to articulate and follow, and that the big danger of cloud deployment is not having insight.

The cloud is complex. Many people still struggle to understand exactly what the cloud can mean, and many IT parties still treat the cloud as if it were a classic on-premise solution, while payments are often by the minute or second, and an average cloud invoice is made up of as many as a hundred different components.

As far as we’re concerned, the solution direction that fits these challenges is to have a “best practice” setup. An out of the box solution that already ticks a lot of boxes, regarding things we often forget when setting up a cloud environment. For example, a good Continuous Integration, Delivery and (where appropriate) Deployment pipeline for the Infrastructure as Code, well-designed backups, observability of serverless functions, but also for the roll-out of applications on a Kubernetes platform within which everything has been thought of.

We have chosen a port as the metaphor to describe this platform. The containers in the port symbolise the workloads (projects) we run in the cloud (often in the form of Docker containers). Containers have a standardised outer size, and a considerable variety of specialised additions. Consider, for example, a container with built-in cooling, or a container fitted with special doors. The content of a container is flexible, just like the workloads of our customers. The container ships that are moored waiting to be transhipped indicate the volume of “work” that the platform has to handle at that moment. The volume of containers being transited at night is significantly less than during the day, similar to a platform having to handle considerably more or less traffic at set times, or when a new fully loaded container ship is offered.

So, the working name of our cloud platform is De Rotterdamse Haven (Port of Rotterdam). In the next blog, you will read more about what cloud hosting can do for your business.